Passive immunization against phosphorylated tau improves features of Huntington's disease pathology

Abstract

Huntington's disease is classically described as a neurodegenerative disorder of monogenic aetiology. The disease is characterized by an abnormal polyglutamine expansion in the huntingtin gene, which drives the toxicity of the mutated form of the protein. However, accumulation of the microtubule-associated protein tau, which is involved in a number of neurological disorders, has also been observed in patients with Huntington's disease. In order to unravel the contribution of tau hyperphosphorylation to hallmark features of Huntington's disease, we administered weekly intraperitoneal injections of the anti-tau pS202 CP13 monoclonal antibody to zQ175 mice and characterized the resulting behavioral and biochemical changes. After 12 weeks of treatment, motor impairments, cognitive performance and general health were improved in zQ175 mice along with a significant reduction in hippocampal pS202 tau levels. Despite the lack of effect of CP13 on neuronal markers associated with Huntington's disease pathology, tau-targeting enzymes and gliosis, CP13 was shown to directly impact mutant huntingtin aggregation such that brain levels of amyloid fibrils and huntingtin oligomers were decreased, while larger huntingtin protein aggregates were increased. Investigation of CP13 treatment of Huntington's disease patient-derived induced pluripotent stem cells (iPSCs) revealed a reduction in pS202 levels in differentiated cortical neurons and a rescue of neurite length. Collectively, these findings suggest that attenuating tau pathology could mitigate behavioral and molecular hallmarks associated with Huntington's disease.

Graphical abstract

Tau and hyperphosphorylated tau are associated with multiple neurodegenerative conditions, including Huntington's disease. In the current study, the authors show that reduction of phosphorylated tau through antibody treatment improved behavior and altered huntingtin aggregation kinetics in a mouse model as well as restoring neurite length in a human cell-culture model.

Introduction

Huntington's disease (HD) is an autosomal dominant neurodegenerative disorder characterized by severe motor, cognitive and psychiatric impairments.^1,2 While motor deficits are most commonly associated with the condition, cognitive dysfunctions often manifest early in disease progression, are clinically very difficult to treat and are consistently among the greatest contributors to reduced quality of life.³ HD is caused by a mutation in the huntingtin (HTT) gene characterized by an expansion of over 35 CAG repeats in exon 1, resulting in an elongated polyglutamine stretch at the N-terminus of the huntingtin protein (HTT). The mutant form of HTT (mHTT) is prone to misfolding, oligomerization and aggregation, thus generating large intracellular and extracellular aggregates in the central nervous system (CNS) and peripheral organs of patients.4, 5, 6, 7, 8

In addition to accumulation of mHTT, cardinal features of tauopathies, namely misfolding, hyperphosphorylation and aggregation of the tau protein into neurofibrillary tangles (NFTs) and neuropil threads (NTs), have been described in the brains of HD patients.^9,10 Tau is primarily expressed in neurons¹¹ and is involved in essential cellular processes, such as microtubule stability, neurite outgrowth, axonal transport and synaptic plasticity.12, 13, 14 Post-translational modifications are key molecular mechanisms responsible for regulating these activities and include phosphorylation and dephosphorylation at multiple residues.^15,16 Under pathological conditions, disruption of biochemical conversions between phosphorylated and dephosphorylated states leads to increased tau hyperphosphorylation,^10,17, 18, 19 destabilized microtubules,^16,20,21 and consequent assumption of aggregation-prone conformations, ultimately leading to cellular dysfunction and toxicity.^15,22 Hyperphosphorylated tau, either in its soluble or aggregated form, has been detected in late disease stages^23,24 and appears to correlate with stage-related cognitive deficits in HD.^25,26 Furthermore, well-validated transgenic and knock-in mouse models of HD acquire features of tau pathology, with increased hyperphosphorylated tau levels in the cortex, hippocampus and striatum.^19,27 There is compelling evidence to suggest that HD may be a secondary tauopathy, urging the evaluation of tau's contribution to behavioral and molecular signatures of this neurodegenerative disorder.

Immunotherapies, i.e., harnessing the immune system to clear pathological proteins, can efficiently decrease phosphorylated tau levels in animal models of tauopathies,^10,28 suggesting that such approach could be used in HD to untangle the contribution of tau pathology to disease progression. Some studies have now reported that injection of anti-phosphorylated tau antibodies into animal models of tauopathy is associated with improved behavioral outcomes and favorable toxicology and safety profiles.^29,30 For example, the monoclonal antibody CP13 has been engineered to recognize tau phosphorylated at serine 202 (pS202) and has been reported to reduce tau phosphorylation in the cortex and hindbrain of the JNPL3 mouse model of severe tauopathies.²⁸ This study provides a proof of concept that anti-phosphorylated tau antibodies are not only therapeutic avenues to explore but also tools that can be leveraged to dissect the roles of pathological proteins in disease onset and progression. Therefore, our study implemented an in vivo immunization-based approach to evaluate the contributions of hyperphosphorylated tau to HD behavioral and cellular pathology. We found that weekly CP13 antibody treatments reduced tau phosphorylation at serine 202 in the hippocampus, improved motor and cognitive deficits and attenuated hallmarks of HD pathology, such as the production of toxic HTT oligomers within the brain. We also show that CP13 reduced tau phosphorylation and rescued neurite length in iPSC-derived cortical neurons produced from patients with HD, thus confirming its efficacy in a human cellular model of HD.

Results

CP13 treatment ameliorates cognitive and motor performance in HD mice

To determine whether a reduction in phosphorylated tau levels could mitigate HD-related cognitive and motor deficits, we used the zQ175 knock-in mouse model that is not only characterized by the progressive appearance of HD-related molecular and behavioral features³¹ but, importantly, recapitulates aspects of tau pathology, including accumulation of pS202 and pS199.¹⁹ Specifically, we administered the CP13 antibody, which binds pS202 and has previously been shown to reduce phosphorylated tau levels in a mouse model of tauopathy,²⁸ or saline to zQ175 mice and wild-type (WT) littermate controls via weekly intraperitoneal injections of 10 mg/kg of antibody. To ensure that tau pathology and behavioral impairments were mild at the start of the experiment, treatment began at 6 months of age and was pursued for a total of 12 weeks. A battery of cognitive and motor tests were conducted prior to the first injection and monthly until the end of the experiment (Figure 1A). Memory impairments are generally well correlated with tau pathology32, 33, 34, 35, 36 and we accordingly began our behavioral analysis by assessing working memory using spontaneous alternation in the Y maze. Three months after the start of treatment, when mice were 9 months of age, zQ175 animals demonstrated a decrease in both the total (Figure 1B) and correct number of entries on this test (Figure 1C). To ensure that these impairments result from a cognitive decline and not motor dysfunction, cognition was assessed by calculating the percentage of correct alternations. This measure was significantly increased in zQ175 mice treated with CP13 at the 3-month time point of testing (Figure 1D), indicating an improvement in working memory. No beneficial effects were observed at earlier time points. Distance traveled in the open field yielded similar results to those observed with theY maze. Specifically, CP13 injections significantly increased distance traveled by zQ175 mice, demonstrating a reduction in motor impairments (Figure 1E). To further assess cognitive changes, distance traveled in the open field was evaluated as a measure of habituation.³⁷ In our cohort, HD mice displayed decreased habituation within (Figure 1F) and between testing sessions (Figure 1G), which is indicative of impairment in both working and long-term memory. The beneficial effect on long-term memory was observed after only 2 months of treatment. All other behavioral improvements were measured after 3 months of treatment (Figure 1G).

Figure 1.

CP13 antibody treatment improves motor and cognitive performance in zQ175 mice

(A) Timeline of antibody treatment and behavioral evaluations. (B–D) Cognitive performance was assessed as the total number of entries (B), the total number of correct entries (C) and the percent of correct entries (D) in the Y maze. (E–G) Motor performance was assessed as the distance traveled in the open field (E). Short- and long-term memory was evaluated by measuring intrasession (F) and intersession (G) habituation in the open field. Data are presented as mean ± SEM with individual animal results indicated as data points. WT (S) n = 8–9; WT (CP13) n = 6–7; zQ175 (S) n = 7–10; zQ175 (CP13) n = 9–10. For all graphs, statistics were performed using a two-way ANOVA with Sidak's post-hoc test. ^#For (E), we confirmed a genotype difference between saline-treated WT and zQ175 mice by performing a Student's t test. ∗p ≤ 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. m, meters; S, saline; WT, wild type.

Following the last dose of treatment, nesting behavior was evaluated as a measure of overall health and well-being. Once again, antibody-treated zQ175 mice displayed higher scores as compared with their saline zQ175 littermates (Figure S1A), reflecting an amelioration of overall health in these animals. Collectively, these observations suggest that enhancement of general locomotor activity is accompanied by an improvement in short- and long-term memory. The absence of a significant difference in anxiety-like behavior between groups in the open field test (Figures S1B and S1C) suggests that the CP13 antibody specifically recovers features associated with motor and cognitive performance in HD mice.

