Huntington disease (HD) is a fatal autosomal dominant neurodegenerative disease that affects approximately 1–2 individuals per 10,000 (Fisher and Hayden, 2014). HD is caused by an expansion of the CAG repeat in the first exon of the gene that encodes the huntingtin protein. The presence of the resultant mutant huntingtin protein, and perhaps in combination with reduced expression levels of non-expanded (i.e., non-pathogenic) huntingtin (Cattaneo et al., 2005; Dietrich et al., 2017), results in the progressive deterioration of much of the neuroaxis, with particularly devastating effects observed in the striatum. The symptoms of HD are described as a clinical triad that includes progressive motor and cognitive decline as well as psychiatric disturbances. The monogenic nature of HD has resulted in the identification of presymptomatic gene carriers (pre-HD); from this pre-HD population, it was discovered that subtle impairments in cognitive function are often evident as many as 15 years prior to a formal diagnosis based on unequivocal motor signs (Paulsen et al., 2008). In fact, in a population of pre-HD individuals predicted to be at least 14 years away from a formal diagnosis of HD, approximately 40% reached the criteria for mild cognitive impairment, with higher rates observed in those closer to a motor diagnosis (Duff et al., 2010). Deficits in episodic memory, processing speed, executive function and visuospatial perception were all found to be associated with pre-HD (Duff et al., 2010). While the bulk of preclinical HD research focuses on striatal dysfunction and motor impairment in HD animal models, it is well-accepted that extrastriatal dysfunction can make key contributions to the non-motor symptoms of HD (Ruocco et al., 2006; Ross et al., 2014). Of particular interest to the present study is HD-associated pathophysiology in the hippocampus, a brain region critically involved in cognitive function.

Prior to neurodegeneration, impairments in synaptic structure and function are readily detected in the striatum in HD mouse models (Milnerwood and Raymond, 2010; Raymond et al., 2011). Like Alzheimer disease (Selkoe, 2002), the early stages of HD have been associated with numerous “synaptic failures” (Selkoe, 2002) that are thought to promote subsequent neurodegeneration (Li et al., 2003; Okamoto et al., 2009; Milnerwood and Raymond, 2010; Milnerwood et al., 2010; Raymond et al., 2011; Parsons and Raymond, 2014). Various forms of synaptic plasticity, the cellular underpinnings of learning and memory, are impaired in mouse models of HD (Usdin et al., 1999; Murphy et al., 2000; Cummings et al., 2006, 2007; Milnerwood et al., 2006; Lynch et al., 2007; Simmons et al., 2009; Brito et al., 2014; Kolodziejczyk et al., 2014; Giralt et al., 2017; Sepers et al., 2017; Zhang et al., 2018). N-methyl-D-aspartate receptor (NMDAR)-dependent long-term potentiation (LTP) at the Schaffer collateral-CA1 synapse in the hippocampus is well-accepted to underlie cognitive processes such as learning and memory (Bliss and Collingridge, 1993; Bruel-Jungerman et al., 2007a,b; Citri and Malenka, 2008). While this particular form of LTP has previously been shown to be impaired in mouse models of HD, many of these studies were conducted at a symptomatic age in N-terminal fragment (Murphy et al., 2000; Anglada-Huguet et al., 2014; Giralt et al., 2017), overexpression (Hodgson et al., 1999) or homozygous knock-in HD mouse models (Lynch et al., 2007; Simmons et al., 2009); thus, while it is well-understood that HD is associated with poor synaptic plasticity, the onset and progression of plasticity impairments are poorly documented. As cognitive decline can emerge in the HD prodrome, it is important to understand what forms of synaptic plasticity, if any, are perturbed at presymptomatic ages in preclinical disease models. In the present study, we provide the most thorough characterization to date of hippocampal plasticity deficits in a mouse model of HD. Q175FDN mice express an expanded CAG tract in the gene encoding the huntingtin protein, and express higher levels of the mutant huntingtin protein than previous knock-in models of HD (Pouladi et al., 2013; Southwell et al., 2016). Altered electrophysiological properties, including reduced excitatory postsynaptic currents, have been detected in the striatum from both heterozygous and homozygous Q175FDN mice (Southwell et al., 2016); to the best of our knowledge, hippocampal synaptic plasticity has yet to be investigated in Q175FDN mice. We chose the novel Q175FDN knock-in mouse model for the present study due to its high expression level of mutant huntingtin and clear HD-like phenotype in both heterozygous and homozygous mice. Our results demonstrate clear deficits in both short- and long-term synaptic plasticity that exhibit differential sensitivities to disease progression, plasticity induction paradigm as well as mutant huntingtin expression levels and/or the loss of non-pathogenic huntingtin.

