A series of N-terminal epitope tagged Hdh knock-in alleles expressing normal and mutant huntingtin: their application to understanding the effect of increasing the length of normal huntingtin’s polyglutamine stretch on CAG140 mouse model pathogenesis

Prior in vitro observations, cell culture studies, and experiments in Drosophila have demonstrated that mutant htt can sequester normal htt in protein aggregates, but recent analyses of human postmortem brain extracts and protein extracts derived from HD mouse models using blue native polyacrylamide gel eletrophoresis (BNP) have shown that full-length normal and mutant htt are predominantly monomeric [33]. To confirm that epitope-tagged normal htt exists predominantly as a monomer in our mouse models, we immunoprecipitated 3xFlag7Q-htt from cerebellar and striatal protein extracts prepared from 1 month-old Hdh ^{3xFlag7Q/HA7Q} mice, and forebrain extracts prepared from 12 month-old Hdh ^{3xFlag7Q/HA7Q} mice (Figure 4). Flag antibody-bound and non-bound fractions were then analyzed by western blotting using FLAG and HA antibodies. In both 1 month and 12 month-old mice, Flag3x7Q-htt was efficiently immunoprecipitated. HA7Q-htt, in contrast, was detected only the antibody non-bound fractions. Thus, in both young and older mice, epitope-tagged htt exists primarily as a monomer in our protein extracts using our co-immunoprecipitation conditions.

We next prepared striatal and cerebellar protein extracts from 6 month-old Hdh ^{140Q/3xFlag7Q} mice to determine if soluble full-length mutant htt can interact with normal htt. Following immunoprecipitation of 3xFlag7Q-htt with anti-FLAG agarose beads, antibody-bound and non-bound fractions were analyzed by western blotting using Flag and expanded polyQ (1C2) antibodies (Figure 5A). While 3xFlag7Q-htt was enriched in the antibody bound fractions, 140Q-htt was detected only in the antibody non-bound fraction. We interpret this result to suggest that little, if any, stable interaction between normal and mutant htt occurs in the 6 month-old Hdh ^{140Q/3xFlag7Q} brain. To determine if an association between normal and mutant htt can be detected in older Hdh ^{140Q/3xFlag7Q} or Hdh ^{3xFlag140Q/HA7Q} mice, co-immunoprecipitation experiments were performed using whole brain extracts prepared from 13 month-old Hdh ^{3xFlag140Q/HA7Q} , and 27 month-old Hdh ^{140Q/3xFlag7Q} mice. Cytoplasmic extract from the Hdh ^{3xFlag140Q/HA7Q} and Hdh ^{140Q/3xFlag7Q} brains was immunoprecipitated with anti-expanded polyQ (1C2) or Flag antibodies, using Protein G-agarose beads, and western blots of antibody-bound and non-bound fractions were probed with 1C2 (recognizing 3xFlag140Q-htt and 140Q-htt), and MAB2166 (recognizing both normal and mutant htt) antibodies (Figure 5B, C). HA7Q-htt did not co-immunoprecipitate with 3xFlag140Q-htt, and similarly 140Q-htt did not co-immunoprecipitate with 3xFlag7Q-htt, suggesting that in older mice, a stable interaction between full-length 7Q-htt and 140Q-htt does not occur. To control for non-specific association of htt with the Protein G agarose-beads, whole extract was incubated with Protein G-agarose beads in the absence of 1C2 or Flag antibodies, and the antibody-bound and non-bound fractions were probed with 1C2 and 2166 antibodies (Figure 5B, C). Htt was not detected in the control bound fractions, indicating that little, if any, non-specific binding of htt to the agarose beads occurred.

