Methods

Yeast strains and culturing: The BY4741 and rnq1Δ deletion strains (MAT a, his3Δ1, leu2-3,112, trp1-1, ura3-1, ade2-1) employed carry integrated constructs encoding a human HTT fragment consisting of the first 17 N-terminal amino acids followed by either 25 or 103 glutamines⁷ . Yeast strains were typically grown on YPD (1 % yeast extract, 2 % peptone, 2 % glucose, 2.5 % agar). Recombinant strains were selected on synthetic medium lacking relevant amino acids (0.68 % yeast nitrogen base without amino acids, 2.5 % agar, 2 % glucose, 0.6 % leucine, 0.3 % lysine, 0.09 % dropout powder lacking the respective amino acids²³ ).

Yeast growth assays: Yeast colonies were inoculated into 100 µl of medium in 96-well plates and incubated at 30 ºC for 16 to 18 hours. Cultures were serially diluted by 5 or 10-fold in distilled water and spotted (5 μl) onto selective plates containing either 2 % galactose or 2 % glucose as a carbon source and incubated for 60 to 72 hours at 30 ºC.

Fluorescence microscopy: Colonies were inoculated in 5 ml of glucose containing selective medium and incubated at 30 ºC with shaking for 24 hours. Cultures were washed once with water and diluted to an OD₆₀₀ of 0.2 in 5 ml selective media containing 2 % raffinose. After a further 6 hours at 30 ºC with shaking, galactose was added to a final concentration of 2 % to induce expression of constructs under the control of GAL1 promoters. After a further 12 hours the number of cells containing aggregates was scored using a Zeiss Axioscope 2 fluorescent microscope with a Zeiss 100 X Plan-NEOFLUAR (1.30/oil, ∞/0.17) objective and a chromomycin filter.

Determination of yeast [PIN] and [PSI] status: The presence of the [PIN⁺] prion was detected by a previously described protocol²⁴ using a primary antibody against Rnq1p (Santa Cruz Biotechnology Inc., 1:10,000 dilution). Samples for the determination of [PSI] status were grown the same way as for the [PIN] status assay, and 4 OD units harvested by centrifugation and suspended in 100 µl of ice-cold lysis buffer [1X PBS without Ca²⁺and Mg²⁺(PAA Laboratories GmbH, Austria), 100 mM NaCl, 2 mM PMSF and 1x EDTA-free protease inhibitors (Roche)]. Samples were lysed with 100 µl of acid washed glass beads (425-600 μm, Sigma-Aldrich, USA) in a bead-beater for 1 minute at maximum speed. After adding an additional 100 µl of ice-cold lysis buffer, samples were left on ice for 1 minute to allow the beads to sediment and the supernatant was removed for further analysis. The amount of protein was quantified using a NanoPhotometer™ Pearl (IMPLEN GmbH, Germany) and 10 μg of protein in 50 μl of lysis buffer were spun at 50,000 rpm at 4 ºC for 15 minutes in a BECKMAN TL-100 ultracentrifuge (Beckman, USA). The supernatant was removed and used as the “Soluble” fraction, while the pellet was resuspended in 50 μl of lysis buffer and used as the “Pellet” fraction. Protein from the total, soluble and pellet fractions were mixed with 2 µl of 5 X protein loading dye [50 mM Tris-HCl (pH 6.8), 2 % SDS, 10 % glycerol, 1 % β-mercaptoethanol, 12.5 mM EDTA, and 0.02 % bromophenol blue] to give a final volume of 10 μl. Samples were denatured at 95 ºC for 5 minutes, separated by SDS-PAGE, and transferred to PVDF membranes. Sup35p was detected with a primary antibody against yeast Sup35p (1:10,000 dilution), a generous gift from Mick Tuite (University of Kent, UK).

Statistical Analyses: Data was analysed by unpaired, two-tailed Mann-Whitney tests using Prism 6 (GraphPad Software).