Here, we sought to thoroughly characterize the onset and progression of synaptic plasticity deficits in a slowly progressing mouse model of HD. We focused on the novel Q175FDN HD model, a knock-in model in which a clear HD-like phenotype begins to develop around 6–9 months of age in heterozygous mice due to high expression levels of the mutant huntingtin protein. In Q175FDN homozygous mice, which completely lack non-pathogenic huntingtin and are therefore less relevant to the large majority of human HD patients, the HD-like phenotype appears earlier and is more severe compared to that observed in Q175FDN heterozygous mice (Southwell et al., 2016). In the present study, experiments were performed at different ages and data were separated into the following three age groups: 3 months (101.5 ± 1.8 days), 6–9 months (215.8 ± 8.4 days), and 12 months (367.9 ± 2.4 days). For Q175FDN heterozygous mice, these ages represent presymptomatic, early symptomatic and late symptomatic disease stages, respectively, while the HD-like phenotype begins to emerge in Q175FDN homozygous mice as early as 3 months (Southwell et al., 2016).

First, we examined basal synaptic connectivity at the Schaffer collateral-CA1 synapse in acute brain slices obtained from WT, Q175FDN heterozygous and Q175FDN homozygous mice at 3, 6–9 or 12 months of age. At 3 months, input–output curves revealed that the slope of the fEPSP response increased to stimulus pulses of increasing intensity as expected, but to the same extent in all mice and no significant genotype or interaction effects were observed (Figures 1A,B, WT n = 10, heterozygous n = 24, homozygous n = 8; repeated-measures two-way ANOVA, p_(genotype) = 0.280, p_(intensity) < 0.001, p_{(interaction)} = 0.484). At 6–9 months, we found that at higher stimulus intensities, mean fEPSP slopes were elevated in heterozygous and homozygous mice compared to WTs, and a significant interaction effect was observed (Figures 1A,C, WT n = 23, heterozygous n = 30, homozygous n = 22; repeated-measures two-way ANOVA, p_(genotype) = 0.051, p_(intensity) < 0.001, p_{(interaction)} = 0.031). At 12 months, we also observed larger fEPSP basal responses in heterozygous mice compared to WT mice (Figures 1A,D, WT n = 7, heterozygous n = 9, repeated-measures two-way ANOVA, p_(genotype) = 0.047, p_(intensity) < 0.001, p_{(interaction)} < 0.001). Homozygous mice were not included in this 12 month cohort, as their survival curves begin to drop sharply after approximately 9 months of age (Southwell et al., 2016).

As we observed significant interaction and genotype effects in the older cohorts, we sought to determine how age affects fEPSP responses within a given genotype. WT fEPSP response slopes significantly diminished with age (Figures 1A,E, 3 months n = 10, 6–9 months n = 23, 12 months n = 7, repeated-measures two-way ANOVA, p_(age) < 0.001, p_(intensity) < 0.001, p_{(interaction)} < 0.001), in agreement with a previous study on the effects of age on Schaffer collateral-CA1 basal connectivity (Weber et al., 2015). Surprisingly, the effect of age on fEPSP slope was not observed in either heterozygous (Figures 1A,F, 3 months n = 24, 6–9 months n = 30, 12 months n = 9, repeated-measures two-way ANOVA, p_(age) = 0.711, p_(intensity) < 0.001, p_{(interaction)} = 0.534) or homozygous mice (Figures 1A,G, 3 months n = 8, 6–9 months n = 22, repeated-measures two-way ANOVA, p_(age) = 0.984, p_(intensity) < 0.001, p_{(interaction)} = 0.890). Therefore, the significant differences observed between genotypes appears to result from an abnormal maintenance of strong basal connectivity with age in heterozygous and homozygous mice that is not seen in WT mice.

Next, we assessed presynaptic release probability through standard paired-pulse ratio measurements (Figure 2A). At 3 months of age, paired-pulse ratios decreased with increasing inter-pulse intervals as expected, although no genotype differences were observed (Figure 2B, WT n = 10, heterozygous n = 24, homozygous n = 8, repeated-measures two-way ANOVA, p_(genotype) = 0.280, p_(interval) < 0.001, p_{(interaction)} = 0.484). Similar results showing no significant genotype or interaction effects were obtained for the 6–9 month cohort (Figure 2C, WT n = 18, heterozygous n = 29, homozygous n = 21, repeated-measures two-way ANOVA, p_(genotype) = 0.227, p_(interval) < 0.001, p_{(interaction)} = 0.441) and the 12 month cohort (Figure 2D, WT n = 13, heterozygous n = 7, repeated-measures two-way ANOVA, p_(genotype) = 0.281, p_(interval) = 0.339, p_{(interaction)} = 0.664). As previous research demonstrates a significant paired-pulse ratio difference at short intervals in HD mice (Kolodziejczyk et al., 2014), we isolated the 50 ms interval data and compared the genotypes within each age group with a one-way ANOVA. There were no genotype differences at 3 months of age (Figure 2E, WT n = 10, heterozygous n = 24, homozygous n = 8, one-way ANOVA, p = 0.657), and while mean paired-pulse ratios tended to be lower in homozygous mice at 6–9 months (Figure 2F, WT n = 18, heterozygous n = 29, homozygous n = 21, one-way ANOVA, p = 0.107) and in heterozygous mice at 12 months (Figure 2G, WT n = 13, heterozygous n = 7, unpaired t-test, p = 0.178), these results were not statistically significant. Thus, we conclude that at the Shaffer collateral-CA1 synapse, presynaptic release probability is minimally affected, if at all, in both pre- and postsymptomatic ages in heterozygous and homozygous Q175FDN mice.