To determine if normal htt with a 20Q stretch and human PRR can interact with full-length mutant htt, we performed FLAG immunoprecipitations with Protein-G agarose beads using whole brain extracts prepared from 16 month-old Hdh ^{140Q/3xFlag20Q} mice, and probed the Flag antibody-bound and non-bound fractions with 1C2 and 2166 antibodies (Figure 5C). A very low level of interaction between 140Q- and 3xFlag20Q-htt was detected in the whole brain protein extracts. We estimate that under our co-immunoprecipitation conditions, < 2.5% of soluble full-length mutant htt was associated with 3xFlag20Q-htt (see Methods).

To explore the effect of expressing a version of htt with a normal human polyQ stretch and human PRR on HD mouse model phenotypes, we first compared the number of striatal neuropil and nuclear inclusions in Hdh ^{140Q/3xFlag7Q} and Hdh ^{140Q/3xFlag20Q} mice at 2 months, 4 months, 6 months, and 24 months of age (n=4 of each genotype) (Figure 6). For both genotypes, htt inclusions were first detected using the MW8 aggregation-specific antibody at 6 months of age, and at this age, the number of nuclear inclusions was greater than that of the neuropil aggregates in both genotypes. By 24 months of age, there was a widespread increase in the number of striatal nuclear and neuropil aggregates in both genotypes (Figure 6). Interestingly, the total number of inclusions in the 6 month-old Hdh ^{140Q/3xFlag20Q} striatum was significantly greater than that found in the Hdh ^{140Q/3xFlag7Q} striatum (Figure 6B). This difference was due entirely to an increase in the number of nuclear inclusions in the Hdh ^{140Q/3xFlag20Q} brain, as there was no significant difference in the number of neuropil aggregates between the two genotypes (Figure 6C).

Although the number of nuclear and neuropil aggregates was not significantly different in the 24 month-old Hdh ^{140Q/3xFlag7Q} and Hdh ^{140Q/3xFlag20Q} brains, the size of the inclusions appeared larger in the Hdh ^{140Q/3xFlag20Q} striatum (Figure 6A, 7A). To quantify the size of the aggregates, we determined the average MW8⁺ pixel area per aggregate in confocal images by dividing the total MW8⁺ pixel area by the number of aggregates (Figure 7B). At 6 months of age, no significant difference in the mean size of aggregates was observed in the Hdh ^{140Q/3xFlag7Q} and Hdh ^{140Q/3xFlag20Q} striatum. At 24 months of age, however, we observed a significant increase in the mean aggregate size in the Hdh ^{140Q/3xFlag20Q} striatum in comparison to the Hdh ^{140Q/3xFlag7Q} striatum.

To compare the extent of the co-localization of 3xFlag20Q-htt with 140Q-htt vs. 3xFlag7Q-htt with 140Q-htt in aggregates, we quantified the percentage of htt aggregates that co-stained with both Flag and MW8 antibodies in the Hdh ^{140Q/3xFlag7Q} and Hdh ^{140Q/3xFlag20Q} striatum using confocal microscopy (Figure 8A). At 6 months of age, we observed no significant difference in the percentage of htt aggregates that contained the 3xFLAG epitope in the Hdh ^{140Q/3xFlag7Q} and Hdh ^{140Q/3xFlag20Q} striatum (Figure 8B). At 24 months of age, there was a trend towards increased co-localization of epitope-tagged normal htt with mutant htt in the Hdh ^{140Q/3xFlag20Q} brain in comparison to the Hdh ^{140Q/3xFlag7Q} brain (P=0.06) (Figure 8B). This result suggests that 3xFlag20Q-htt may be recruited into inclusions more efficiently than 3xFlag7Q-htt, and this could contribute to the increased average size of aggregates we observed in the Hdh ^{140Q/3xFlag20Q} striatum.

To confirm that normal htt was sequestered in mutant htt inclusions, we first performed cellulose acetate filter trap assays [34] with the initial 800xg pellet fraction from 24 month-old Hdh ^{140Q/3xFlag7Q} whole brain extracts (containing crude nuclei and insoluble htt aggregates) (Figure 9A). 1C2, mEM48, and Flag-positive htt inclusions trapped on the cellulose acetate were detected by western blotting, while an underlying PVDF membrane was positive for SDS-soluble 140Q- and 3xFlag7Q-htt species, but negative for mEM48-positive aggregates.