CP13 treatment reduces tau phosphorylation in brain structures targeted in HD

The CP13 monoclonal antibody recognizes phosphorylated tau at the S202 residue, a post-translational modification associated with tau pathology and inclusion bodies characteristic of tauopathies.¹⁰ As described above, we observed an amelioration of motor and cognitive deficits in CP13-treated HD mice, suggesting that peripheral antibody administration improved CNS-related pathology. To confirm these improvements were associated with the presence of CP13 in the brain parenchyma, we developed a sandwich ELISA-based assay capable of detecting anti-tau antibodies in perfused brain homogenates (Figure S2A). We found significantly greater levels of anti-tau antibodies in the brains of CP13-treated WT and zQ175 animals as compared with controls (Figures S2B and S2C), suggesting efficient penetrance of the antibody into the brain. We therefore measured the levels of pS202 in three brain structures affected in tauopathies and/or HD, namely the cortex, hippocampus and striatum. The cerebral cortex is a brain region particularly vulnerable in HD and primary tauopathies,^10,38, 39, 40 as demonstrated by a significant neuronal loss and atrophy of specific cortical areas.⁴¹ Cortical tissue was analyzed by western blot to measure pS202 (Figure 2A) and saline-treated zQ175 mice showed higher baseline levels of pS202 tau compared with saline-treated WT animals. However, injection of CP13 in zQ175 mice induced a trend towards decreased pS202 tau levels compared with saline-treated zQ175 animals (Figures 2B and 2C). To understand whether this trend was specific to the targeted tau residue, pS202, we measured two other sites, or combinations of sites, that are commonly phosphorylated in mouse models of tauopathy: pS199 and pS396/404. Neither of these targeted residues was altered in CP13-treated zQ175 animals (Figures 2D and 2E) suggesting that the CP13 antibody specifically targets pS202 in zQ175 mice. In contrast to these findings, a significant increase in pS396 levels was observed in WT mice. The reason for this increase is unclear however the trend towards a decrease in total tau levels may have contributed to this observation. In the hippocampus, the brain region that is generally most affected in tauopathies,^42,43 we observed a significant reduction of pS202 (but not pS199 or pS396) tau levels in zQ175 mice injected with the antibody compared with saline controls (Figures 2F–2J) as well as a significant increase in pS396 in WT mice. This increase seems to be at least partially due to a significant reduction in the amount of total tau present in WT mice. Finally, we analyzed tau phosphorylation in the striatum, the most affected brain region in HD pathology.^10,38 We observed that striatal tissue isolated from zQ175 mice expressed significantly greater levels of pS199 tau compared with WT mice, but antibody injections did not alter phosphorylated tau levels at any of the three analyzed residues (i.e., pS199, pS202, and pS396) in either genotype (Figures 2K–2O). Together, these observations suggest that CP13 specifically targets pS202 tau in zQ175 mice, particularly where levels are elevated and that there is no correlation between pS199, pS202, and pS396 levels following antibody treatment.

Figure 2.

CP13 antibody treatment decreases levels of phosphorylated tau in the cortex and hippocampus of zQ175 mice

(A, F, and K) Representative immunoblots depicting protein levels of tau phosphorylated at serine 202 (pS202), serine 199 (pS199), serines 396/404 (p396/404), total tau (t-tau) and total protein (tot prot) in the cortex (A), hippocampus (F) and striatum (K) of WT or zQ175 mice treated with CP13 or saline. (B–E) Quantification of protein levels in cortical homogenates and calculated ratios of t-tau normalized over tot prot (B), pS202 (C), pS199 (D) or pS396/404 (E) normalized to t-tau levels. (G–J) Quantification of protein levels in hippocampal homogenates and calculated ratios of t-tau normalized over tot prot (G), pS202 (H), pS199 (I) or pS396/404 (J) normalized to t-tau levels. (L–O) Quantification of protein levels in striatal homogenate and calculated ratios of t-tau normalized over tot prot (L), pS202 (M), pS199 (N) or pS396/404 (O) normalized to t-tau levels. Data are presented as mean ± SEM with individual animal results indicated as data points. WT (S) n = 3–4; WT (CP13) n = 3–4; zQ175 (S) n = 4–6; zQ175 (CP13) n = 5–6. For all graphs, statistical analyses were performed using a two-way ANOVA with Sidak's post-hoc tests. ∗p ≤ 0.05 and ∗∗p < 0.01. kDa, kilodalton; t-tau, total tau.

While the specific reduction of pS202 suggests that the antibody directly targets tau, it is also possible that the antibody altered tau phosphorylation through effects on enzymes. Tau phosphorylation state is regulated by kinases and phosphatases in response to intracellular and extracellular cues, and changes in the kinetics of phosphorylation and dephosphorylation can result in tau hyperphosphorylation.44, 45, 46, 47, 48 The roles of the GSK3β kinase and PP2B phosphatase in tau pathology have been extensively described¹⁰ and we sought to determine whether CP13 treatment attenuated tau hyperphosphorylation by affecting either or both of these enzymes. We therefore measured, by western blot, GSK3β and PP2B protein levels but did not find significant changes in CP13-injected zQ175 mice compared with their saline-injected counterparts (Figures S3A–S3F). These findings are consistent with the specific effect of the CP13 antibody on phosphorylation at S202.

Overall, these findings confirm previous observations of tau hyperphosphorylation in a number of brain regions implicated in HD pathology¹⁰ and suggest that immunotherapies can target and reduce tau phosphorylation in the CNS, thereby improving tau-associated behavioral phenotypes (Figure S4).

CP13 injections do not trigger gliosis

Astrocytes and microglia are essential mediators of CNS health and their coordinated action is necessary to maintain brain homeostasis.^49,50 These cells constantly survey their microenvironment and transition from non-reactive to reactive phenotypes in response to contact with foreign material.⁵¹ To ensure that administration of exogenous CP13 did not initiate an inflammatory response as a result of antibody-protein complexes in the brain parenchyma, we evaluated signs of gliosis in the cortex, hippocampusand striatum of treated animals. In all brain regions analyzed, the total number of astrocytes as well as the protein levels of the astrocyte/inflammatory marker GFAP were similar between antibody-treated zQ175 and WT mice (Figures 3A–3G). We then analyzed the microglial response in the three brain regions using the microglial/inflammatory marker Iba1 and found that the total microglia count in the striatum and hippocampus remained unchanged, but we observed an increased number of microglia in the cortex of CP13-injected zQ175 mice compared with antibody-injected WT animals (Figures 3H–3K). In addition, the cortex (but not the hippocampus or striatum) of saline-injected zQ175 mice expressed higher protein levels of Iba1 than saline-injected WT animals (Figures 3L–3N). However, our data show that CP13 injection did not impact Iba1 protein levels in the hippocampus, cortex or striatum of zQ175-versus WT-treated animals. Overall, these findings suggest that zQ175 mice have higher Iba1 expression and microglial numbers than WT mice, but CP13 treatment does not upregulate neuroinflammation-associated proteins (i.e., GFAP and Iba1) or induce brainwide astrocyte and microglia proliferation (Figure S4).

Figure 3.

CP13 antibody treatment does not induce a glial response

(A) Representative images of cortical astrocytes labeled by immunohistochemistry with the GFAP marker. (B–D) Quantification of the number of GFAP-positive astrocytes in brain sections prepared from the cortex (B), hippocampus (C)and striatum (D). (E–G) Quantification of GFAP protein levels by western blot using homogenates prepared from the cortex (E), hippocampus (F) and striatum (G). (H) Representative images of cortical microglia labeled by immunohistochemistry with the Iba1 marker. (I–K) Quantification of the number of Iba1-positive microglia in brain sections prepared from the cortex (I), hippocampus (J) and striatum (K). (L–N) Quantification of Iba1 protein levels by western blot using homogenates prepared from the cortex (L), hippocampus (M) and striatum (N). Data are expressed as mean ± SEM with individual animal results indicated with data points. WT (S) n = 3–4; WT (CP13) n = 3–4; zQ175 (S) n = 3–6; zQ175 (CP13) n = 3–6. For all graphs, statistics were performed using a two-way ANOVA with Sidak's post-hoc test. ∗p < 0.05 and ∗∗p < 0.01. Scale bars: 20 μm. GFAP, glial fibrillary acidic protein; Iba1, ionized calcium binding adaptor molecule; k, thousands; N, number; Tot prot, total protein.