Results

HD model yeast exhibit HTT aggregation independent of toxicity in a [PIN⁺] and [PSI⁺] background: To avoid confounding issues due to variable plasmid copy number we generated yeast strains containing either HTT25Q or HTT103Q constructs integrated at the HIS3 locus in parental BY4741 yeast, as well as in a rnq1Δ strain in the same background. The Rnq1p-deficient strain was generated by deleting RNQ1 using homologous recombination, which was verified by PCR genotyping and immunoblotting (Figure 1b; data not shown). These constructs encode the first 17 amino acids of HTT followed by the respective polyQ length under the control of an inducible GAL1 promoter, and have an N-terminal FLAG epitope tag and a C-terminal CFP fusion. Integration of the HTT constructs was achieved by employing the integrative vector pRS303, and successful integration events were confirmed by PCR and sequencing of the HIS3 locus (data not shown).

The effect of expressing the integrated HTT constructs was assessed via growth assays and fluorescence microscopy. Surprisingly, we observed that expression of HTT103Q in the integrated strain did not impair growth in multiple integrants generated from independent transformations (Figure 1a). As the presence of the [PIN⁺] and [PSI⁺] prions modulates mutant HTT aggregation and toxicity in baker’s yeast²⁵ , we assessed their status in the integrated strains. We observed that Rnq1p is predominantly found as [PIN⁺] in the BY4741 integrated strains, and as expected is not present in the rnq1Δ strains (Figure 1b). Sup35p (Figure 1c) was found as [PSI⁺] in both the BY4741 and rnq1Δ integrated strains. These data confirm that lack of [PIN⁺] or [PSI⁺] is not responsible for the absence of HTT103Q toxicity in the generated strains.

Despite the lack of HTT103Q-dependent toxicity in the newly created integrated strains (Figure 1a), we observed a high level of inclusion body formation in these cells (Figure 2). Indeed, ~90 % of the cells in the integrated HTT103Q strain contained mutant HTT inclusion bodies 12 hours post induction (Figure 2), while in the rnq1Δ strain the number of cells containing inclusions is reduced to ~7 %. Reintroduction of a plasmid expressing Rnq1p into the rnq1Δ strain restored inclusion body formation to control levels. Thus, this novel integrated yeast model of HD dissociates the toxicity caused by HTT103Q expression from its aggregation and allowed us to further dissect the relationship between prion-like yeast proteins and the dynamics of mutant HTT aggregation without the confounding effects of toxicity.

Interestingly, we found that deletion of the endogenous RNQ1 gene leads to the formation of inclusion bodies in a small number of cells expressing the HTT25Q construct, which does not normally form these structures (Figure 2; P=0.0002). It has recently been observed that a network of proteins in yeast modify the formation of mutant HTT aggregates²⁶ , and it is thus possible that deletion of the endogenous RNQ1 disrupts the balance between other yeast genes that govern HTT aggregation, leading to the aggregation of the HTT25Q construct, which is highly enriched for glutamines in comparison to most endogenous yeast proteins.

Ybr016wp and Sup35p modulate HTT aggregation independently of Rnq1p: We next focused on characterizing the role of Ybr016wp and Sup35p in the aggregation dynamics of mutant HTT in yeast. We initially examined the ability of Ybr016wp to modulate the number of cells containing inclusion bodies when overexpressed in the presence and absence of Rnq1p (Figure 3). Interestingly, overexpression of Ybr016wp causes a small but significant increase in the formation of inclusion bodies in HTT25Q-expressing BY4741 (~ 12 %; P=0.0022) and rnq1Δ (~ 7 %; P=0.0007) cells (Figure 3). A possible explanation for this is the high level of Ybr016wp overexpression, which contains five consecutive perfect repeats (GYNQQ) that are enriched for asparagine, glutamine, tyrosine and glycine. Overexpression of such sequences induces aggregation of other proteins, and could seed formation of inclusion bodies by polyQ-rich sequences^22,27 . Analysis of the HTT103Q integrated strain revealed that Ybr016wp overexpression yileds a small but significant increase in the number of BY4741 cells exhibiting inclusion bodies (P=0.008). Ybr016wp overexpression in the rnq1Δ background also significantly increased the number of rnq1Δ cells forming HTT103Q inclusions, from ~ 7 % to ~ 16 % (P=0.0293). These data indicate that Ybr016wp is able to partially compensate for the loss of Rnq1p in the rnq1Δ background with respect to mutant HTT aggregation, and that this aggregation is not polyQ-length dependent.