In both full-length and truncated mutant htt HD mouse models, the predominant species of mutant htt found in macroscopic inclusions consists of N-terminal fragments [35‐37]. To determine if epitope-tagged N-terminal fragments of normal htt were present in the inclusions, we extracted the SDS-insoluble 800xg pellet fraction obtained from 24 month old Hdh ^{140Q/3xFlag7Q} , Hdh ^{140Q/3xFlag20Q} , and Hdh ^{3xFlag140Q/HA7Q} brains with formic acid to solubilize the htt aggregates [38‐40], and performed western blotting using Flag and 1C2 antibodies to visualize full-length and N-terminal fragments of normal and mutant htt (Figure 9B). In the formic acid-solubilized SDS-insoluble nuclear fractions, we could detect 140Q- and 3xFlag140Q-htt fragments with the 1C2 antibody that ranged in size from ~100 kD to ~250 kD, and apparently full-length mutant htt. The Flag antibody detected an increased level of 3xFlag20Q-htt fragments in comparison to 3xFlag7Q-htt fragments. We interpret these data to suggest that sequestration of normal htt N-terminal fragments into inclusions does occur in heterozygous knock-in HD mouse models, and is enhanced by expression of a humanized version of the normal Hdh allele.

The large neuropil and nuclear inclusions that we visualized by standard immunohistochemical methods using the MW8 and mEM48 (MAB5374) antibodies may not represent all the sites where aggregates are forming. To visualize these sites, treatment of tissue sections with formic acid to expose the expanded polyQ epitope within the htt aggregates prior to antibody staining is required [41]. Using this method, small neuropil aggregates and aggregation foci located in striatal neuronal processes resembling chains of immuno-positive punctae were observed [41]. Such 1C2⁺ punctae were detected in both Hdh ^{140Q/3xFlag7Q} and Hdh ^{140Q/3xFlag20Q} striata (Figure 10A). For quantification, we measured the total pixel area of the 1C2⁺ punctae, as they were difficult to resolve clearly. We observed a significant increase in 1C2 staining following formic acid treatment in the Hdh ^{140Q/3xFlag20Q} striatum in comparison to the Hdh ^{140Q/3xFlag7Q} striatum at both 6 months and 24 months of age (Figure 10B). Unfortunately, we could not detect co-localization of the 3xFlag epitope with the expanded polyQ epitope, as we were unable to recover FLAG immunoreactivity following formic acid treatment (data not shown). Nevertheless, our observations suggest that increasing the length of the normal htt polyQ stretch and/or replacing the mouse PRR with the human PRR can enhance the formation of small cytoplasmic htt aggregates.

In HD mouse models and in the HD brain, reactive astrocytosis occurs as a consequence of mutant htt expression [42‐44]. To determine if astrocytosis in Hdh ^{140Q/3xFlag20Q} mice is altered in comparison to that observed in Hdh ^{140Q/3xFlag7Q} mice, we examined glial fibrillary acidic protein (GFAP) expression by immunohistochemistry in the striatum of 24 month-old wild-type, Hdh ^{140Q/3xFlag7Q} , Hdh ^{140Q/3xFlag20Q} , and Hdh ^{140Q/3xFlag140Q} mice (Figure 11A). As expected, GFAP immunostaining was relatively low in the wild type striatum (measured as total pixel area of GFAP signal/field), likely reflecting the basal expression of GFAP in resident astrocytes. In all three HD mouse models, however, GFAP immunostaining was significantly higher in the mutant striatum in comparison to the wild type striatum (Figure 11B). A significant increase in astrocytosis was observed in the Hdh ^{140Q/3xFlag7Q} striatum compared to wild-type striatum, and in the Hdh ^{140Q/3xFlag20Q} striatum compared to the Hdh ^{140Q/3xFlag7Q} striatum.