Reduction in phosphorylated tau levels does not alter neuronal markers in HD mice

Extending on the post mortem analyses, we characterized the expression of four neuronal markers to determine whether CP13 treatment ameliorates deficits previously described in zQ175 animals as well as in the R6/1 mouse model expressing human HTT exon 1 with 116 CAG repeats.^31,52, 53, 54 We first determined cortical and hippocampal protein levels of SMI32 (a.k.a. Neurofilament H) to detect non-phosphorylated neurofilaments and measure axonal damage.⁵⁵ We did not detect differences in hippocampal SMI32 levels between saline-injected zQ175 and WT mice, but we found a significant decrease in cortical tissue samples of HD animals that was not rescued by CP13 treatment (Figures 4A and 4B). We then analyzed DARPP32 striatal levels to assess the effect of antibody injections on medium spiny neurons (the most impacted population of striatal neurons in HD),^56,57 but we did not observe a rescue of the zQ175 phenotype as demonstrated by significantly lower DARPP32 protein levels in antibody-treated zQ175 versus WT mice counterparts (Figure 4C). Finally, we measured changes in the expression of vesicular glutamate transporter 1 (VGlut1), a glutamatergic neuronal marker abundantly expressed in the cortex and hippocampus,⁵⁸ two regions affected by HD and tauopathies.59, 60, 61 HD pathology is in part characterized by the dysfunction of glutamatergic cortical neurons,^61,62 which leads to the disruption of glutamatergic innervation of the striatum.^38,63 Our data indicate that CP13 injections did not increase VGlut1 levels in zQ175 mice in the cortex, hippocampus or striatum compared with saline-treated zQ175 littermates (Figures 4D–4F). In addition, analysis of postsynaptic marker PSD95 revealed that CP13 treatment did not increase the levels of this synaptic protein in any of the three tested brain regions (Figures 4G–4I), although CP13 treatment in WT mice did increase PSD95 in the cortex and reduced PSD95 in the hippocampus. These results suggest that the reduction in phosphorylated tau levels does not mediate behavioral improvement through rescue of neuronal populations or synaptic connections (Figure S4).

Figure 4.

CP13 antibody treatment does not alter neuronal markers

(A and B) Quantification of protein levels and representative immunoblots detecting markers of axon damage (SMI32) in the cortex (A) and hippocampus (B). (C) Quantification of DARPP32 protein levels in striatal tissue homogenate by western blot. (D–F) Quantification of protein levels and representative immunoblots detecting the glutamatergic neuron marker VGlut1 in the cortex (D), hippocampus (E) and striatum (F). (G–I) Quantification of protein levels and representative immunoblots detecting the post-synaptic protein PSD95 in homogenates prepared from the cortex (G), hippocampus (H) and striatum (I). (G–I) Data are expressed as mean ± SEM with individual animal results indicated with data points. WT (S) n = 3–4; WT (CP13) n = 3–4; zQ175 (S) n = 4–6; zQ175 (CP13) n = 4–6. For all graphs, statistics were performed using a two-way ANOVA with Sidak's post-hoc test. ∗p ≤ 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. PSD95, postsynaptic density protein 95; VGLT1, vesicular glutamate transporter 1.

CP13 injections alter mHTT aggregation in HD mice

Patients with HD produce an abnormal form of the HTT protein that acquires an increased propensity to misfold and generates a variety of aggregate species, from small oligomers to large inclusion bodies.^6,64 We therefore analyzed whether our immunotherapy approach could alter levels of soluble and insoluble mHTT. We found that the antibody treatment did not change monomeric HTT and mHTT levels in CP13- versus saline-injected animals when tissue homogenates were analyzed by western blot in the cortex (Figures 5A–5C), hippocampus (Figures 5D–5F) or striatum (Figures 5G–5I). We then performed filter retardation assays to detect the presence of SDS-insoluble aggregates using two different HTT antibodies. Prior to blotting onto a nitrocellulose membrane with a pore size of 0.2 μM, samples were boiled in a solution containing 0.5% SDS and 0.1% DTT in order to solubilize smaller aggregates species. Using this method, our data show a 1.8- to 2.7-fold increase in the levels of aggregated HTT in the cortex and striatum of CP13-injected zQ175 mice compared with their saline-injected counterparts (Figures 6A–6C). To further characterize changes associated with different sub-populations of aggregates, we used stereology to estimate the total number of aggregates as well as the volume of individual aggregates present in each brain region of interest. This analysis was performed using brain sections immunoreacted with an antibody that recognizes HTT aggregates (EM48). We evaluated the distribution of very small (0.25–0.5 μm³), small (0.5–2 μm³), medium (2–4 μm³) and large (>4 μm³) aggregate populations and found a significant increase in the number of small aggregates in the cortex and hippocampus of CP13- versus saline-injected zQ175 mice (Figures 6D–6H). There was also a trend toward an increased number of medium aggregates in the striatum in antibody-treated zQ175 mice (Figure 6I). While the significant increase in the number of small aggregates observed in the cortex was absent in the striatum, collectively, these findings suggest that tau hyperphosphorylation influences the dynamics of mHTT aggregation in the CNS, with reductions in phosphorylated tau correlating with the increased formation of insoluble mHTT aggregates in these two structures (Figure S4).

Figure 5.

CP13 antibody treatment does not affect soluble HTT and mHTT levels in the brain

(A, D, and G) Representative confocal photomicrographs of brain slices labeled by immunofluorescence to detect HTT. (B–I) Quantification of protein levels and representative immunoblots showing soluble HTT (MAB2166 antibody) and soluble mHTT (MAB1574 antibody) in cortical (B and C), hippocampal (E and F) and striatal (H and I) brain homogenates. Data are expressed as mean ± SEM with individual animal results indicated with data points. WT (S) n = 3–4; WT (CP13) n = 3–4; zQ175 (S) n = 4–6; zQ175 (CP13) n = 4–6. For all graphs, statistics were performed using a two-way ANOVA with Sidak's post-hoc test;∗p < 0.05 and ∗∗∗∗p < 0.0001. Scale bars: 10 μm. HTT, huntingtin; mHTT; mutant huntingtin.

Figure 6.

Reduced tau phosphorylation is associated with increased cortical and striatal insoluble HTT and mHTT

(A–C) Representative immunoblots and quantification of filter retardation assays targeted to the detection of insoluble HTT and mHTT in tissue homogenates prepared from the cortex (A), hippocampus (B) and striatum (C) using N18, EM48, 1C2 and D7F7 antibodies. (D–F) Graphs show a stereological-based quantification for the number of HTT aggregates and their distribution across the 0–6 μm³ size range in the cortex (D), hippocampus (E) and striatum (F). The aggregates display a range of volumes and are categorized as very small (0.25–0.5 μm³), small (0.5–2 μm³), medium (2–4 μm³) or large (>4 μm³). (G–I) Graphs showing single data points in the cortex (G), hippocampus (H) and striatum (I) extracted from the respective aggregate distribution (D–F). Data are expressed as mean ± SEM with individual animal results indicated with data points. WT (S) n = 4–5; WT (CP13) n = 3–4; zQ175 (S) n = 3–5; zQ175 (CP13) n = 3–6. Statistical analyses were performed using a two-way ANOVA with Sidak's post-hoc test (A–C) or a Student's t test (G–I). ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Est, estimated.

Reduced tau phosphorylation is associated with decreased oligomers and amyloid fibrils in the CNS of zQ175 mice

Accumulating evidence supports the hypothesis that the process of pathological protein aggregation, from monomeric protein misfolding to the formation of large intracellular aggregates, generates a range of intermediate species of varying toxicity.64, 65, 66, 67 Early in the process, misfolded protein monomers assemble into unstable toxic oligomers and subsequently form fibrillar species that can accelerate pathology.67, 68, 69 Toward the end of the spectrum, large protein aggregates are believed to be the least toxic pathological structures.^70,71 In light of our observations, we sought to determine whether CP13 injections could reduce the load of toxic oligomeric species by shifting the protein aggregation process to an increased production of insoluble protein aggregates. We therefore measured oligomer and amyloid fibril levels in the cortex, hippocampus and striatum using dot blot assays and found that CP13 significantly decreased the levels of protein or peptide oligomers in the cortex of antibody-treated zQ175 mice compared with saline-treated littermates (Figure 7A). However, we did not observe changes in the hippocampus and striatum (Figures 7B and 7C). In addition, dot blot experiments detected lower levels of amyloid fibrils in the cortex, hippocampus and striatum of HD mice as a result of CP13 injections (Figures 7D–7F). These findings suggest a decrease in the overall levels of oligomers and fibrils in the brains of CP13-injected animals. However, as these antibodies are directed against structures common to all misfolded proteins, they do not permit us to identify whether they are HTT oligomers or fibrils. This lack of specificity for HTT could explain why significant differences in oligomer and fibril levels were not detected between saline-treated WT and zQ175 animals. We therefore established a semi-denaturing detergent agarose gel electrophoresis (SDD-AGE) assay that enables the detection of protein polymers from small oligomers to larger aggregates⁷² and that is similar to protocols previously used to assess mHTT aggregates in HD mouse models.⁷³ Using this method, we measured HTT-specific changes as a result of CP13 injections. This technique complements our filter retardation and dot blot experiments and indicates global—but not species-specific—changes in the levels of HTT aggregates. SDD-AGE results show that antibody treatment decreased total aggregated HTT in the hippocampus, but not in the cortex or striatum, of CP13-injected zQ175 mice when compared with saline-treated littermates (Figures 7G–7I). In summary, our biochemical investigations of the hippocampus revealed (1) a reduction in the levels of amyloid fibrils (dot blot), (2) no changes in the amount of larger HTT aggregates (filter retardation assay) and (3) a reduction in overall HTT aggregate species (SDD-AGE). We can therefore infer that HTT oligomers are likely the most decreased species upon CP13 treatment of zQ175 animals. In addition, analysis of the cortex and striatum of antibody-treated zQ175 mice showed (1) an increase in large HTT aggregates (filter retardation assay), (2) a decrease in protein oligomers and/or amyloid fibrils (dot blot) and (3) no changes in total levels of aggregated HTT (SDD-AGE). These findings suggest that antibody injections may have preferentially decreased cortical and striatal HTT oligomers and fibrils and increased the formation of larger HTT aggregates (Figure S4).