As the presence of Rnq1p is crucial for the de novo formation of [PSI⁺] as well as the formation of mutant HTT aggregates in yeast, we next investigated the connection between Sup35p expression and mutant HTT aggregation by overexpressing Sup35p. As with Ybr016wp, overexpression of Sup35p leads to a small but significant increase in the formation of HTT25Q inclusion bodies in the BY4741 background (Figure 4; P < 0.0001). It is possible that these inclusion bodies form due to the presence of [PSI⁺] in the cells, as formation of [PSI⁺] can be triggered by overexpression of Sup35p²⁸ , which is able to cause aggregation of heterogeneous polyQ-rich sequences²⁹ . Interestingly, in the rnq1Δ strain overexpressing Sup35p the formation of HTT25Q inclusion bodies is dramatically increased (~ 68 %) compared with the parental BY4741 strain (~ 12 %; P=0.0264). As mentioned earlier a network of proteins is able to modify aggregation of mutant HTT in yeast²⁵ and has also been found to trigger the formation of [PSI⁺]. It is therefore possible that this network is activated by the absence of endogenous Rnq1p in the rnq1Δ strain, and is further stimulated by the presence of [PSI⁺]. Interestingly, the HTT25Q and HTT103Q strains exhibit a similar number of cells with aggregates in the rnq1Δ background with Sup35p expression, indicating that this substitution is not polyQ-length dependent, and potentiates the aggregation of shorter polyQ stretches.

Discussion

Our studies indicate that overexpression of Rnq1p, Sup35p and Ybr016wp can modulate the formation of mutant HTT aggregates in yeast. Furthermore, the induction of mutant HTT inclusion body formation by both Sup35p and Ybr016wp is in part dependent upon the presence of Rnq1p in the cell. The mechanism(s) by which these two aggregation-prone proteins modulate mutant HTT aggregation is not clear, nor is it known if they are acting in a similar manner. In the case of Sup35p it has been proposed that there are two groups of factors that govern both the formation of mutant HTT aggregates and the Sup35p to [PSI⁺] switch in yeast²⁶ . The first group is involved in the formation of large mutant HTT aggregates and the switch between Sup35p and [PSI⁺], while the factors in the second group are crucial for the formation of soluble oligomeric species of mutant HTT, as well as the formation of diffuse [PSI⁺] aggregates. These findings further support the hypothesis that yeast prions and glutamine-rich amyloidogenic proteins interact with similar molecular partners upon forming larger amyloid structures in yeast²² . Based upon this observation it is possible that mutant HTT interacts directly with Sup35p. Indeed, a recent investigation indicates that mutant HTT can sequester Sup35p in [PSI⁺] cells²⁷ . As most established yeast prions have a Q/N-rich sequence and mutant HTT can sequester Q-rich sequences in yeast²⁸ , it is possible that mutant HTT can sequester other yeast prions. Ybr016wp is poorly characterised, and the relationship between this protein and mutant HTT described here is novel. Although Ybr016wp has an unknown cellular function it contains a CYSTM domain that is characteristic of eukaryotic proteins involved in stress tolerance³⁰ and is anchored to the plasma membrane and predicted to be palmitoylated, similar to the human prion protein PrP^{C 31}.

Many highly conserved yeast proteins have prion-like features²² , suggesting that human proteins with prion-like domains may behave in a similar manner. Indeed, a recent study has implicated a large number of glutamine and asparagine rich human RNA-binding proteins in the pathology of a range of neurodegenerative diseases³². Thus, it is possible that a suite of aggregation prone proteins that modify the aggregation of mutant HTT exists not only in yeast but in humans as well, which may have implications for HD pathogenesis. Indeed, such proteins could modify the aggregation and misfolding of not only mutant HTT, but other amyloidogenic proteins as well, thus informing the development of therapies for HD as well as for other neurodegenerative diseases caused by protein misfolding.

Material and methods

Culture Media. Dulbecco’s modified Eagle’s medium (DMEM) with high glucose (4.5 g/liter) was from Sigma (Madrid, Spain), Ham’s F12 nutrient mixture, Eagle’s minimal essential medium (EMEM) with Earl’s salts, Leibovitz’s L-15 medium; and supplements as L-glutamine, fetal bovine serum (FBS), sodium pyruvate, geneticin and L-glutamine, purchased from Invitrogen (Madrid, Spain). Glucose 45 %, trypsin-EDTA, insulin, putrescine, progesterone, and sodium selenite were from Sigma (Madrid, Spain), and human transferrin, 30 % iron saturated, was from Boehringer- Mannheim (Barcelona, Spain).