In addition to the gliosis observed in HD, elevated levels of lipofuscin are detected in HD mouse models, and in the HD brain [43, 45, 46]. Lipofuscin is an autofluorescent aging pigment that accumulates over time in neurons and other postmitotic cells. It is composed primarily of cross-linked lipid, and is generated by the autophagic catabolism of membrane and organelles [47]. In HD, oxidative stress can result in the increased accumulation of perinuclear lipofuscin deposits in neurons. We assessed lipofuscin accumulation in the striatum of 24 month-old wild-type, Hdh ^{140Q/3xFlag7Q} , and Hdh ^{140Q/3xFlag20Q} , to Hdh ^{140Q/3xFlag140Q} mice, and found that there was a trend towards increased lipofuscin accumulation in the striatum: wild-type<Hdh ^{140Q/3xFlag7Q} <Hdh ^{140Q/3xFlag20Q} <Hdh ^{140Q/3xFlag140Q} . Both Hdh ^{140Q/3xFlag20Q} and Hdh ^{140Q/3xFlag140Q} mice had significantly more striatal lipofuscin in comparison to wild-type mice, but there was no significant difference in the amount of striatal lipofuscin observed between wild-type and Hdh ^{140Q/3xFlag7Q} mice (P=0.056), between Hdh ^{140Q/3xFlag20} and Hdh ^{140Q/3xFlag7Q} mice (P=0.6), or between Hdh ^{3xFlag140Q/140Q} and Hdh ^{140Q/3xFlag20Q} mice (P=0.46) (Figure 11C, D).

Partially complementary oligonucleotides containing sequence encoding an HA or 3xFlag N-terminal epitope tag, an AlwN I restriction site at the 5´end, and an Xmn I restriction site at the 3´end were annealed, repaired with the Klenow fragment of DNA polymerase I, and then digested with AlwN I and Xmn I restriction enzymes to generate a synthetic DNA fragment that was used to replace an endogenouse Hdh exon 1 AlwN I–Xmn I fragment encoding the N-terminal amino acids of htt. Oligonucleotides used were, HA-f: 5´-GTCTTCAGGGTCTGTCCCATCGGGCAGGAAGCCGTCATGTACCCATACGACGTCCCAGACTACGCT-3´, HA-r: 5´-CGACTCGAAAGCCTTCATCAGCTTTTCCAGGGTTGCAGCGTAGTCTGGGACGTCGTATGGGTA-3´, Flag3x-f: 5´-GTCTTCAGGGTCTGTCCCATCGGGCAGGAAGCCGTCATGGACTACAAAGACCATGAGGGTGATTATAAAGATCATGACATCGACTACAAGGACGACGATGACAAG-3´, Flag3x-r: 5´-CGACTCGAAAGCCTTCATCAGCTTTTCCAGGGTTGCCTTGTCATCGTCGTCCTTGTAGTC-3´. For assembly of the Hdh ^3xFlag20Q gene targeting vector, An Xmn I–Kpn I fragment containing human wild-type HTT sequence was obtained by PCR from genomic DNA according to the method described in [32]. This fragment contains the human Xmn I restriction site, a normal 20Q stretch, the human proline-rich region (PRR), human sequence extending to the end of exon 1, a 100 bp intron 1 deletion near the 5´-splice site of intron 1, and a Kpn I restriction site. For assembly of the Hdh ^3xFlag140Q gene targeting vector, the 3xFlag epitope tag AlwN I–Xmn I restriction fragment was used to replace the corresponding AlwN I–Xmn I fragment in our CAG140 targeting vector. Thus, this vector also contains both the human PRR and intron 1 deletion. In contrast, insertion of the epitope tag sequences was the only modification made to the Hdh ^HA7Q and Hdh ^3xFlag7Q gene targeting vectors. Following electroporation of the targeting vectors into the W9.5 ES cell line, targeted clones were identified by Southern blot analysis as described [32]. Southern positive clones were then screened for expression of epitope-tagged full-length htt using anti-HA (HA.11, Covance) and anti-Flag (FLAG M2, Sigma) antibodies. Mice were generated from the targeted ES cell clones using standard procedures, and germline transmission was obtained from at least two independently-targeted ES cell clones for each construct. All lines were backcrossed with C57BL/6J mice for at least six generations.

