Increased 90-kDa ribosomal S6 kinase (Rsk) activity is protective against mutant huntingtin toxicity

Background The 90-kDa ribosomal S6 kinase (Rsk) family is involved in cell survival. Rsk activation is regulated by sequential phosphorylations controlled by extracellular signal-regulated kinase (ERK) 1/2 and 3-phosphoinositide-dependent protein kinase 1 (PDK1). Altered ERK1/2 and PDK1 phosphorylation have been described in Huntington's disease (HD), characterized by the expression of mutant huntingtin (mhtt) and striatal degeneration. However, the role of Rsk in this neurodegenerative disease remains unknown. Here, we analyzed the protein levels, activity and role of Rsk in in vivo and in vitro HD models. Results We observed increased protein levels of Rsk1 and Rsk2 in the striatum of HdhQ111/Q111 and R6/1 mice, STHdhQ111/Q111 cells and striatal cells transfected with full-length mhtt. Analysis of the phosphorylation of Rsk in Hdh mice and STHdh cells showed reduced levels of phospho Ser-380 (dependent on ERK1/2), whereas phosphorylation at Ser-221 (dependent on PDK1) was increased. Moreover, we found that elevated Rsk activity in STHdhQ111/Q111 cells was mainly due to PDK1 activity, as assessed by transfection with Rsk mutant constructs. The increase of Rsk in STHdhQ111/Q111 cells occurred in the cytosol and in the nucleus, which results in enhanced phosphorylation of both cytosolic and nuclear Rsk targets. Finally, pharmacological inhibition of Rsk, knock-down and overexpression experiments indicated that Rsk activity exerts a protective effect against mhtt-induced cell death in STHdhQ7/Q7 cells transfected with mhtt. Conclusion The increase of Rsk levels and activity would act as a compensatory mechanism with capacity to prevent mhtt-mediated cell death. We propose Rsk as a good target for neuroprotective therapies in HD.

Briefly, sequential phosphorylations are initiated by ERK1/2 at Thr-573/574 of CTKD leading to the autophosphorylation of Rsk at Ser-380. This phosphorylation allows the dockage of PDK1 to the hydrophobic motif and enables PDK1-dependent phosphorylation in the NTKD of Rsk at Ser-221, resulting in its maximal activation [1,4]. When activated, Rsk promotes the phosphorylation of many cytosolic and nuclear targets. In the cytosol, Rsk induces the inactivation of certain proapoptotic proteins, such as Bad [5], glycogen synthase kinase 3β (GSK-3β) [6] or death-associated protein kinase (DAPK) [7], whereas in the nucleus it activates transcription factors involved in the synthesis of antiapoptotic proteins, namely cAMP response element binding protein (CREB) [8], serum response factor (SRF) [9], and IBα [10,11]. Although the function and the mechanism of Rsk activation have been well studied in non-neural cells, in neurons there are few studies about Rsk, and they associate its activity with the anti-apoptotic effect of trophic factors [12][13][14]. However, no data exists about the possible role of Rsk in neurodegenerative diseases.
Huntington's disease (HD) is a neurodegenerative disorder caused by a dominantly heritable expansion of a trinucleotide CAG repeat in the huntingtin (htt) gene [15], and characterized by the preferential neurodegeneration of striatal medium-sized spiny neurons [16]. Although the brain areas affected by the disease are well established, the mechanisms by which neural dysfunction and neurodegeneration occurs are not well defined yet. Interestingly, previous data from a HD cellular model show a de-regulation of both kinases that control Rsk activity. Knock-in striatal cells expressing full-length mutant huntingtin (mhtt) (STHdh Q111/Q111 ) show increased levels of active PDK1 [17] and reduced levels of ERK1/2 activity [18] compared with striatal cells expressing wild-type htt (STHdh Q7/Q7 ). Moreover, stimulation of these kinases and their pathways has been proposed as good therapeutic approaches for HD [19][20][21]. These results suggest a de-regulation of Rsk activity in HD models and that modulation of its activity could be a good therapeutic strategy. Therefore, here we studied whether the protein levels and activity of Rsk1 and Rsk2, the two isoforms with higher expression levels [1], are modified in the presence of mhtt. To this end, we analyzed striatal protein levels and activity of Rsk in knock-in mhtt mouse and cellular models. In addition, we studied the contribution of ERK1/2 and PDK1 to the activation of Rsk in the presence of mhtt. Finally, we evaluated the potential protective role of Rsk against mhtt toxicity.
To know whether increased Rsk1 and Rsk2 protein levels also occur in exon-1 mhtt mice we analyzed by western blot these proteins in the striatum of R6/1 mice at 8-and 12-week of age, when they do not show motor symptoms [22]. Similar to that observed in the striatum of knock-in mice, R6/1 mouse striatum displayed higher Rsk1 and Rsk2 levels compared to wild-type (WT) mice at both ages ( Figure 1E and 1F). Altogether, these results indicate that an increase of Rsk1 and Rsk2 protein levels is an event that occurs in full-length and exon-1 models of HD at presymptomatic stages. In addition, changes in Rsk1 and Rsk2 are not dependent on mhtt levels since we observed a similar response in striatal cells expressing low (STHdh Q111/Q111 cells), normal (knock-in mice striatum) or very high (R6/1 mouse striatum) levels of mhtt.