Figure 7.

CP13 antibody treatment decreases brain levels of oligomers and amyloid fibrils

(A–F) Quantification of protein oligomers detected by dot blot and representative membranes probed using an oligomer- (A11) or amyloid-fibril-specific (OC) antibody in tissue homogenates prepared from the cortex (A and D), hippocampus (B and E) and striatum (C and F). (G–I) Quantification of HTT aggregates by SDD-AGE and representative membranes shown for the cortex (G), hippocampus (H) and striatum (I). Data are expressed as mean ± SEM with individual animal results indicated with data points. WT (S) n = 3–4; WT (CP13) n = 3–4; zQ175 (S) n = 4–6; zQ175 (CP13) n = 4–6. For all graphs, statistical analyses were performed using a two-way ANOVA with Sidak's post-hoc test. ∗p ≤ 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001.

Overall, these results suggest that decreased tau hyperphosphorylation is associated with reduced oligomeric and fibrillar forms of proteins in the brains of HD animals, which could be mediated by an accelerated aggregation process, reduced protein misfolding and/or increased protein degradation. Furthermore, the combination of dot blot, SDD-AGE and immunohistochemistry suggests that reduced oligomer and fibril levels are, at least partially, attributable to changes associated with the HTT protein.

CP13 reduces tau hyperphosphorylation in human induced-pluripotent-stem-cell-derived cortical neurons

To understand whether the beneficial properties of CP13 could extend to vulnerable human cell populations, we tested the ability of our antibody to alter levels of phosphorylated tau in HD-patient-derived cortical neurons. We therefore produced human iPSCs-derived neurons prepared from sex-matched control individuals or patients with HD (control, 17/18 HTT CAG repeats; HD, 60 or 180 HTT CAG repeats). Following a well-established protocol by Shi et al.,⁷⁴ iPSCs were differentiated into neural progenitor cells (NPCs) vial dual SMAD inhibition, followed by maturation into cortical neurons (Figures 8A and 8B). This protocol was selected as previous studies have reported a significant reduction in neurite length in neurons generated from HD patient iPSCs.⁷⁵ After 45 days in culture, the cells were assessed for expression of the pan-neuronal marker microtubule-associated protein 2 (MAP2) and the glutamatergic neuronal marker VGlut1. We obtained a yield of 99% and 96% glutamatergic neurons (defined as MAP2⁺/VGlut1⁺) in cultures differentiated from control and HD lines, respectively (Figure 8C). Prior to performing the experiments aimed at characterizing the effects of passive immunization in these cultures, HD neurons were treated with CP13 and assessed for antibody internalization. We found that CP13 localized to the neurites and cell body, suggesting its efficient uptake by neuronal cells (Figure 8D). Previous work with HD neurons conducted in our laboratory had demonstrated that a significant increase in the levels of phosphorylated tau could be observed when compared with control neurons. We therefore took advantage of this feature and measured the ability of our antibody to reduce tau hyperphosphorylation in a human-based model. Specifically, iPSC-derived cortical neurons prepared from control or HD donors were treated with CP13 or an immunoglobulin G (IgG) control antibody at a concentration of 15 μg/mL for 72 h. Quantification of pS202 levels by western blot showed a 25% reduction in CP13-treated, but not IgG-treated, HD neurons (Figure 8E). To evaluate the effects of CP13 on neuronal health, we then treated HD neurons with CP13 or an IgG control for 14 days. Morphological assessment revealed a significant rescue of neurite retraction in HD neurons treated with CP13 compared with IgG or untreated HD neurons (Figure 8F). These results show that the anti-phosphorylated tau antibody is able to efficiently target human iPSC-cortical neurons, decrease neuronal levels of pathological proteins and ameliorate neuron morphology (Figure S4).

Figure 8.

CP13 antibody treatment reduces tau hyperphosphorylation in human iPSC-derived cortical neurons

(A) Graphical representation of the experimental procedure implemented to differentiate iPSCs into cortical neurons. (B) Representative confocal images of control and HD iPSC-derived neurons immunostained for MAP2 (red), VGlut1 (green)and nuclear stain DAPI (blue). (C) Quantification of the proportion of iPSC-derived neurons immunopositive for VGlut1 and MAP2 neuronal markers. Data are from three biological replicates and a total of 600 control and 572 HD neurons were counted. For both controls and HD, two independent iPSC lines each were analyzed (control lines, 17/18 CAG repeats; HD lines, 180 CAG repeats and 60 CAG repeats). (D) Representative confocal images of HD iPSC-derived neurons immunostained for MAP2 (green), anti-mouse secondary antibody (red) and nuclear stain DAPI (blue). (E) Quantification of pS202 tau normalized to total tau and beta actin protein levels as well as representative immunoblots showing pS202 tau, total tau and the loading control β-actin in iPSC-derived cortical neurons. Data are from four biological replicates prepared using two independent control iPSC lines (17/18 CAG repeats) and one HD line (180 CAG repeats). (F) Quantification of neurite length after 14 days in culture in absence or presence of 15 μg/mL IgG or CP13. Data are from two biological replicates and more than 200 neurites were measured in each experimental condition (control lines, 17/18 CAG repeats; HD lines, 180 CAG repeats and 60 CAG repeats). Statistical analysis was performed using a one-way ANOVA with Dunnett's multiple comparisons test (E and F); ∗p ≤ 0.05 and ∗∗p < 0.01. Outliers in (E) were identified using the Grubbs test (alpha = 0.05) and excluded from analysis. Scale bars: 25 μm (B) and 10 μm (D). IF, immunofluorescence; iPSCs, induced pluripotent stem cells; NPCs, neural progenitor cells; ns, not significant.

Discussion

Passive immunization against pS202 tau ameliorates motor and cognitive impairments in the zQ175 mouse model

This study demonstrates that peripheral injection of an anti-tau pS202 antibody ameliorates a number of behavioral measures in the zQ175 knock-in mouse model without resulting in changes in WT mice treated with the same antibody. Weekly intraperitoneal injections resulted in improved short- and long-term memory, a greater inclination of mice to explore their environment and an overall improvement in locomotor activity and general health (Figures 1 and S1). While the behavioral improvements extended to multiple cognitive and motor tests, they did not extend to anxiety-like behavior. In Alzheimer's disease (AD), the relationship between tau and cognition has been shown in multiple clinical studies through the demonstration of a correlation between the severity of cognitive impairments and the degree of tau accumulation.⁷⁶ The correlation between tau and cognition extends beyond AD and has additionally been described for Parkinson's disease,⁷⁷ amyotrophic lateral sclerosis⁷⁸ and mild cognitive impairment.⁷⁹ With regards to motor symptoms, the greatest evidence for an involvement of tau comes from primary tauopathies, such as progressive supranuclear palsy and frontotemporal dementia with parkinsonism-17, which are characterized by a combination of motor and cognitive impairments.^80,81 Furthermore, an important study by Fernandez-Nogales and colleagues⁸² showed that decreasing expression of the microtubule-associated protein tau (Mapt) gene in the R6/1 mouse model of HD (i.e., Mapt^−/+ or Mapt^−/− R6/1) reduced striatal tau levels and ameliorated motor features measured in the rotarod and activity cage. These previous reports support our observations that improved behavioral outcomes in zQ175 animals are correlated with reduced tau pathology in the cortex and hippocampus (Figures 1 and 2). Furthermore, our findings suggest that CP13 antibodies injected to the periphery are detectable in the brain and able to mediate central effects, which corroborates previous studies indicating that antibodies injected to the periphery can reach their targets in the CNS and promote the degradation and clearance of pathological proteins.^29,83, 84, 85 These effects included a specific reduction of pS202 tau in the hippocampus of zQ175 mice and an increase in pS396 in the cortex and hippocampus of WT mice, as well as decreased total tau levels in the hippocampus of WT mice. While the latter result was unexpected, there is no indication that the reduction of total tau or the increase in pS396 resulted in any behavioral changes in these animals. Longer term treatment would be required to better understand both the duration of these changes and their lasting consequences.