For genotyping the Hdh ^HA7Q and Hdh ^3xFlag7Q alleles, HDepi-1: 5´-GCGTAGTGCCAGTAGGCTCCAAG-3´and HDepi-2: 5´-CTGAAACGACTTGAGCGACTCGAAAG-3 amplify a portion of the Hdh sequence flanking the epitope tag insertion site between mouse htt amino acids 1 and 2. A wild-type Hdh allele will generate a 112 bp PCR product, while the Hdh ^3xFlag7Q allele will generate a 178 bp product due to the insertion of the 22 amino acid 3xFlag epitope tag. The Hdh ^HA7Q allele generates a 139 bp PCR product due to the insertion of the nine amino acid HA epitope tag. For genotyping the Hdh ^3xFlag20Q and Hdh ^3xFlag140Q alleles, HDepi-1 and HuHDepi-2: 5´-GAAGGACTTGAGGGACTCGAAGG-3´are used. HuHDepi-2 is homologous to the human sequence located 3´of the Xmn I restriction within exon 1 of both the Hdh ^3xFlag20Q and Hdh ^3xFlag140Q alleles. Both the Hdh ^3xFlag20Q and Hdh ^3xFlag140Q alleles will generate a 176 bp PCR product with this primer pair, while no product will be generated from the wild-type Hdh allele. Alternatively, the Hdh ^3xFlag20Q and Hdh ^3xFlag140Q alleles can be genotyped by using 140for: 5´-CTGCACCGACCGTGAGTCC-3´ and 140rev: 5´-GAAGGCACTGGAGTCGTGAC-3´. These primers flank the 100 bp intron 1 deletion that is present in the Hdh ^3xFlag20Q , Hdh ^3xFlag140Q , and Hdh ^140Q alleles, and will generate a 150 bp PCR product, while the wild-type Hdh, Hdh ^3xFlag7Q , and Hdh ^HA7Q alleles lacking the intron 1 deletion generate a 235 bp product. The CAG repeat lengths in the Hdh ^3xFlag20Q (23Q), Hdh ^3xFlag140Q (136Q) alleles were determined by PCR sequencing (Laragen, Inc.) of genomic DNA isolated from tail biopsies.