Next, we assessed another form of presynaptic short-term plasticity, PTP, as well as long-term postsynaptic plasticity at Shaffer collateral-CA1 synapses in WT, heterozygous and homozygous Q175FDN mice at the aforementioned age groups. In these experiments, plasticity was induced by TBS consisting of 10 bursts of 4 pulses separated by an inter-burst interval of 200 ms. At 3 months, we found no significant difference in PTP induced by TBS (herein referred to as TBS–PTP), quantified by averaging the fEPSP slope for the first 3 min after TBS and expressing this average as a percentage of the baseline fEPSP slope prior to TBS (Figures 3A,B, WT n = 5, heterozygous n = 13, homozygous n = 7, one-way ANOVA, p = 0.220). While short-term PTP was intact, we found that LTP induced by TBS (herein referred to as TBS–LTP), quantified 55–60 min post-TBS, was highly variable in heterozygous mice while mean responses tended to be lower in homozygous mice. At this age, we observed a significant effect of genotype on LTP magnitude (Figures 3A,C, WT n = 5, heterozygous n = 13, homozygous n = 7, one-way ANOVA, p = 0.042), although no significance difference between any two particular genotypes was detected with post hoc comparisons (Tukey post hoc p > 0.05 for all comparisons).

At 6–9 months, TBS–PTP was significantly affected by genotype (Figures 3D,E, WT n = 7, heterozygous n = 15, homozygous n = 19, p < 0.001). Mean TBS–PTP values were highest for WT mice and lowest for homozygous mice, with intermediate levels observed in heterozygous mice (Tukey post hoc: WT vs. heterozygous p > 0.05; WT vs. homozygous p < 0.001; heterozygous vs. homozygous p < 0.01). A similar pattern was observed for TBS–LTP, with LTP strength being highest in WT mice, lowest in homozygous Q175FDN mice and intermediate in heterozygous Q175FDN mice (Figures 3D,F, WT n = 7, heterozygous n = 15, homozygous n = 19, one-way ANOVA, p < 0.001, Tukey post hoc: WT vs. heterozygous p > 0.05; WT vs. homozygous p < 0.001; heterozygous vs. homozygous p < 0.01). These data clearly demonstrate a greater impairment of both short- and long-term plasticity in homozygous mice compared to heterozygous mice.

As before, only WT and heterozygous mice were tested at 12 months of age, as a result of the short lifespan in homozygous mice (Southwell et al., 2016). PTP was significantly impaired in 12 month old heterozygous mice compared to age-matched WT littermates (Figures 3G,H, WT n = 8, heterozygous n = 7, unpaired t-test, p = 0.012). While LTP tended to be reduced in 12 month HETs, this result did not quite reach statistical significance (Figures 3G,I, WT n = 8, heterozygous n = 7, unpaired t-test, p = 0.076).

To further clarify how disease progression affects short- and long-term synaptic plasticity, we correlated the magnitude of TBS–PTP and TBS–LTP with age (in days) for each genotype. TBS–PTP was not significantly affected by age in WT mice (Figure 4A, n = 20, r = −0.316, p = 0.755). In contrast, a significant negative correlation was observed between TBS–PTP and age for both heterozygous (Figure 4B, n = 35, r = −0.669, p < 0.001) and homozygous mice (Figure 4C, n = 26, r = −0.713, p < 0.001). That is, TBS–PTP significantly decreased with disease progression in both heterozygous and homozygous mice (Figure 4D). Surprisingly, and in contrast to TBS–PTP, TBS–LTP was not significantly affected by age in WT (Figure 4E, n = 20, r = 0.348, p = 0.132), heterozygous (Figure 4F, n = 35, r = −0.187, p = 0.286) or homozygous mice (Figure 4G, n = 26, r = −0.217, p = 0.288). These data demonstrate that deficits in presynaptic short-term plasticity are more closely linked to disease progression than deficits in postsynaptic long-term plasticity. As age had no significant effect on TBS–LTP measures in either genotype, we pooled the data from all ages (Figures 4I,J). Note that the data presented in Figures 4I,J represent the combined data from Figures 3A,D,G. By combining the three different age groups, the higher n-values now revealed significant post hoc LTP deficits for both heterozygous and homozygous mice, with the deficit being more pronounced in the latter (Figures 4I,J, WT n = 20, heterozygous n = 35, homozygous n = 26, one way-ANOVA, p < 0.001, Tukey post hoc: WT vs. heterozygous p < 0.05; WT vs. homozygous p < 0.001; heterozygous vs. homozygous p < 0.01). In all, these results suggest that while deficits in both short- and long-term plasticity exist in the HD hippocampus, short-term plasticity deficits may be more closely linked to disease progression than long-term plasticity deficits.