Neuronal and glial cultures. Conditionally immortalized, STHdh^Q7/Q7 (Q7), and Q111 striatal neuronal progenitor cell lines expressing endogenous levels of huntingtin with 7 and 111 glutamines, respectively, were kindly provided by Prof. Alberch (Universitat de Barcelona, Spain) and Prof. Marcy E. Macdonald (Massachusetts General Hospital, Boston, USA) and were described previously by Trettel²⁵ . Striatal cells were grown at 33 °C in Dulbecco’s Modified Eagle’s Medium (DMEM), supplemented with 10 % fetal bovine serum (FBS), 1 % streptomycin-penicillin, 2 mM L-glutamine, 1 mM sodium pyruvate, and 400 μg/ml geneticin (G418). Twenty-four hours after plating, the cells were changed to serum-free defined medium (EF12) as previously reported ³¹³². EF12 consisted of a 1:1 (v/v) EMEM and nutrient mixture of Ham’s F-12, supplemented with D-glucose (6 mg/ml), insulin (25 µg/ml), transferrin (100 µg/ml), putrescine (60 µM), progesterone (20 nM), and sodium selenite (30 nM).

Glial striatal cultures were obtained from 129SV/C57BL/6 wild-type (WT) mice as previously described ¹². Glia-enriched striatal cultures were obtained from sibling cells kept in culture from 10–15 days to 3 months in DMEM-FBS. Positive staining with anti-GFAP antibody identified the astrocytes in these cultures. After 10–15 days in culture, the number of astrocytes was about 80–90 % of total cells ²². The cultures were fed once per week. To obtain the glia conditioned medium (GCM), DMEM-FBS medium was discarded, and the cells were washed out three times with Leibovitz’s L-15 and subsequently cultured in serum free defined medium. After 24 hours of culture under such conditions, the medium was collected and stored frozen. This medium was considered GCM.

Antibodies. The following antibodies were used: Anti-mouse IgG fluorescein from Jackson (PA, USA) and anti-rabbit IgG Alexa Fluor from Molecular Probes (Eugene, OR, USA). Mouse monoclonal anti-HSP-70 and rabbit polyclonal anti-CHIP antibodies were from Santa Cruz Biotechnology (Heidelberg, Germany). Mouse monoclonal anti-PARP was from Cell Signalling Technology (Denver, USA). Mouse monoclonal anti-ubiquitin, anti-synaptophysin and anti-ß-tubulin class III antibodies were from Chemicon, Millipore (Temecula, CA, USA). Anti-mouse and anti-rabbit horseradish peroxidase secondary antibodies were from Amersham. ß-Actin secondary antibody was an anti-mouse phosphatase alkaline conjugated from Sigma (Madrid, Spain).

Chemicals. 5,5-Dithio-bis-2-nitrobenzoic acid (DTNB), hydrogen peroxide, sodium glutamate, 3-nitropropionic acid, trypan blue reagent and reduced glutathione (GSH) were from Sigma (Madrid, Spain). NADPH and GSH reductase (GR) were from Boehringer-Mannheim (Barcelona, Spain). The cytotoxicity detection kit for lactate dehydrogenase (LDH) and the cell proliferation kit I [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT)] were from Roche Diagnostics. The BCA protein assay kit was from Pierce (Rockford, Ill, USA). All other reagents were of the highest purity commercially available from Merck or Sigma.

Treatment of the cellsExperiments were performed in 12 and 24-well culture plates for western blot and for survival and death assays, respectively. For the immunocytochemistry detection, the cells were cultured in 13 mm-diameter glass cover slips with a density of 5000 cells. After 3 days in vitro (DIV) in the growth medium (DMEM-FBS), Q7 and Q111 cells reached confluence. At this time, the medium was discarded and the cells were subsequently cultured and treated in a chemically serum-free defined medium (EF12) or in striatal glia-conditioned medium (GCM) (embryonic E16, postnatal of 13 days and postnatal of 2 months).