Whole brain tissue was dounce homogenized in extraction buffer {50 mM Tris–HCl pH 8.8, 100 mM NaCl, 5 mM MgCl₂, 1 mM EDTA pH 8.0, 0.5% (v/v) NP40, protease inhibitor cocktail (Thermo Scientific), with 5 mM NaF and 1 mM sodium vanadate} on ice, and then centrifuged at 16,100×g for 10´at 4°C to obtain a crude cytoplasmic supernatant fraction. Co-immunoprecipitation was performed by incubating protein extracts with either FLAG M2 antibody-agarose beads or by incubating antibody (FLAG M2 or 1C2) with the extract, followed by capture of the antigen-antibody complex with Protein G-agarose beads. In the first method, 0.5–1.0 mg of the protein lysate was incubated by rotation o/n at 4°C with 12–20 μl FLAG M2 antibody-agarose beads (Sigma) that were pre-washed three times with 0.5 ml extraction buffer according to the manufacturer’s instructions. Following incubation of the antibody-agarose beads with the extract, the beads were pelleted by centrifugation at 8,200×g for 30 sec at 4°C, and the supernatant was carefully removed and saved as the antibody non-bound fraction (NB). The beads were then washed three times with 0.5 ml ice-cold extraction buffer. Antibody-bound protein was eluted with two sequential incubations in 30 μl 0.4 mg/ml 3xFlag peptide (Sigma) dissolved in extraction buffer. Both elution fractions were then combined as the antibody-bound (B) fraction. NB- and B-samples were fractionated on 5% SDS-PAGE and then transferred electrophoretically o/n at 30V in 25 mM Tris base, 192 mM glycine, 10% methanol, 0.025% SDS to 0.45 μm PVDF membranes (Millipore). In the second method, 500 μg cytoplasmic extract was incubated o/n at 4°C with 5–6 μg FLAG M2 or 1C2 antibody. 20 μl Protein-G agarose beads (Upstate Biotechnology; washed once in extraction buffer, blocked for 1 h at 4°C in 1% blocking powder {Boehringer} dissolved in TBS, and then washed 3X in extraction buffer prior to resuspending in the original volume of the same buffer), was then added to the mixture of protein extract and antibody. This mixture was then rotated for 3 hr at 4°C to bind the antibody-antigen complexes to the beads. The beads were then pelleted by centrifugation, the supernatant was saved as the NB-fraction, and then the beads were washed as described previously for the anti-FLAG M2-agarose beads. The B-fraction was eluted from the beads by adding 50 μl 3× SDS-PAGE sample loading buffer to the beads, and incubating for 5 min at 99°C in a heating block. The beads were then pelleted, and the supernatant transferred to a fresh tube. 5% of the NB-fraction and 50% of the B-fraction were analyzed by SDS-PAGE and western blotting as described above. Western blot membranes were blocked in 5% milk in TBS-0.05% Tween 20 (TBST) prior to incubation o/n at 4°C with either FLAG M2, (Sigma), HA.11 (Covance), the expanded polyQ-specific monoclonal antibody 1C2 (MAB1574, Millipore) or MAB2166 (Millipore). Membranes were washed five times with TBST before incubation with secondary goat anti-mouse IgG-HRP conjugated antibodies (Pierce) for 1 h at room temperature. Following an additional five washes, the membranes were incubated with West-Dura chemiluminescence substrate (Pierce) and then exposed to film. For quantification of the blots, exposures within the linear range of the film were scanned, and then analyzed by Image J software (http://imagej.nih.gov/ij/download.html). The percentage of 140Q-htt co-immunoprecipitating with 3xFlag20Q-htt was determined by dividing the pixel intensity of the B-fraction 140Q-htt signal by the NB-fraction signal, and then adjusting for the fraction of each sample loaded, and also for the efficiency of 3xFlag20Q-htt immunoprecipitation as determined from the 2166 western blots (e.g. ~20% of 3xFlag20Q-htt was immunoprecipitated in the experiment shown in Figure 5C, and the % of 140Q-htt associating with 3xFlag20Q-htt would then be corrected by a factor of 100/20).

Brains were cut in half (sagittally), and each half brain was dounce-homogenized in 2 ml extraction buffer on ice, before centrifugation at 800×g for 15´ at 4°C to obtain a crude nuclear pellet fraction. Formic acid-solubilized nuclear htt aggregates were obtained following the procedures of [39] and [40] with modifications. The crude nuclear fraction was first resuspended in extraction buffer, and then boiled 15 min in the presence of SDS prior to sonication with a Fisher Scientific 550 probe sonicator (power level 5 for 20 sec with pulses of 0.6 sec on/0.4 sec off). Following sonication, the sample was centrifuged at 16,100xg for 15´at RT, and the SDS-insoluble pellet was resuspended in formic acid and incubated for 1 h at 37°C at 350 rpm. The sample was then dried in a speed-vac and resuspended in 1M Tris base to neutralize residual formic acid. An equal volume of 2X SDS-PAGE loading buffer was added, and then one third of the sample was incubated 5 min at 97°C prior to fractionation on 4%–15% gradient SDS-PAGE (BioRad) and blotting onto 0.45 μm PVDF membrane. Blots were probed with Flag (FLAG M2, Sigma) and expanded polyQ-specific (1C2, Millipore) antibodies, and imaged using a ChemiDoc XRS+ (BioRad) with West-Dura chemiluminescence reagents (Pierce).