Overexpression of full-length mhtt increases Rsk protein levels
In order to confirm that increased levels of Rsk were dependent on mhtt expression, we looked at the protein levels of Rsk1 and Rsk2 in STHdh Q7/Q7 cells, M213 cells and striatal primary neurons transfected with a plasmid expressing full-length wild-type (FL-17Q htt) or mutant (FL-75Q htt) htt. The quantification of Rsk1 and Rsk2 levels was performed by confocal microscopy due to the low efficiency of transfection (15-20% approximately). In all cell types examined, transfection with FL-75Q htt increased Rsk1 and Rsk2 protein levels compared to those registered in cells expressing FL-17Q htt ( Figure  2), indicating that the increase in Rsk1 and Rsk2 protein levels are due to the presence of mhtt.
Rsk phosphorylation in HD knock-in models: ERKdependent residues versus PDK1-dependent residues To study whether the phosphorylation levels of Rsk were altered by changes in total Rsk protein levels, we analyzed its phosphorylation at Ser-380 (dependent on ERK1/2) and at Ser-221 (dependent on PDK1), in the striatum of 10-month old Hdh Q7/Q7 and Hdh Q111/Q111 mice. We detected reduced levels of phospho-Rsk (Ser-380; reduction of 63 ± 13%; Figure 3A) and increased levels of phospho-Rsk (Ser-221; increase of 190 ± 19%; Figure 3A) in Hdh Q111/Q111 respect to Hdh Q7/Q7 mice. Similar results were obtained in STHdh Q111/Q111 cells ( Figure 3B). These results indicate that in the presence Figure 1 Rsk levels are elevated in HD mouse and cellular models. Rsk1 (A, C and E) and Rsk2 (B, D and F) protein levels were analyzed by western blot of protein extracts obtained from the striatum of 6-and 10-month old wild-type (Hdh Q7/Q7 ) and knock-in (Hdh Q111/Q111 ) mice (A and B), from wild-type (STHdh Q7/Q7 ) and mutant (STHdh Q111/Q111 ) htt knock-in striatal cells (C and D), and from the striatum of 8-and 12-week old WT and R6/1 mice (E and F). Htt protein levels were also analyzed by western blot in knock-in models (A-D). Results (mean ± SEM; n = 4-6) represent the ratio between Rsk and actin levels obtained by densitometric analysis of western blot data, and are expressed as a percentage of Hdh Q7/Q7 levels at 6 months (A and B), as a percentage of protein levels in STHdh Q7/Q7 cells (C and D), or as a percentage of protein levels in WT mice at 8 weeks (E and F). Data were analyzed by two-way ANOVA followed by Bonferroni's post hoc test (A, B, E and F) or by Student's ttest (C and D). *p < 0.05 and **p < 0.01 as compared with Hdh Q7/Q7 mice (A and B), **p < 0.01 and ***p < 0.001 as compared with STHdh Q7/Q7 cells (C and D), and *p < 0.05 and **p < 0.01 as compared with WT mice (E and F). Representative immunoblots are presented.
of mhtt the phosphorylation of Rsk at ERK-and PDK1dependent residues is altered in an opposite way.