In order to test the protection of fetal GCM against different neurotoxins, we added the following chemicals to Q7 and Q111 cells cultured in defined medium, supplemented or not with fetal GCM: sodium glutamate (5 mM), H₂O₂ (2 µM), 3NP acid (2.5 mM) for 24 hours. In order to study the protection of postnatal GCM against H₂O₂ induced cell death in Q111 cells we used postnatal striatal GCM, from 13-day and 2-month old mice and the experiments were performed as described in the previous paragraph.

To investigate the mechanisms of neuroprotection by fetal GCM, we studied the ubiquitination of proteins, and the expression of death and chaperone proteins by Western blot assays. To compare the effects of GCM with the well-known neuronal growth factors, we treated Q111 cells with fetal GCM versus different growth factors such as GDNF (50 and 100 ng/ml), bFGF (10 and 20 ng/ml) and BDNF (50 and 100 ng/ml). We initially studied the effects of each growth factor on Q111 cells and compared them with the fetal GCM. Then, we studied the protection of the growth factors and fetal GCM on H₂O₂ induced cell death in Q111 cells.

Cell survival, proliferation assay and cell differentiation. Necrotic cell death was measured according to LDH activity in the culture medium and by trypan blue dye exclusion in cells. LDH activity was measured by using a cytotoxicity detection kit ¹²³³ . Apoptotic cell death was measured by nuclei stained with bis-benzimide (Hoechst 33342). Mitochondrial activity was analyzed with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay in Q7 and Q111 cells. The MTT assay determines the ability of cells to metabolize MTT. At the end of the cell treatment period, 300 µl of culture medium were removed from total 500 µl of each well, and 20 µl of MTT solution (5 mg/ml) added and incubated for 2 h. At this time, 200 µl of solubilization solution (10 % SDS in 0.01 M HCl) were then added to the wells, and, after 24 h of incubation at 37 °C, 100 µl transferred into a 96-well microtiter plate. The absorption value at 540 nm was measured in an automatic microtiter reader (Spectra Fluor; Tecan, Männedorf, Switzerland). To test the differentiation of the cells, we counted the number of multiprocess cells as the number of cells with 3 or more prolongations.

Immunocytochemistry. After the experimental treatments, the cells were fixed with 4 % paraformaldehyde. Then, cells were postfixed and permeabilized in ethanol-acetic acid (19:1) for 10 min at -20 ºC and incubated in a blocking solution followed by overnight incubation at 4 ºC with the cytoplasmatic class III ß-tubulin primary antibody (1:300) and with a protein of the presynaptic vesicle exocytosis, synaptophysin primary antibody (1:200). Fluorescein and Alexa Fluor-conjugated secondary antibodies were used to visualize positive cells under fluorescent microscopy. Nuclei were stained by bis-benzimide (Hoechst 33342) and immunostaining was visualized under fluorescent microscopy. The number of immunoreactive cells was counted in predefined parallel strips.

Measurement of GSH and ROS levels.

Total glutathione levels were measured by the method of Tietze ³⁴. Intracellular levels of ROS were detected using the non-fluorescence DCF dye, which is oxidized by intracellular ROS to form the highly fluorescence DCF. The cells were incubated with 5 µM DCFH-DA 30 min before H₂O₂ treatments. After washing the excess of DCF with a medium free of phenol red (PBS with 1mM glucose), the fluorescence was measured in a laser-scanning confocal microscope (Nikon eclipse Ti) or in an automatic microtiter reader (Spectra Fluor, Tecan) at excitation and emission of 485 nm and 582 nm, respectively. Fluorescence was quantified by automated image analysis with Image Pro software. For each section mean fluorescence was calculated from 7 separate single fields.

Western blot analysis and detection of ubiquitinated proteins.

Q111 extracts for western blot analysis were prepared in ice-cold extraction buffer consisting of 20 mM Tris–HCl (pH 7.4), 10 mM potassium acetate (AcK), 1 mM dithiothreitol (DTT), 0.25 % NP-40, 1 mM EDTA, 2 mM EGTA, 1 mM PMSF, protease inhibitors cocktail (SIGMA) and a cocktail of phosphatase inhibitors (100 mM sodium fluoride, 20 mM sodium molybdate and 20 mM ß-glicerophosphate). The samples were homogenized, centrifuged at 4 ºC and protein content was determined by the BCA protein assay kit. Total protein (30 µg) was electrophoresed in 10 % SDS-PAGE gels and transferred to 0.45 mm nitrocellulose membranes (Amersham), as described previously¹²³³. After blocking, blots were incubated overnight at 4 ºC in 5 % non- fat dried milk with primary antibodies: the chaperone anti-HSP-70 (1/700); anti-CHIP (1/1000) and anti-PARP (1/1000).