Whole brains were dounce homogenized in extraction buffer and centrifuged at 800×g for 15´at 4°C to obtain a crude nuclear pellet fraction. Aliquots of this fraction were treated for 60 minutes on ice with 0.1 mg/ml DNAse I, SDS was then added to 2% final concentration, and incubated 5 minutes at 99°C prior to spotting onto a cellulose acetate/PVDF membrane sandwich. SDS-resistant htt inclusions are trapped on the cellulose acetate membrane while soluble 3xFlag7Q-htt and 140Q-htt species are retained on the underlying PVDF membrane. Htt inclusions were detected by western blotting with the 1C2 and MAB5374 antibodies, while sequestered 3xFlag7Q-htt was detected with the FLAG M2 antibody.

Dissected brains were flash-frozen in isopentane on dry ice, and then serially sectioned at 14 μm on a cryostat (Bright Instrument Co.). Sections were either used immediately or stored at −80°C until use. Sections were washed briefly in PBS, fixed for 10 min in 4% paraformaldeyde in 0.1M phosphate buffer pH 7.4, and washed in PBS three times before blocking with a 1:100 dilution of monovalent Fab fragments (donkey anti-mouse IgG (H+L), Jackson ImmunoResearch) in blocking buffer (5% donkey serum, 0.1% Triton X100 in PBS) for 1 h at RT, followed by three washes in PBS, and then incubated o/n at 4°C with primary antibody diluted in blocking buffer. Primary antibodies used were mouse monoclonal MW8 (1:70, Developmental Studies Hybridoma Bank), mouse monoclonal MAB5375 (1:100, mEM48, Millipore), mouse monoclonal FLAG M2 (1:100, Sigma), mouse monoclonal HA.11 (1:100, Covance), and rabbit anti-GFAP (1:1000, AB5804, Millipore). Following incubation with the primary antibody, sections were washed three times in PBS, and then incubated with secondary antibody (Cy3- or FITC-conjugated donkey anti-mouse or donkey anti-rabbit-IgG, Jackson ImmunoResearch) together with the fluorescent DNA stain To-Pro-3 iodide (Invitrogen) in blocking buffer for 1 h at RT. Sections were then washed with PBS before treating with Autofluorescence Eliminator Reagent (Millipore) following the manufacturer’s instructions. Sections were then mounted with Vectashield (Vector Laboratory) and examined using either an Olympus BX51 microscope equipped with a MagnaFire CCD camera or a Nikon C1 confocal system.

The method described in [41] was adapted for fresh frozen sections. Sections were washed briefly in PBS, and fixed in 4% paraformaldehyde in 0.1M PB for 15 min at RT. Sections were then washed twice in PBS, once with water (10 min each), and then treated with formic acid (Sigma) three times for 10 min each. The sections were rinsed three times in water and washed twice for 10 min each in PBS, followed by 30 min incubation in 1% sodium borohydride (w/v) in PBS, and then washed three times for 10 min each in PBS prior to immunohistochemical staining using the monoclonal 1C2 antibody (1:100, Millipore).

Tissue sections were imaged with a 60X objective using a Nikon C1 confocal microscope, and the number of MW8⁺ huntingtin aggregates was counted blind to genotype from sagittal sections through the striatum (8 images from sections separated by 280 μm through the striatum, n=4 mice of each genotype) using ImagePro 4.5 software (Media Cybernetics). Neuropil and nuclear inclusions were counted separately. To determine the mean size of inclusions, the total pixel area was divided by the number of inclusions in each imaging field.

Lipofuscin analysis . Imaging and quantification of lipofuscin was performed as described in [46].

χ² analyses of the genotypes were performed using SigmaStat (Systat software). Student’s t-test was used to evaluate significance in the quantification of aggregate numbers, aggregate mean size, and co-localization of MW8 and 3xFlag epitopes in htt aggregates.