STHdh Q111/Q111 cells show increased Rsk activity that is mainly regulated by PDK1
Our next goal was to know whether increased Rsk1 and Rsk2 protein and phosphorylation levels were associated with elevated Rsk activity. To this end, we analyzed Rsk activity in knock-in cells by using an in vitro activity assay. We observed that Rsk activity was higher in STHdh Q111/Q111 than in STHdh Q7/Q7 cells (270 ± 15%; Figure 4A). Moreover, overexpression of Rsk by transfection of HA-Rsk1 in STHdh Q7/Q7 cells increased Rsk activity (STHdh Q7/Q7 cells: 100 ± 9%; STHdh Q7/Q7 + HA-Rsk: 222 ± 13%; p < 0.0002; Student's t-test) indicating that one of the parameters that regulates Rsk activity is its protein levels.
To address the importance of phosphorylation by ERK1/ 2 and PDK1 on elevated Rsk activation in STHdh Q111/Q111 cells, we measured Rsk activity in knock-in cells transfected with two mutant forms of Rsk: HA-RskT574A and HA-RskS380E, which cannot be phosphorylated by ERK1/ 2 and PDK1, respectively. Transfection with HA-RskT574A or HA-RskS380E similarly reduced Rsk activity in STHdh Q7/Q7 cells (reduction of 28% and 33% respectively; Figure 4B). Interestingly, and supporting a main role for PDK1 in the increased Rsk activity observed in STHdh Q111/Q111 cells, transfection with HA-RskS380E induced a stronger decrease of Rsk activity (47%; Figure  4B) compared with the transfection with HA-RskT574A (19%; Figure 4B). Note that co-transfection with both mutant forms reduced the activity of Rsk only by 30-40% probably because the efficiency of transfection was not maximal ( Figure 4B).
Rsk levels are increased in both cytosol and nucleus of STHdh Q111/Q111 cells Phosphorylated and activated Rsk can translocate from the cytosol to the nucleus. In these compartments, it regulates different targets [1]. Thus, we studied Rsk1 and Rsk2 levels in cytosolic and nuclear fractions of knock-in cells by western blot. When compared with control cells, STHdh Q111/Q111 cells displayed enhanced levels of Rsk1 and Rsk2 in both compartments, with a more pronounced effect in the nucleus ( Figure 5A). To confirm these data, we analyzed the localization of Rsk by immunocytochemistry. We detected three different patterns of expression: homogeneous expression, and exclusive cytosolic or nuclear localization ( Figure 5B). Analysis of STHdh Q7/Q7 cells revealed a predominant homogeneous distribution of Rsk1, whereas Rsk2 was mainly located in the nucleus. In STHdh Q111/Q111 cells, Figure 3 Different regulation of phospho-Rsk residues in HD knock-in mice and STHdh Q111/Q111 cells. Lysates from the striatum of 10month old wild-type (Hdh Q7/Q7 ) and knock-in (Hdh Q111/Q111 ) mice (A) or from wild-type (STHdh Q7/Q7 ) and mutant (STHdh Q111/Q111 ) htt knock-in striatal cells (B) were subjected to western blot to analyze phosphor-Rsk (Ser-380; Ser-221), Rsk1, Rsk2 and tubulin protein levels. Results (mean ± SEM; n = 4-6) represent the ratio between phospho-Rsk levels and Rsk1 plus Rsk2 levels obtained by densitometric analysis of western blot data, and are expressed as a percentage of Hdh Q7/Q7 (A) or STHdh Q7/Q7 (B) levels. Representative immunoblots are presented. All data were analyzed by Student's t-test. *p < 0.05, **p < 0.01 and ***p < 0.001 as compared with Hdh Q7/Q7 mice (A) or STHdh Q7/Q7 cells (B).
Rsk1 changed its distribution as it was located only in the nucleus, while the nuclear expression of Rsk2 was even more evident than in STHdh Q7/Q7 cells. Note that we did not observe exclusive cytosolic localization of either Rsk isoforms in STHdh Q111/Q111 cells ( Figure 5B). Altogether, these results indicate that although the increase of Rsk protein levels in STHdh Q111/Q111 cells occurs in both compartments, this increase is more pronounced in the nucleus.