To determine changes in ubiquitination, cell cultures treated or untreated with H₂O₂ (2 µM) in serum-free defined medium and in GCM for 24 h, were scraped in 150 µl of lysis buffer [50mM Tris–HCl, 150 mM NaCl, 20 mM EDTA, 1 % TritonX-100, 50 mM sodium fluoride, 20 mM N-ethyl-maleimide, 100 µM sodium ortovanadate, 1 mM PMSF and protease inhibitors cocktail (SIGMA)] and boiled for 5 min. The lysates were centrifuged at 12 000 x g at 4 ºC for 30 min and 7.5 µg of protein were conducted to immunoblot assay with a rabbit polyclonal antibody to ubiquitin.

The secondary antibodies (1/2000) followed by ECL detection reagents (Amersham) were used for immunodetection. Immunoblot of ß-actin diluted (1/20000) was performed to demonstrate equal protein loading. The blots were quantified by computer-assisted video densitometry.

Characterization of defined and glia-conditioned medium protein profile. Defined E12 and embryonic E16 glia-conditioned media were lyophilized and concentrated in 5 % of initial volume. Protein concentration was determined by BCA method and 5 µg of protein from the supernatants of both media, were used for the electrophoresis in 4-20 % polyacrylamide gel. Electrophoresis was carried out at 150V. After completion of electrophoresis, the gel was developed using silver staining to observe protein profile.

Statistical analysis. The statistical analysis was performed by one-way analysis of variance (ANOVA), followed by Newman–Keuls multiple comparison test. Differences were considered statistically significant when p< 0.05. Analysis of data was performed using the Graph Pad Software (San Diego, CA) Prism 4 software.

Results

Fetal striatal GCM prevents cell death produced by sodium glutamate in Q7 and more so in Q111 cells. Sodium glutamate, 5 mM x 24h, reduced cell proliferation and mitochondrial activity and increased cell death in Q7 and Q111 cells in serum-free defined medium (Fig 1A-C). After treatment with GCM the number of high affinity GABA uptake sites, a marker of striatal neurons, was reduced to around 75 % of controls (Fig 1D). Co-treatment, with fetal GCM completely reverted these effects of glutamate in Q111 cells (Fig 1A-D).

The fetal striatal GCM prevents cell death produced by 3-nitropropionic acid in Q111 cells. The mitochondrial complex II inhibitor (3NP, 2.5 mM x 24h) induced cell death in Q111 cells, as shown by reducing the neuronal area of the culture (Fig 2A and B) and increasing the number of apoptotic cells of around 50 % more than in the untreated cultures (Fig 2C and D). Co-treatment with fetal GCM increased the neuronal area of the cultures, expressed by the area of ß-tubulin (Fig 2A and B), and reduced the levels of apoptosis not only to levels of the cultures untreated with 3NP, but to levels below 50 % of those of the controls (Fig 2C and D).

Fetal striatal GCM decreases H₂O₂ induced cell death in Q111 cells. The treatment with H₂O₂, 2µM for 24 hours, decreased the ß-tubulin area of the cultures to around 2/3 of the controls (Fig 3A and B), almost doubled the percentage of apoptotic cells in the cultures (Fig 3C) and more than tripled the number of necrotic cells, as shown by trypan blue staining (Fig 3D and E). Co-treatment with fetal GCM completely prevented the toxic effects of H₂O₂, on ß-tubulin staining (Fig 3A and B) and apoptosis (Fig 3C) and reduced the percentage of necrosis to levels even below those of the controls (Fig 3D and E).

The treatment with H₂O₂, induced greater accumulation of intracellular ROS in Q111 cells than in Q7 cells (data not shown). The accumulation of ROS in Q111 cells was time (Fig 3F and G) and dose dependent (Fig 3H). Co-treatment with fetal GCM completely prevented the elevation of ROS induced by H₂O₂(Fig 3I and J).