While these changes in tau levels indicate the antibody does reach the brain, published literature suggests that the amount entering the brain is likely limited by the blood-brain barrier (BBB), a major obstacle to overcome for peripherally injected antibodies. Studies have shown that less than 0.2% of intravenously injected antibodies enter the brain.^86,87 Despite this limited penetrance and low levels in cerebral tissues, CNS cells, such as neurons and microglia, possess the molecular machinery to respond to the small number of antibodies that do reach the brain. These cells express Fcγ II/III receptors that recognize Fc fragments and thus enable antibody internalization via receptor-mediated endocytosis.^88,89 The probability of such a mechanism occurring in our experimental model is strengthened by the ELISA data, which indicate that CP13 antibody is present in the hippocampal parenchyma of treated WT and zQ175 mice. Furthermore, our in vitro cell culture experiments demonstrate that the CP13 antibody can be taken up by neurons. However, even if the antibody does not enter neurons in vivo, tau secretion is a physiological process and pathological hyperphosphorylation leads to the secretion of phosphorylated tau as free-floating proteins or enclosed within extracellular vesicles.⁹⁰ As a result, extracellular pathological tau can be internalized by neighboring cells and exacerbate and extend pathology to surrounding cells.^90,91 It is therefore reasonable to suggest that, upon crossing the BBB and entering the CNS, CP13 can recognize, bind and promote degradation of intracellular as well as extracellular tau proteins.

However, the presence of CP13 or CP13-tau complexes in the brain parenchyma could induce an immune response and we therefore investigated changes related to gliosis in the three previously analyzed brain regions. We did not detect brainwide signs of glial reactivity (i.e., increase in GFAP or Iba1 levels) in CP13-injected mice (Figure 3), suggesting that astrocytes and microglia could have contributed to extracellular antibody-protein removal without inducing a significant inflammatory response. In support of our findings, internalization of antibody-tau complexes by microglia does not exacerbate the inflammatory response,⁸⁹ although brain-localized monomeric IgG has been reported to promote mild but neuroprotective microglial inflammatory signaling.⁹²

The amelioration of motor and cognitive measures observed in zQ175 mice treated with CP13 suggests that reduced tau phosphorylation improves neuronal function. We therefore explored whether the antibody treatment restored protein levels of markers characteristic of vulnerable neuronal populations (VGlut1 and DARPP32) and neuronal health (SMI32 and PSD95). However, we did not detect any changes in these neuronal populations, the general pre-synaptic marker PSD95 or the SMI32 marker of axonal damage. To further characterize CP13 neuroprotective properties, future experiments could focus on the effects of CP13 on a larger panel of neuronal sub-populations. Given that behavioral improvements were not observed prior to 3 months of treatment, it is also possible that our study was not long enough to observe specific changes in neuronal populations. Additional work evaluating longer duration treatment beginning prior to 6 months of age could help test this. Another possibility is that our use of mouse brain homogenates and western blotting was not sufficiently sensitive and that subtle effects may have been detected if more cell specific assays had been used. As a result, two questions remain to be addressed to determine the specific impact of CP13 on neurons and whether hyperphosphorylated tau could be targeted in a human-based model of the disease. We therefore prepared human iPSC-derived cortical neurons from control and HD donors. Early characterization of these cells revealed that our differentiation replicated previously reported decreases in neurite length in HD neurons⁷⁵ and additionally displayed an increase in pTau levels. The presence of increased pTau enabled us to treat neurons with CP13 and to demonstrate that treatment with CP13 (but not an IgG control) efficiently reduced pS202 tau levels in HD cortical neurons and rescued neurite retraction (Figure 8). These experiments are a proof of concept that human HD neurons can be targeted by protective antibodies and serve as disease models to study the effects of abnormal tau phosphorylation in HD. Findings in other iPSC-based models of tauopathies further support the relevance and feasibility of such a model.^93,94

Clearance of pathological tau is associated with reduced HTT oligomers and amyloid fibrils

An important aspect of the molecular pathology of HD is the misfolding of mHTT into toxic oligomers and fibrils that further aggregate into large inclusion bodies but also acquire spreading properties suggested to exacerbate disease progression.95, 96, 97 Reducing the cellular load of these small and reactive mHTT aggregates is therefore an attractive strategy to attenuate cellular toxicity. To this end, we evaluated the effects of CP13 injections on soluble and aggregated HTT and mHTT levels in the cortex, hippocampus and striatum of antibody-treated zQ175 mice compared with their WT counterparts. Our results suggest that reduced tau hyperphosphorylation does not correlate with a decrease in soluble HTT or mHTT levels (Figure 5). However, we observed a shift in the protein aggregation profile as a result of CP13 injections, with an increase in small- to medium-sized HTT aggregates and a decrease in small oligomers and amyloid fibrils (Figures 5, 6, and 7). This finding is particularly important because accumulating evidence indicates that the formation of large aggregates may be a protective mechanism through which toxic oligomeric and fibrillar species are sequestered.^65,97, 98, 99, 100, 101 This significant shift in aggregate size in the brains of antibody-treated zQ175 mice suggests that attenuation of tau hyperphosphorylation could help neutralize harmful oligomers and fibrils into less reactive aggregates. The mechanisms involved in this seemingly protective effect are unclear and future studies could concentrate on understanding the biological relationships between tau and HTT and their respective and conjoint roles in supporting cell survival and function. For example, an important study of the HTT interactome network revealed 747 HTT-interacting proteins and a significant enrichment in proteins involved in actin skeleton organization (e.g., MAP1 and MAP2) but also protein transport and molecular chaperones.¹⁰² This study also suggested that HTT interacts with tau in the striatum and cortex of mice at the ages of 2 and 12 months, respectively. It therefore seems reasonable to suggest that, upon mHTT misfolding, a number of processes related to maintenance of the cytoarchitecture and targeted transport for protein degradation begin to malfunction. As a result, cells may increasingly depend on tau to stabilize microtubules, which could trigger rapid conversions between phosphorylated and dephosphorylated states and eventually lead to tau dysfunction and misfolding. If protein degradation pathways become impaired as a result of mHTT and tau pathology,^103,104 damaged proteins would accumulate and aggregate to form protein inclusions as observed in the brains of HD patients.4, 5, 6^,9,10 In this context, one could speculate that preventing tau hyperphosphorylation could stabilize mHTT-induced microtubule dysfunction and favor the neutralization of toxic oligomers through inclusion in amorphous deposits.¹⁰¹ One potential caveat that we did not address in this study is the possibility that such large deposits may become toxic over time. There is still significant controversy surrounding the effects of large aggregates on cell function and some studies have shown that they may interfere with normal cellular functions¹⁰⁵ or disrupt organelle membranes.¹⁰⁶ However, previous work with an HD model known as the shortstop mouse (a transgenic mouse model expressing human HTT truncated at amino acid 117) has demonstrated that large numbers of HTT aggregates can be present in the mouse brain without causing neurodegeneration or behavioral changes, even at 12 months of age.¹⁰⁷ Importantly, cell culture work using the same truncated form of mHTT showed that the shortstop mutation led to rapid conversion of oligomers into larger aggregated forms.¹⁰⁸ Such a process would be consistent with our findings and would suggest that the aggregates are unlikely to induce pathology, even at later time points.

Understanding the interactions between pathological proteins and disease onset and progression is an essential milestone in the search for safe and effective drug targets. Our study provides evidence of the complex relationships between tau and HTT and mHTT, and supports a framework whereby tau hyperphosphorylation is a key contributor to HD pathological features. These observations also suggest that therapeutic strategies, such as immunotherapy, could be leveraged to target both tau and HTT and mHTT and improve motor and/or cognitive functions in patients with HD.

Material and methods

Animals

zQ175 (B6J.129S1-Htt^tm1Mfc/190ChdiJ) and WT littermate control mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) and subsequently maintained in an in-house colony at the Center de Recherche du CHU de Québec (Québec, QC, Canada). ZQ175 mice are a knock-in mouse model where mouse Htt exon 1 is replaced by human HTT exon 1, containing an expansion of approximately 190 CAG repeats, and display an HD-related neurological phenotype. These animals also develop features of tauopathies, including increased tau phosphorylation at serine 202 and 199.¹⁹ All mice were maintained in a temperature-controlled room (∼23°C) with a 12/12 h light/dark cycle and ad libitum access to food and water. Animal handling, injections and all other procedures were completed in accordance with the guidelines of the Canadian Council on Animal Care and were approved by the Comité de Protection des Animaux du CRCHUQ-UL.

Injections

Twenty zQ175 and sixteen WT male littermates received weekly intraperitoneal injections of either CP13 (10 mg/kg; n = 10 zQ175 and n = 7 WT) or saline (Teligent, no. 0195AF01; n = 10 zQ175 and n = 9 WT) beginning at 6 months of age. Injections were administered using 28-gauge syringes (BD Microfine TM IV Needle U-100 Insulin, Thermo Scientific, no. 14-829-1B) in a volume of 100 μL/10 g of body weight. The treatment protocol was conducted over 12 weeks and the injection sites were alternated between the left and right sides of the abdomen. On the day of the injection, an aliquot of frozen antibody stock was defrosted and diluted in sterile saline to reach a final concentration of 1 mg/mL. The addition of more animals to the study was prevented by shutdowns resulting from coronavirus disease 2019 (COVID-19) as well as the passing of Dr. Davies and the subsequent rupture of antibody supply.