Increased activity of Rsk in STHdh Q111/Q111 cells enhances the phosphorylation of both cytosolic and nuclear targets
Rsk plays its protective role through the inactivation of cytosolic pro-apoptotic proteins and/or the activation of transcription factors that mediate the synthesis of antiapoptotic proteins. Thus, we determined in knock-in cells expressing wild-type or mhtt the phospho-levels of two Rsk substrates, Bad at Ser-112 (cytosolic target), and SRF at Ser-103 (nuclear target). According with the elevated Rsk activity observed in STHdh Q111/Q111 cells, we found increased levels of phospho-Bad ( Figure 6A) and phospho-SRF ( Figure 6B) respect to STHdh Q7/Q7 cells. Then, to determine whether the increased phosphorylation of Bad and SRF was due to the action of Rsk, we treated knock-in cells with a pharmacological and specific inhibitor of Rsk, BI-D1870 (0.1 μM) [23]. The presence of BI-D1870 reduced the phosphorylation levels of both Bad and SRF in STHdh Q111/Q111 cells (Figure 6A and 6B). In STHdh Q7/Q7 cells, we did not observe changes in phospho-Bad levels in the presence of BI-D1870 ( Figure 6A), whereas phospho-SRF levels were slightly decreased ( Figure 6B). To corroborate that BI-D1870 efficiently inhibited Rsk, we tested the activity of Rsk in both cell lines after treatment with Rsk inhibitor. Addition of BI-D1870 (0.1 μM) completely inhibited Rsk activity in STHdh cells (STHdh Q7/Q7 cells: 100 ± 14%; STHdh Q7/Q7 + BI: 7 ± 3%; STHdh Q111/Q111 cells: 286 ± 22%; STHdh Q111/Q111 + BI: 13 ± 5%). These results show that increased Rsk activity in STHdh Q111/ Q111 cells results in augmented phosphorylation of both cytosolic and nuclear targets.

Increased Rsk activity contributes to prevent mhttinduced cell death
To evaluate whether elevated Rsk activity could exert a protective effect against mhtt-induced cell death, we study its protective capacity against mhtt-induced toxicity using pharmacological inhibition, knock-down and overexpression approaches. The expression of endogenous mhtt in immortalized STHdh Q111/Q111 cells does not Figure 4 Increased Rsk activity in STHdh Q111/Q111 cells is mainly regulated by PDK1. (A) Lysates from STHdh Q7/Q7 and STHdh Q111/Q111 cells were subjected to immunoprecipitation with anti-Rsk1 and anti-Rsk2 antibodies and the resulting immunocomplexes were used to determine Rsk activity as indicated in material and methods. Kinase activity was expressed as percentage of Rsk activity in STHdh Q7/Q7 cells and data are the mean ± SEM of three independent experiments. Results were analyzed by Student's t-test. **p < 0.01 as compared with STHdh Q7/Q7 cells. (B) STHdh Q7/Q7 and STHdh Q111/Q111 cells were transfected with the following constructs: HA-RskT574A (to study the role of ERK1/2 on Rsk activity) and HA-RskS380E (to analyze the role of PDK1 on Rsk activity). Twenty-four hours after transfection, both STHdh Q7/Q7 and STHdh Q111/Q111 cells were subjected to Rsk activity assay. Results are the mean ± SEM of three independent experiments and are expressed as percentage of control cells (cells transfected with HA alone). Data were analyzed by two-way ANOVA followed by Bonferroni's post hoc test. **p < 0.01 and ***p < 0.001 as compared with STHdh Q7/Q7 control cells, + p < 0.05 and +++ p < 0.001 as compared with STHdh Q111/Q111 control cells and # p < 0.05 as compared with STHdh Q111/Q111 cells transfected with HA-Rsk T574A. produce cell death. Thus, to induce mhtt toxicity, STHdh Q7/Q7 cells were transfected with wild-type (FL-17Q htt) or mutant (FL-75Q htt) htt and cell death was assessed by Hoechst 33258 staining 72 hours after transfection. Overexpression of FL-75Q htt induced 16 ± 3% apoptotic cell death versus 6 ± 2% apoptotic cell death observed in FL-17Q htt-transfected cells ( Figure 7A). In parallel experiments we treated transfected cells with the Rsk inhibitor BI-D1870 (0.1 μM). The inhibition of Rsk exacerbated the toxic effect of FL-75Q htt expression and increased apoptotic cell death to 28 ± 4% (Figure 7A). In contrast, addition of BI-D1870 to FL-17Q htt-transfected cells did not alter cell death ( Figure 7A). Our next goal was to know whether the protective role of Rsk was mediated by Rsk1, Rsk2, or by both isoforms. To address this issue, STHdh Q7/Q7 cells were co-transfected with FL-75Q htt and with siRNAs against Rsk1 (siRsk1), Rsk2 (siRsk2) or both (siRsk1 + siRsk2). First, we checked that transfection with siRsk1 or siRsk2 separately decreased the protein levels of each isoform, and that the co-transfection with siRsk1 and siRsk2 reduced the expression of both isoforms ( Figure 7B). The analysis of cell death showed that inhibition of Rsk1 or Rsk2 separately was not enough to increase the toxic effect of mhtt ( Figure 7C). In contrast, the knock-down of both isoforms enhanced FL-75Q htt-mediated cell death (Figure 7C), similar to that observed after treatment with BI-D1870 ( Figure 7A). To confirm the beneficial effect of Rsk in cells expressing mhtt, we overexpressed Rsk in cells transfected with FL-75Q htt by the co-transfection with HA-Rsk. The analysis of cell death 72 hours later revealed that Rsk overexpression reduced two-fold the cell death induced by mhtt ( Figure 7D). Thus, we conclude that Rsk activity exerts a protective effect against mhtt-induced toxicity, and that both Rsk1 and Rsk2 isoforms are involved in this protective effect.