Mechanisms implicated in the GCM neuroprotection in Q111 cells

Free radicals. In order to study the mechanisms implicated in the GCM neuroprotection from H₂O₂ induced cell death, we investigated the differential pattern of proteins between defined medium (DM) and fetal GCM and the differences in the free radical scavengers. Gel electrophoresis of GCM and DM revealed the presence of seven bands in the GCM, which were absent or much weaker in DM. These bands have a relative molecular weight of 250, 123, 38, 36, 33, 29 and 12 KDa (Fig 4A). GSH levels were many folds greater in fetal GCM (0.58 ± 0.04) than in DM (0.013± 0.01). This may explain, at least in part, the GCM protective effects (Fig 4B). Treatment with H₂O₂, 2µM, decreased the intracellular levels of GSH and co-treatment with fetal GCM increased these levels even above the levels of the controls (Fig 4C).

Apoptotic proteins. The cleavage of PARP by caspases is recognized as a hallmark of apoptosis ³⁵ . The H₂O₂ activated the accumulation of this cleavage and the fetal GCM prevented this effect. H₂O₂, 2µM x 24h, tripled the ratio PARP cleaved/full, an index of apoptosis, in Q111 cells and co-treatment with GCM greatly diminished this index (Fig 4D).

Chaperones. The chaperone HSP 70 levels were increased by treatment with H₂O₂, 2µM x 24h (Fig 4E) and these effects were prevented by fetal GCM (Fig 4E). The levels of CHIP were not changed by treatment with H₂O₂, 2µM x 24h, GCM or both (data not shown).

Polyubiquitinated proteinsH₂O₂, 2µM x 24 hours, induced the accumulation of polyubiquitinated proteins in Q111 cells (Fig 4F and G). Co-treatment with GCM reverted this accumulation (Fig 4F and G).

Postnatal GCM prevents cell death produced by H₂O₂ in Q111 cells. We have also studied the protective effects of postnatal GCM, of 13 days and 2 months in the huntingtin mutant cells (Fig 5). Postnatal GCM prevents H₂O₂ induced cell death in Q111 cells. GCM prevented the H₂O₂-induced reduction of the ß-tubulin area and the increased apoptosis in Q111 cells (F 5A-H).

Comparative effects of neurotrophic factors and GCM on Q111 cells

The growth factor bFGF has beneficial effects in Q111 cells but fewer than fetal GCM. Basic FGF 10 and 20 ng/ml, did not change the neuronal area of the cultures, as measured by the ß-tubulin staining, a parameter which is increased by GCM (Fig 6 A and B). Both doses of bFGF decreased the percentage of apoptotic cells (Fig 6 C and D) and increased the cell differentiation measured by the number of cells with multiprocesses (Fig 6 E and F), but the effects of bFGF, at the doses used in these experiments, were less potent than those of GCM.

The growth factor GDNF has beneficial effects in Q111 cells but fewer than fetal GCM. GDNF, 50 and 100 ng/ml, did not increase, as GCM does, the ß-tubulin area of the cultures (Fig 7A and B) and had a significant but very mild protective effect on apoptosis, which is strongly suppressed by GCM (Fig 7C and D). GDNF 50 ng/ml failed to increase the percentage of cells with multiple processes, GDNF, 100ng/ml increased it by around 25 %; however, GCM almost doubled it (Fig 7E and F). As a consequence, GCM was a much more potent neurotrophic agent for Q111 cells than GDNF at doses used in this study.

Effects of growth factor BDNF and fetal GCM in Q111 cells. BDNF, 50 and 100 ng/ml, did not increase, as GCM does, the ß-tubulin area of the cultures (Fig 8A and B) and had a significant but smaller protective effect on apoptosis than GCM (Fig 8C and D). BDNF, 50 ng/ml, failed to increase the number of differentiated cells with multiple processes. BDNF, 100 ng/ml, increased this number significantly, with respect to control, but was less potent than GCM (Fig 8E and F). As a consequence, GCM was a much more potent neurotrophic agent for Q111 cells than BDNF at doses used in this study.

Comparative effects of growth factors and fetal striatal GCM on H₂O₂ induced cell death. We compared the effects of bFGF, GDNF, BDNF and fetal GCM on H₂O₂ induced toxicity. BDNF and bFGF failed to revert the H₂O₂ induced reduction of the ß-tubulin area of the cultures while GDNF and GCM did (Fig 9A and B). BDNF, GDNF and GCM prevented the H₂O₂ induced apoptosis, being GCM the most effective (Fig 9C) while bFGF failed to decrease apoptosis. None of the neurotrophic factors, at the doses used in this study, reverted the reduction of differentiated cells induced by H₂O₂, but GCM completely normalized it (Fig 9D and E). As a consequence, GCM was a more potent neurotrophic agent for Q111 cells than any of the neurotrophic factors at the doses used in this study.