Behavioral testing

All behavioral tests were performed during the light phase of the light/dark cycle after 1 h of habituation to the testing room by an experimenter blind to the genotype and treatment received by the animals. To preserve experimental blinding and reduce the level of anxiety experienced by the animals, the behavioral tests and weekly injections were performed by two different experimenters. All tests were performed four times: once before the start of the CP13 injections followed by monthly testing, with the exception of nesting, which was performed only at the last time point as mice need to be individually housed for this assessment. All testing arenas were cleaned with 70% ethanol between animals.

Open field

Four mice were placed individually in one of four adjacent non-porous plastic arenas (30″ × 30″) equipped with a camera overhead. Mice were allowed to freely explore for 60 min without the experimenter in the room. Analysis occurred live during testing using AnyMaze tracking software (v.4.8; Stoelting, Wood Dale, IL). From the tracked trajectories, distance traveled was collected in 5-min bins and was evaluated at minutes 30–35 and 55–60 of testing. The time spent in the corners during the first 5 min of the test was also recorded for each animal. Intrasession habituation was measured by comparing the distance traveled between minutes 30–35 and minutes 55–60, while intersession habituation was calculated by comparing total distance traveled between sessions. Both types of habituation were calculated using an activity change ratio (distance traveled final)/(distance traveled initial + distance traveled final), with a score of 0.5 indicating no habituation.³⁷ The number of fecal pellets were counted after removing the animals from the arena.¹⁰⁹ The first 30 min in the open field were excluded from analysis, as animals displayed indications of increased anxiety during this time period.

Y maze

Testing for the Y maze was completed between 8 a.m. and 12 p.m. for all time points. The Y maze apparatus is made of clear acrylic covered by black laminated paper and composed of three equal arms measuring 32.5 cm in length, 8.5 cm in width and 15 cm in height. The walls of the maze are identical and opaque to prevent animals from using visual or spatial clues. Mice were initially placed at the center of the maze and left to explore for 7 min with the experimenter remaining in the room. All trials were filmed and the recorded performances were evaluated by a blinded experimenter after completion of the test. A correct alternation was defined as the successive entrance into each of the three arms, in any order, without re-entering one of the arms. The percentage of correct alternations was calculated as the number of alternations divided by the total number of entries minus two. The aim of this test is to evaluate short-term spatial working memory.¹¹⁰

Nesting

Two to three days after the last injection, the animals were transferred to clean individual testing cages with a nestlet. The next morning, the quality of the nests was scored based on a previously published rating scale from 0 to 4.¹¹¹ A score of 0 was given to a nestlet that was left untouched (more than 95% intact). A score of 1 was given to a nestlet that was only slightly shredded (50%–95% intact). A score of 2 was given if greater than 50% was shredded but no discernible nest had been formed. A score of 3 was given when up to 90% of the nestlet had been shredded but the nest was flat. A score of 4 denoted a near-perfect nest where nearly 100% had been shredded and the mouse had formed high walls from the torn pieces.¹¹¹ The nestlet was then removed, replaced and scored 1 h later. Nest building is a response to the evolutionary need for heat conservation, shelter and reproduction, and its assessment provides a general indication of well-being.¹¹¹

Post mortem analyses

Tissue processing

At the end of the protocol, mice were sacrificed using two different methods: (1) lethal injection with ketamine/xylazine with 1% ketamine (30 mg/kg) and xylazine hydrochloride (4 mg/kg) followed by phosphate-buffered saline (PBS) perfusion (n = 4 WT saline, n = 3 WT CP13, n = 4 zQ175 saline, and n = 4 zQ175 CP13) or (2) decapitation (n = 5 WT saline, n = 4 WT CP13, n = 6 zQ175 saline, and n = 6 zQ175 CP13). The PBS perfusion protocol served for immunofluorescence and immunohistochemistry, while decapitation was performed for experiments measuring tau pathology, as earlier reports have suggested that anesthetics can affect the phosphorylation state of tau.¹¹²

Tissue processing after perfusion

The brains were collected, placed on ice and cut to separate the right and left hemispheres. One hemisphere was dissected on ice to isolate the striatum, hippocampus, cerebellum, cortex and frontal cortex. The other hemisphere was post-fixed in 4% paraformaldehyde solution (PFA) (Electron Microscopy Sciences, no. 19210) overnight and subsequently stored in 20% sucrose (Sigma, no. S9378-5KG) in PBS at 4°C for cryoprotection. The tissue was processed within 1 week of sacrifice by cutting 25-μm-thick coronal sections using a sliding microtome (Leica Microsystems, no. SM 2000R) and then collecting and pooling the sections in the wells of a 48-well plate (Falcon, no. 353078) filled with anti-freeze solution. The 48-well plates were stored at −20°C until use.

Tissue processing after decapitation

The brains were immediately removed and dissected on ice to isolate the structures described above that were frozen on dry ice prior to storage at −80°C. The tissue isolated from the right hemisphere was homogenized in a five times v/w ratio of radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl, pH 7.4 [Sigma, no. T3253-1KG], 1% NP-40 [Sigma, no. I8896-50ML], 150 mM NaCl [Sigma, no. S7653-5KG], 0.25% Na-deoxycholate [EMD, no. SX0480-2], 1 mM EDTA [Fisher BioReagent, no. BP282-500] and 10 μL/mL of Halt Proteases Inhibitor Cocktail [Thermo Scientific, no. 1861281]) on ice with a Dounce homogenizer potter. Samples were then centrifuged (Sorvall mX150 + micro ultracentrifuge, Thermo Scientific) for 20 min at 20,000 × g at 4°C. The supernatant was recovered and stored at −80°C. The dissected brain structures from the left hemisphere were homogenized in eight volumes of lysis buffer (150 mM NaCl [Sigma, no. S7653-5KG], 10 mM NaH₂PO₄ [Sigma, no. S0751-1KG], 1% Triton X-100 [Sigma-Aldrich, no. T8787], 0.5% sodium dodecyl sulfate (SDS) [Sigma, no. L3771-1KG], and 0.5% deoxycholate) containing 10 μL/mL Halt Proteases Inhibitor Cocktail. After homogenization, all lysates were briefly sonicated (3 times for 10 s) and centrifuged (Sorvall mX150 + micro ultracentrifuge, Thermo Scientific) at 100,000 × g for 20 min at 4°C. The supernatant was withdrawn and stored at −80°C, while the pellet obtained was homogenized on ice with a Dounce homogenizer potter in RIPA solution, sonicated (3 times for 10 s) and passed through a 28G needle (BD Microfine TM IV Needle U-100 Insulin, Thermo Scientific, no. 14-829-1B) five times. The homogenized insoluble fraction was then stored at −80°C.

Total protein in the soluble fraction obtained from both right and left hemisphere structures as well as from the pellet derived from the processing of the left hemisphere was quantified using a BCA Protein Assay Kit (Thermo Fisher Scientific, no. 23225).

Preparation and treatment of iPSC-derived cortical neurons

Human iPSCs produced from patients with HD and age- and sex-matched healthy donors were obtained from the Coriell Institute for Medical Research via the National Institute of Neurological Disorders and Stroke (NINDS) repository (control lines: ND 41658, male, 17/18 HTT CAG repeats; ND 38554, female, 17 CAG repeats and HD lines: ND 36999, male, 180 HTT CAG repeats; ND 36998, female, 60 HTT CAG repeats). The cells were routinely cultured in mTesR E8 media (STEMCELL Technologies, Canada) and regularly passaged using Versene (Thermo Fisher Scientific, no. 15040066) on Matrigel-coated 35-mm plates (Corning, no. 354277). To initiate the differentiation of iPSCs into cortical neurons, cells were harvested with Versene and plated in E8 media supplemented with 10 μM Y-27632 (day −1). The next day, iPSCs reached 100% confluence and media was changed to neural induction media (NIM) consisting of neural maintenance media (NMM) (50% v/v DMEM/F12 plus Glutamax, 50% v/v Neurobasal media, 2.5 μg/mL insulin, 100 μM 2-mercaptoethanol, 0.5× non-essential amino acids, 500 μM sodium pyruvate, 0.5× N2, 0.5× B27, and 1 mM L-glutamine) supplemented with 10 μM SB431542 (Tocris Bioscience, no. 1614) and 1 μM LDN-193189 (STEMCELL Technologies, no. 72147; day 0). NIM was replaced daily until day 12, at which time cells were harvested in Accutase, centrifuged at 1,000 rpm for 5 min, resuspended in NIM supplemented with 10 μM Y-27632 and replated at a 1:2 ratio on Matrigel-coated plates. At day 13, cells were washed with PBS and media was changed to NMM supplemented with 20 ng/mL basic fibroblast growth factor (bFGF) (STEMCELL Technologies, no. 78003.1) and replaced daily until the formation of neural rosettes around day 20. When rosettes became visible, cells were considered to be at the neural progenitor stage and were harvested in Accutase and passaged at a 1:2.5 ratio in NMM, on Matrigel-coated dishes. Media was changed with fresh NMM every 48 h and neurons were matured until at least day 35. iPSC-derived neurons were then harvested in Accutase and replated on Matrigel-coated dishes for subsequent experiments. The studied phenotypes were reproducible between 36998 and 36999 iPSC-derived neurons and therefore data collected using both lines were pooled.