Discussion
In this work, we provide for the first time evidence for a role of Rsk in HD. Within Rsk family, we analyzed Rsk1 and Rsk2, two isoforms that are broadly expressed in the brain, including the striatum, and whose expression levels are higher respect to other Rsk isoforms [1]. We observed increased protein levels of both Rsk1 and Rsk2 in the striatum of Hdh Q111/Q111 and R6/1 mice, and in STHdh Q111/Q111 cells, which are not dependent on mhtt protein levels as these HD models express different levels of mhtt. Thus, increased Rsk1 and Rsk2 protein levels is the result of the presence of mhtt, and we confirmed this hypothesis by showing that transfection of FL-75Q htt in STHdh Q7/Q7 cells, M213 cells or striatal primary cultures elevated the levels of both Rsk isoforms. In addition, increased Rsk1-2 protein levels correlated with higher basal Rsk activity in STHdh Q111/Q111 cells than in cells expressing wild-type htt. Interestingly, the inhibition of Rsk activity enhanced striatal cell death induced by transfection of mhtt (FL-75Q htt) in STHdh Q7/Q7 cells. Moreover, we show that the overexpression of Rsk reduces considerably cell death in STHdh Q7/Q7 cells transfected with FL-75Q htt. Altogether, these results indicate that elevated Rsk1-2 activity is an efficient mechanism to protect cells against mhtt toxicity. Thus, increased activity of Rsk1-2 could be a compensatory mechanism occurring in HD striatum.
Compensatory responses activated at early phases of HD are considered interesting targets to design neuroprotective therapies to inhibit the progression of neurodegeneration. Here, we analyzed Hdh Q111/Q111 mice at 6 and 10 months of age, and R6/1 at 8 and 12 weeks of age, when they do not show motor dysfunction, but express cellular and molecular markers of HD pathology [22,[24][25][26][27][28]. In addition, we studied the activity and the neuroprotective role of Rsk in STHdh cells, which derive from Hdh Q111/Q111 embryos [29], and reproduce early events in the disease cascade [17]. We observed that increased Rsk activity is neuroprotective against mhttinduced cell death since inhibition of its activity in FL-75Q htt-transfected STHdh Q7/Q7 cells increased cell death. Moreover, we determined that this neuroprotective effect is due to an increase of Rsk activity in striatal cells expressing mhtt, since the inhibition of Rsk did not affect the viability of FL-17Q htt-transfected STHdh Q7/ Q7 cells. Confirming the protective role of Rsk against mhtt, we overexpressed Rsk in STHdh Q7/Q7 cells and showed a protective effect against mhtt-mediated cell death. In addition, we show that both Rsk1 and Rsk2 are necessary to exert this protective role. Different studies in HD models have shown the regulation of other kinases as compensatory mechanisms activated in response of mhtt toxicity. These include Akt [17,30,31] and proteins closely related to Akt, such as the serumand glucocorticoid-induced kinase [32]. Here, we show that in addition to these kinases, Rsk activity is also upregulated in the presence of mhtt, and more importantly, that Rsk activity is neuroprotective against mhttinduced cell death.