Discussion

These studies reveal that fetal and postnatal striatal GCM protects from spontaneous degeneration and cell death induced by exposure to glutamate, 3-NP acid and H₂O₂ in striatal cells with polyglutamines expansions of huntingtin which could be considered as an in vitro models of HD. GCM is rich in small antioxidants, like GSH and ascorbic acid, and peptidic neurotrophic factors, such as GDNF, BDNF, neurotrophin-3 (NT-3), NGF, and bFGF as well as novel DA neurotrophic factors ^{223637383940414243}. The effects of GCM observed in these experiments are, however, more potent than those of any known neuroprotective or neurotrophic factor ever tested in vitro models of HD.

All the mechanisms of neurotoxicity triggered by the neurotoxins used in these experiments, including abnormal excitatory neurotransmission, abnormal mitochondrial function, and excessive oxidative stress have been demonstrated in patients with HD and animal models of this disease ⁴⁴⁴⁵⁴⁶. Glutamate is a major excitatory amino acid neurotransmitter in the central nervous system. A high concentration of glutamate induces neuronal cell damage under in vitro conditions ⁴⁷⁴⁸⁴⁹. Striatal medium spiny neurons, which receive a strong excitatory input in their dendritic spines from cortical neurons, through the cortico striatal pathways, are neurons of projection mostly to the pallidum, which contain GABA and other neurotransmitters, and are particularly vulnerable to HD. Their progressive depletion is in strong correlation with symptom severity ⁵⁰. Our data shows that the GCM reverted cell death produced by glutamate and increased the levels of GABA uptake sites, a marker of GABA neurons, in STHdhQ111/111 cells.

In HD there is a mitochondrial dysfunction that produces a significant decrease in the activity of mitochondrial respiratory chain complexes (MCC) II, III and IV, among other factors⁵¹. 3-Nitropropionic acid is a natural toxin that irreversibly inhibits the succinate dehydrogenase enzyme, which is the main constituent of MCC II. Administered to rodents, 3-NP reproduces the brain lesions observed in HD patients, which consist of degeneration of the striatum with particular impact on medium-sized spiny projection neurons ⁵². In Q111 cells, we have observed that the 3-NP produces a decrease in cell survival and an increase in necrotic and apoptotic cell death. This study shows the protection by GCM of 3-NP induced cell death.

We challenged Q111 cells with H₂O₂ to study the effects of the fetal and postnatal GCM on free radical metabolism and ubiquitinated proteins in Q111 cells. H₂O₂ produces an accumulation of free radicals and polyubiquitinated proteins in Q111 cells and fetal GCM reverts both nearly to control levels. Fetal GCM also prevents the H₂O₂ induced cell death in Q111 cells. The cleavage of PARP by caspases is recognized as a hallmark of apoptosis ³⁵. H₂O₂ activated the accumulation of this cleavage and fetal GCM prevented this effect. The chaperone HSP70 is increased in the presence of H₂O₂ as a marker of cell stress but the fetal GCM reduced it. Fetal GCM increased the levels of the antioxidant GSH in Q111 cells and prevented the H₂O₂ induced depletion of GSH. The 13-day and 2-month postnatal GCM also decreased the H₂O₂ induced cell death in this in vitro model of HD.

Since GCM is rich in neurotrophic factors we investigated whether the effects of GCM could be related to the presence of these factors. We compared the effects of GCM and bFGF, GDNF and BDNF, at different concentrations, on one of the models of toxicity for Q111 cells used in this study; the challenge with H₂O₂. We found that GCM was more potent than any of the above mentioned factors for the rescue of Q111 cells, the reduction of apoptosis and the induction of arborisation, after treatment with H₂O₂. However, glia-expressing mutant huntingtin can also secret toxic molecules ^53545556

We conclude that the GCM is a powerful neuroprotective agent for Q111 striatal cells. GCM should be tested on animal models of HD and, if the results obtained in this study are confirmed, on patients with HD.