To determine the effect of CP13 on iPSC-derived neurons, cells were plated on 24-well plates at a density of 300,000 to 500,000 cells/well. Forty-eight hours after plating, iPSC-derived neurons were treated for a total of 72 h with 15 μg/mL CP13 or mouse IgG2b (R&D Systems, no. MAB0041) antibodies diluted in NMM. At the end of the treatment, neurons were washed once with PBS and harvested in 50 μL Laemmli buffer supplemented with 1× Halt phosphatase and protease inhibitor cocktail. The samples were heated at 70°C for 10 min, sonicated 3 times for 5 s and processed for western blot, as described in the next section. In addition, we used an immunofluorescence approach to assess the proportion of glutamatergic neurons produced by control and HD iPSCs. To this end, we plated and cultured iPSC-derived neurons on Matrigel-coated coverslips in NMM, fixed the cultures in 4% PFA for 20 min and processed the samples for immunofluorescence as described in the “Immunofluorescence” section below. The western blot experiments were performed on two independent control iPSC lines (17/18 HTT CAG repeats) and one HD iPSC line (180 HTT CAG repeats). Immunofluorescence was performed on two control iPSC lines (17/18 HTT CAG repeats) and both HD lines (180 and 60 HTT CAG repeats).

Western blotting

Cortical, hippocampal and striatal homogenates containing 10 μg protein were mixed with 1% (v/v) Laemmli buffer (312.5 mM Tris-HCl, 30% glycerol [Sigma, no. G5516-4L], 12,5% β-mercaptoethanol [Sigma, no. M3148-100ML], 10% SDS, 0.025 M EDTA and 0.01% Bromophenol blue [Sigma, no. B0126-25G]), heated at 70°C for 10 min and separated on an SDS 6% or 10% polyacrylamide gel for 1 h and 15 min at 110 V in migration buffer (25 mM Tris HCl, 190 mM glycine [Sigma, no. G7126-5KG], and 0.1% SDS). The gel was then transferred onto a 0.2-μm nitrocellulose membrane (Bio-Rad, no. 1620112) in transfer buffer (25 mM Tris HCl, 190 mM glycine, 20% methanol [Fisher Chemical, no. A452-4]) at 20 V and 4°C overnight followed by a boost for 20 min at 100 V. After completion of the transfer, total protein was detected using a REVERT total protein stain (LI-COR Biosciences, no. 926-11011) according to manufacturer's instructions. After reversal of total protein stain, non-specific binding sites were blocked with 3% gelatin extracted from cold water fish skin (Sigma, no. G7041-500G) in PBS for 1 h at room temperature (RT), followed by overnight incubation at 4°C with the following primary antibodies: total tau (1:10,000, Dako, no. A0024); phosphorylated tau (1:1,000 pS202 and 1:1,000 PHF1 provided by the late Peter Davies; CP13 1:1,000 and 1:5,000 pS199, Invitrogen, no. 44734G); VGlut1 (1:1,000, NeuroMAB, no. 75-066); PSD95 (1:1,000, NeuroMab, no. 75028); GSK3β (1:1,000, BD Transduction Laboratories, no. 610202); PP2B (1:1,000, Cell Signaling Technology, no. 2614); GFAP (1:1,000, Sigma-Aldrich, no. G3893); DARPP32 (1:1,000, Cell Signaling Technology, no. 2306); and SMI-32 (1:1,000, BioLegend, no. SMI-32P) diluted in 3% fish gelatin in PBS supplemented with 0.1% Tween 20 (PBST) (Fisher Bioreagent, no. BP337-500). The following day, the membranes were washed three times for 10 min, incubated for 45 min at RT with IRDye 800CW (LI-COR Biotechnology, no. 926-32212) or IRDye 680RD (LI-COR Biosciences, no. 926-68073) antibodies and quantified using Odyssey CLx imaging system (LI-COR Biosciences).

Detection of Iba1

A total of 20 μg of protein was mixed with 1% (v/v) Laemmli buffer and heated for 5 min at 95°C. Samples were migrated on a 12% SDS-PAGE for 1.5 h and proteins transferred to a 0.2 μm nitrocellulose membrane overnight at 20 V, followed by a boost for 20 min at 100 V. Following the boost, total protein was assessed as described above. The membrane was then blocked with 2.5% (w/v) BSA (Bioshop, no. ALB001) in PBS for 1 h and incubated overnight at 4°C with primary antibody Iba1 (1:500, Wako, no. 016-20001) diluted in blocking solution (2.5% [w/v] BSA in PBST). The following day, membranes were washed three times in PBS and incubated for 45 min at RT in secondary antibody IRDye 800CW (LI-COR Biosciences, no. 926-32212) diluted in blocking solution. After three washes in PBST, the membranes were revealed and analyzed using Odyssey CLx imaging system (LI-COR Biosciences).

Detection of HTT

A total of 50 μg of protein was diluted in 1% (v/v) Laemmli buffer and heated at 60°C for 5 min, followed by separation on a 6% SDS-polyacrylamide gel for 3 h at 110 V and subsequent overnight transfer on a low fluorescence polyvinylidene difluoride (PVDF) membrane (Immobilon-FL Merck-Millipore no. IPFL00010) at 20 V in transfer buffer containing 0.01% (w/v) SDS and 16% (v/v) methanol. REVERT total protein stain was performed as described above. The membranes were then blocked with 2.5% (w/v) BSA prepared in tris-buffered saline (TBS) and sequentially incubated overnight at 4°C with anti-HTT (1:1,000, Merck-Millipore, no. MAB2166) and anti-polyglutamine expansion (1:500, Merck-Millipore, no. MAB1574) primary antibodies diluted in TBS containing 0.1% (v/v) Tween 20 (TBST). The membranes were washed three times and incubated with IRDye 800CW (LI-COR Biosciences, no. 926-32212) secondary antibody for 45 min at RT. Membrane revelation and quantification were performed as detailed above.

Filter retardation assay

The insoluble fractions extracted from cortical, hippocampal and striatal brain tissue were passed five times through a 28G needle. Cortical (20 μg), hippocampal (10 μg) and striatal (10 μg) homogenates were then diluted in PBS to make up a volume of 70 μL, to which 30 μL of SDS:DTT (2% [v/v] SDS and 100 mM DTT, Invitrogen, no. 28025-013) was subsequently added.⁶⁸ Samples were boiled at 100°C for 10 min, cooled to RT and passed through a 0.2-μm cellulose acetate membrane (Steriltech, no. 1480025) using a vacuum-filtration system (HYBRI DOT Manifold, BRL Bethesda Research Laboratories, no. 1050 MM). Each sample was processed in triplicates to ensure the validity of the results. The membrane was then washed twice with 0.1% (w/v) SDS in PBS and rinsed once with PBS to remove excess SDS. The membrane was subsequently blocked with 5% (w/v) BSA in PBS for 1 h at RT prior to overnight incubation at 4°C with the following antibodies: anti-HTT (1:500, Millipore Sigma, no. MAB5374), anti-total HTT N18 (1:2,500, kindly provided by Dr. Ray Truant, McMaster University, Canada) and anti-polyglutamine expansion (1:500, Millipore Sigma, no. MAB1574) diluted in 2.5% (w/v) BSA in PBST. The membranes were washed three times in PBST for 5 min, incubated in IRDye 800CW (LI-COR Biosciences, no. 926-32212) and IRDye 680RD (LI-COR Biosciences, no. 926-68073) secondary antibodies for 45 min, and visualized using the Odyssey CLx imaging system.

Dot blot

Levels of oligomeric and fibrillar protein species in brain homogenates were assessed using a dot blot assay. A total of 5 μg of brain lysates were spotted onto nitrocellulose membranes in triplicate and left to dry for 20 min. Dry membranes were washed twice for 10 min in PBS and once for 10 min in PBST before blocking with 2.5% (w/v) BSA in PBS for 1 h. Membranes were then incubated overnight at 4°C with primary antibodies oligomer A11 (1:1,000, Invitrogen, no. AHB0052) or anti-amyloid fibrils OC (1:1,000, Sigma, no. AB2286) diluted in PBST supplemented with 2.5% (w/v) BSA. The next day, membranes were washed three times for 10 min in PBST, followed by one wash in PBS and they were subsequently incubated with IRDye 680RD secondary antibody (LI-COR Biosciences) for 1 h at RT. Finally, the membranes were washed two times for 10 min in PBST followed by one wash in PBS and visualized using the Odyssey CLx imaging system.