Present results also show that the increased activity of Rsk in striatal neurons expressing mhtt is mainly due to the action of PDK1, a kinase whose activity is independent of extracellular factors [3]. We detected increased phosphorylation of Rsk at Ser-221 (dependent on PDK1) and reduced phosphorylation of Rsk at Ser-380 (indirectly dependent on ERK1/2) in Hdh Q111/Q111 mice striatum and STHdh Q111/Q111 cells. In accordance with our results, STHdh Q111/Q111 cells display elevated levels of phospho-PDK1 [17] and reduced levels of phospho-ERK1/2 [18] respect to STHdh Q7/Q7 cells. In addition, using mutant forms of Rsk, we show that the activity of Rsk in STHdh Q111/Q111 cells was considerably inhibited when PDK1-regulated, but not when ERK1/2-regulated residues, were mutated. Consistent with our observations, it has been suggested that PDK1 has the capacity to activate Rsk in an ERK1/2-independent manner [33]. Furthermore, and supporting the important role of PDK1-mediated phosphorylation to Rsk activity, PDK1 deficiency results in Rsk inactivation [34]. Although the neuroprotective role of Rsk has been classically associated with extracellular stimuli induced by trophic factors such as brain-derived neurotrophic factor [13] or epidermal growth factor [12] through the activation of ERK1/2, Rsk phosphorylation by PDK1 increases its activity to a higher extend than ERK1/2-dependent phosphorylation [3]. In HD, neurotrophic deprivation has been proposed as one of the mechanisms involved in the preferential loss of striatal neurons [28,35,36]. Thus, in the absence of trophic support, activation of Rsk through the basal activity of PDK1 could be a crucial mechanism to prevent cell death in HD.
The neuroprotective activity of Rsk is basically due to the wide range of proteins that it regulates. In the nucleus, Rsk phosphorylates and activates several transcription factors, some of them implicated in neuronal survival such as SRF [9], CREB [37] or NFB [10,11]. Studies in non-neural cell lines showed that in the cytosol Rsk phosphorylates and inactivates pro-apoptotic proteins such as Bad [5], GSK-3β [6] or DAPK [7]. In this way, our data indicate that Rsk1-2 activity is elevated in the cytosol and in the nucleus of STHdh Q111/ Q111 cells. In the cytosol, increased Rsk activity correlated with an enhancement of phosphorylated Bad, whereas in the nucleus we observed increased levels of phospho-SRF. Changes in phospho-Bad and phospho-SRF were due, at least in part, to Rsk activation, since inhibition of Rsk significantly reduced the phosphorylation levels of both proteins in STHdh Q111/Q111 cells. In STHdh Q7/Q7 cells, the inhibition of Rsk produced a slight effect on SRF phosphorylation levels, and we did not detect an effect on Bad phosphorylation. Probably, this lack of effect on Bad phosphorylation is due the predominant nuclear activity of Rsk in unstimulated cells [1,2]. Overall, we propose that the neuroprotective effect of Rsk observed in the models studied here could be mediated by the inactivation of pro-apoptotic factors in addition to the activation of transcription factors that regulate the expression of anti-apoptotic proteins.