Semi-denaturing detergent agarose gel electrophoresis (SDD-AGE)

Brain lysates were sonicated once for 10 s and a total of 30 μg of protein was diluted in sample buffer (20% [v/v] glycerol, 0.01% [v/v] bromophenol blue and 0.08% [w/v] SDS in 2× Tris-acetate buffer), followed by separation on an agarose gel (1.8% [w/v] agarose, 0.02% [w/v] SDS, 0.24% [w/v] Tris base and 1.5% [w/v] glycine in water) for 2 h at 110 V on ice and subsequent transfer on a nitrocellulose membrane using the Trans-Blot Turbo system (Bio-Rad). The membrane was then heated in PBS to 30°C, allowed to return to RT, and incubated for 40 min in blocking buffer (3% [w/v] fish gelatin in PBS). The membrane was then incubated overnight at 4°C in anti-HTT primary antibody (1:1,000, Merck-Millipore, no. MAB5492) prepared in TBST containing 2.5% (w/v) BSA, followed by incubation in IRDye secondary antibody (LI-COR Biosciences) for 1 h at RT. Finally, the membranes were visualized using the Odyssey CLx imaging system.

Sandwich ELISA

ELISA microplates (Invitrogen, cat no. 44-2404-21) were first coated with 100 μL capture antibody (anti-tau, 1:1,000, Dako, cat no. A0024) and incubated overnight at RT. The next day, plates were washed three times using 1× wash buffer (Thermo Fisher Scientific, cat no. 28352), incubated with 300 μL of blocking buffer (1% [w/v] BSA in Dulbecco’s PBS [DPBS]) for 1 h and washed again three times using 1× wash buffer. Each sample was added in technical duplicates and consisted of 100 μL brain homogenate prepared from perfused mice (258 μg/mL diluted in blocking buffer) and the plate was incubated overnight on a rocking platform at 4°C. Plates were subsequently washed three times using wash buffer, incubated with 100 μL horseradish peroxidase (HRP)-conjugated goat anti-mouse secondary antibody (1:4,000, Jackson ImmunoResearch, cat no. 115-035-166) for 45 min at RT and washed again three times using 1× wash buffer. Lastly, 100 μL 3,3′, 5,5’-tetramethylbenzidine (TMB) chromogen solution (G-Bioscience, cat no. 00-4201-56) was added to the plate and incubated for 4 min and reaction was stopped by adding 50 μL of 0.18 M sulfuric acid. Absorbance was read using a Synergy microplate reader (BioTek Instruments, Winooski, VT, USA) set at 450 and 540 nm. Brain homogenates prepared from tau knockout animals were used as negative controls, while WT and zQ175 brain homogenates supplemented with CP13 during the overnight incubation step served as positive controls. In addition, wells coated with capture antibody and incubated with HRP-conjugated goat anti-mouse secondary antibody only (no brain samples) served as blank values. To analyze the data, wavelength correction was first performed by subtracting readings at 540 nm from the readings at 450 nm, and absorbance readings were then corrected using the blank values.

Immunofluorescence

Free-floating sections covering the entire brain were washed three times in potassium phosphate-buffered saline (KPBS), incubated in 3% (v/v) H₂O₂ for 30 min, washed three times with KPBS, blocked in blocking buffer (10% [v/v] donkey serum [Sigma-Aldrich, no. D9663], 0.01% [v/v] Triton X-100 [Sigma-Aldrich, no. T8787] and 0.5% [w/v] BSA) and incubated overnight at 4°C with anti-HTT antibody (1:1,000, Millipore, no. MAB2170). The next day, the sections were washed three times with KPBS, and incubated with Alexa Fluor 546 antibody (1:500, Thermo Fisher Scientific, no. A10036) diluted in blocking buffer for 2 h at RT. The sections were washed, incubated for 7 min with DAPI nuclear stain (Molecular Probes, no. D3571) diluted in KPBS, washed again and mounted using Fluoromount G mounting medium (Thermo Fisher Scientific, no. 00-4958-02). All sections were stored at 4°C to preserve the fluorescent signal.

iPSC-derived neurons were fixed in 4% PFA for 20 min at RT, washed once with PBS, and incubated in blocking/permeabilizing buffer (0.3% [v/v] TX-100, 1% [w/v] BSA, and 10% [v/v] fetal bovine serum [FBS] in PBS) for 1 h at RT. The neurons were washed once with PBS and incubated overnight at 4°C with the following primary antibodies: MAP2 (1:1,000, LS Biosciences, no. LS-B290) and VGlut1 (1:500, Sigma, no. V0389) diluted in PBS with 0.1% (w/v) BSA. The following day, the coverslips were washed twice with PBS and incubated with Alexa-conjugated secondary antibodies (Thermo Fisher Scientific) diluted at 1:1,000 in PBS with 0.1% (w/v) BSA for a total of 1 h at RT and in the dark. The neurons were subsequently washed once with PBS and incubated in a DAPI nuclear stain (Molecular Probes) for 10 min at RT and in the dark. At the end of the incubation, the cells were washed two times and mounted on slides using ProLong antifade or Fluoromount G mounting medium (Thermo Fisher Scientific) and stored at 4°C until analysis.

Immunohistochemistry

Free-floating sections representative of the entire brain were washed three times in 0.1 M PBS, incubated in 3% (v/v) H₂O₂ for 30 min, washed three times and blocked in blocking buffer (5% [v/v] normal goat serum [NGS] [Wisent Bioproducts, no. 053-150] and 0.1% [v/v] Triton X-100 in PBS) and incubated overnight at 4°C with the following primary antibodies: anti-mHTT (1:500, Millipore Sigma, no. MAB5374); anti-GFAP (1:1,000, Dakocytomation, no. Z0334); and anti-Iba1 (1:1,000, Wako Pure Chemicals Industries, no. 019-19741). The next day, samples were washed three times and incubated with secondary antibodies goat anti-mouse IgG biotinylated (1:500, Vector Laboratories, no. BA9200) or goat anti-rabbit IgG biotinylated (1:500, Vector Laboratories, no. BA1000) diluted in blocking buffer for 1 h at RT. Three washes with 0.1 M PBS were performed before and after the sections were incubated in ABC-HRP complex solution (Vector Laboratories, no. PK-6100). An additional step was performed for mHTT staining only and consisted of staining with Nickel DAB (DAB substrate kit, peroxide [with nickel], Vector Laboratories, no. SK-4100) according to the manufacturer's protocol. The free-floating sections were then washed and dehydrated in ascending grades of ethanol, cleaned in Citrisolv (Decon Laboratories, no. 1601) and mounted in DPX mounting media (Electron Microscopy Sciences, no. 13512).

Image acquisition and quantification

Fluorescent photomicrographs were acquired using an AXI0 Imager Z.2 upright confocal microscope equipped with a plan-apochromatic 20× objective lens (Zeiss, numerical aperture [NA] = 0.8) and Zen software (Zeiss, Oberkochen, Germany). All images were prepared using Fiji¹¹³ or Photoshop CS5 (Adobe, San Jose, CA), and brightness and contrast were adjusted, when necessary, to improve image visualization.

Quantification of the number and size of mHTT aggregates

Stereology was performed using the Optical Fractionator probe (Stereo Investigator, Microbrightfield) installed on an E800 Nikon microscope in order to count and measure the size of mHTT-positive aggregates in the striatum, cortex, and hippocampus. Perimeters of the brain structures were outlined under a 4× objective lens, using the tracing contour option. Counts were performed by a blinded investigator under a 60× objective lens within the defined perimeter every 10^th section, in a series of EM48-stained sections. For all three structures, six sections were counted. For the striatum and cortex, these sections spanned bregma 1.54 to bregma −0.1, while for the hippocampus they spanned bregma −1.06 to −2.56.¹¹⁴ The counting frame was set to 15 × 15 and the grid layout to 300 × 300. Aggregates for which the size was too small to be accurately measured were placed in the “very small” aggregate category. All the other aggregates were divided in four groups according to their size: very small (0–0.5 μm³); small (0.5–2 μm³); medium (2–4 μm³) and large (4–6 μm³).

Quantification of Iba1-positive microglia and GFAP-positive astrocytes

An Optical Fractionator installed onto an E800 Nikon microscope was used to count the number of Iba1-positive microglia and GFAP-positive astrocytes. The contours were traced at 4× and the counting was performed using a 20× objective. The counting frame was set at 200 × 200 μm and the sampling grid at 500 × 500 μm.

Quantification of neurite length

Neurite length of iPSC-derived cortical neurons was measured after 14 days in culture in the absence or presence of 15 μg/mL IgG control or CP13 using the Neurite Tracer plugin available in Fiji. Data were collected from two biological replicates and more than 200 neurites were measured in each experimental condition.

Statistical analysis

Statistical analyses were performed using Prism 7 for Mac OS X (La Jolla, CA). Experiments consisting of two groups were assessed using two-tailed unpaired Student's t test assuming equal standard variation. Multiple groups (i.e., mouse genotype versus treatment) were analyzed by two-way ANOVA followed by Sidak's post-hoc test or one-way ANOVA followed by Dunnett's post hoc tests for cell culture experiments. All data are expressed as mean ± SEM with individual data points shown for each graph. For all experiments, a significance cutoff of 0.05 was used. Outliers were calculated for each measure of behavior separately and were defined as data points that exceeded the 1.5 interquartile ranges below the first quartile or 1.5 interquartile ranges above the third quartile. Identification of outliers revealed little to no overlap of animals between tests. For cell culture experiments, outliers were identified using the Grubbs test (alpha = 0.05). Outliers were excluded from all statistical analyses.