Conclusions
In conclusion, here we provide evidences that the increase of Rsk1-2 levels is an early event taking place in striatal cells expressing full-length mhtt. Increased Rsk1-2 levels contribute to enhance Rsk activity. Interestingly, our results strongly support that increased Rsk activity in the presence of mhtt is mainly regulated by the basal activity of PDK1 and not by ERK1/2. Moreover, we show that the increase of Rsk1-2 activity observed in cells expressing mhtt could contribute to prevent mhtt-induced cell death. This is the first work showing a role for Rsk in HD, and we propose that therapies targeted to maintain Rsk activity would be a good approach for neuroprotection in HD.

HD mouse models
Homozygous mutant HdhQ 111/Q111 and wild-type Hdh Q7/Q7 knock-in mice were obtained from mating between male and female Hdh Q111/Q7 heterozygotes as described previously [23]. We also used R6/1 mice (B6CBA background) expressing the exon-1 of mhtt with 145 CAG repeats [38]. Mouse genotype was determined as described elsewhere [22]. CAG repeat length was determined by PCR amplification of the repeat using HD1 and HD2 fluorescently labeled primers as previously describe by the Huntington's Disease Collaborative Research Group [15], and subsequent size determination in an ABI 3100 analyzer. These results were double checked by Laragen, Inc. (Los Angeles, CA). All mice used in the present study were housed together in numerical birth order in groups of mixed genotypes, and data were recorded for analysis by microchip mouse number. Experiments were conducted in a blind-coded manner respect to genotype. Mice were genotyped by polymerase chain reaction as described previously [23]. The animals were housed with access to food and water ad libitum in a colony room kept at 19-22°C and 40-60% humidity, under a 12:12 hours light/ dark cycle. All procedures were performed in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and approved by the Local Animal Care Committee of Universitat de Barcelona (99/01), and Generalitat de Catalunya (99/ 1094), in accordance with the Directive 86/609/EU of the European Commission.

Protein extraction and subcellular fractionation
STHdh Q7/Q7 and STHdh Q111/Q111 cells, with or without BI-D1870 treatment or Rsk knock-down, were washed once with phosphate-buffered saline (PBS), and total cellular proteins were extracted by incubating cells in lysis buffer containing 1% Triton X-100, 50 mM Tris-HCl (pH 7.5), 10 mM EGTA, 150 mM NaCl, protease inhibitors (2 mM phenylmethylsulfonyl fluoride (PMSF), 10 μg/μL aprotinin, 1 μg/μL leupeptin) and phosphatase inhibitor sodium orthovanadate (2 mM). Hdh Q7/Q7 and Hdh Q111/Q111 mice were deeply anesthetized and killed by decapitation at the age of 6 and 10 months and wt and R6/1 mice at 8 and 12 weeks of age. The brain was quickly removed and the striatum was dissected out and homogenized in lysis buffer (as above). All samples were centrifuged at 16 100 × g for 20 minutes at 4°C, the supernatants were collected and protein concentration was measured using the Dc protein assay kit (Bio-Rad Laboratories, Hercules, CA).
For quantification of Rsk in htt-transfected cells, immunocytochemistry was performed 24 (striatal primary cultures) or 48 (striatal knock-in and M213 cells) hours after transfection. Quantification of Rsk1 and Rsk2 was performed by confocal microscopy (Leica, Mannheim, Germany) as previously described [42]. Values were expressed as a ratio between the sums of Rsk1 or Rsk2 positive pixels versus cell area. For each condition, 30-40 cells were randomly selected. To study the localization of Rsk1 and Rsk2, STHdh Q7/Q7 and STHdh Q111/Q111 cells were fixed at 80% confluence and processed for immunocytochemistry against Rsk1 or Rsk2. At least 250 cells were evaluated for each condition.

Rsk activity assay
To measure Rsk activity, the assay was performed in STHdh wild-type or mutant cells at 80% confluence or 24 hours after transfection with Rsk constructs, as described previously [23]. Briefly, immunoprecipitation of Rsk was performed by incubation of total protein extracts (100 μg) with anti-Rsk1 and anti-Rsk2 antibodies, 1 μg each. Then, immunoprecipitates were incubated for 15 minutes at 30°C under continuous agitation with the assay mixture buffer containing the substrate peptide and the mixture of ATP and [γ-32 P] ATP (Perki-nElmer, Boston, MA). Reactions were terminated and analyzed as described elsewhere [43]. Incubation with BI-D1870 was used to assess the specificity of Rsk activity assay.

Quantification of apoptosis
STHdh Q7/Q7 cells transfected with FL-17Q htt or FL-75Q htt, with or without Rsk siRNAs, or Rsk DNA plasmid transfection or BI-D1870 treatment were processed for immunocytochemistry against htt as described above. Finally, cells were washed twice in PBS and stained with Hoechst 33258 (1 μg/mL; Molecular Probes, Inc, Eugene, OR) for 5 minutes. After washing twice with PBS the coverslips were mounted with mowiol. Nuclear DNA staining was observed with a fluorescence microscope (Olympus). Transfected cells were detected by the overexpression of htt respect to non-transfected cells. Condensed or fragmented nuclei were counted as apoptotic. At least 100 cells were evaluated for each condition in each independent experiment.

Statistical analysis
Statistical analysis was performed by using the one-or two-way analysis of variance (ANOVA) followed by Bonferroni's post-hoc test, or the unpaired Student's ttest, as appropriate and indicated in the figure legends.