It was previously shown that symptomatic homozygous knock-in Hdh^Q111 mice exhibit a reduced CA1 stratum radiatum population response to TBS of the Schaffer collateral pathway (Lynch et al., 2007). Thus, we were interested in determining whether Q175FDN mice exhibited abnormal population responses during the duration of the TBS itself. We analyzed the area under the curve of the fEPSP response to each burst of the TBS in WT as well as heterozygous and homozygous Q175FDN mice at the aforementioned presymptomatic, early symptomatic and late symptomatic ages. At 3 months, the fEPSP response to TBS was not significantly different between WT, heterozygous and homozygous mice, and no significant genotype or interaction effects were observed (Figure 5A, WT n = 10, heterozygous n = 24, homozygous n = 8, repeated-measures two-way ANOVA, p_(genotype) = 0.292, p_(burst) < 0.001, p_{(interaction)} = 0.062). The trending interaction effect largely reflects the non-significant tendency for homozygous mice to maintain larger mean fEPSP responses throughout the duration of the TBS. In contrast, at 6–9 months, heterozygous and homozygous mice exhibited lower mean fEPSP responses early in the TBS, and significant genotype and interaction effects were observed (Figure 5B, WT n = 10, heterozygous n = 16, homozygous n = 19, repeated-measures two-way ANOVA, p_(genotype) = 0.041, p_(burst) < 0.001, p_{(interaction)} < 0.001; Bonferroni post hoc: p > 0.05 for all comparisons). At 12 months, heterozygous responses to TBS tended to be lower than WT responses early in TBS but exhibited less desensitization throughout; this pattern resulted in a significant interaction effect in the absence of a genotype effect (Figure 5C, WT n = 7, heterozygous n = 8, repeated-measures two-way ANOVA, p_(genotype) = 0.494, p_(burst) < 0.001, p_{(interaction)} < 0.001; Bonferroni post hoc: p > 0.05 for all comparisons).

While less physiologically relevant than TBS, high-frequency stimulation (HFS; 100 Hz, 1 s) is the most common stimulation paradigm used to evoke LTP. Different patterns of plasticity-inducing stimuli, including TBS and HFS, are known to produce LTP via distinct mechanisms (Zhu et al., 2015). As the mechanisms underlying the plasticity deficit in HD are not well-understood, any different sensitivities to one form of LTP over the other may help reveal abnormal signaling pathways in the HD hippocampus. Lastly, we sought to determine whether HFS–LTP was more or less susceptible to the HD-causing mutation compared to TBS–LTP. To address this question, we focus our efforts on WT and the more relevant heterozygous mice, and chose to conduct these experiments at 6 month of age (±2 weeks). At this age, and including mice up to 9 months of age, no significant TBS–LTP deficit was observed in heterozygous mice; homozygosity for the HD-causing mutation was required to observe a TBS–LTP deficit at this age. Interestingly, while 6 month heterozygous mice exhibited perfectly normal PTP following HFS (HFS–PTP, Figures 6A,B, WT n = 8, heterozygous n = 9, unpaired t-test, p = 0.715), a significant impairment of HFS–LTP was observed (Figure 6C, WT n = 8, heterozygous n = 9, unpaired t-test, p = 0.023). This result suggests that pathways required for HFS–LTP may be affected earlier in HD than those required for TBS–LTP, and also demonstrates that a clear LTP impairment can be observed in HD mice in the absence of a PTP deficit.

Hippocampal LTP deficits have been previously described in numerous HD mouse models (Hodgson et al., 1999; Murphy et al., 2000; Lynch et al., 2007; Simmons et al., 2009; Anglada-Huguet et al., 2014; Brito et al., 2014; Kolodziejczyk et al., 2014; Giralt et al., 2017; Zhang et al., 2018). Many of these studies were conducted at a symptomatic age in N-terminal fragment, overexpression or homozygous knock-in HD mouse models; thus, while it is well-understood that HD is associated with poor synaptic plasticity, the onset and progression of short- and long-term plasticity impairments are poorly documented. The present study extends on the existing literature by quantifying basal connectivity, presynaptic release probability as well as short-term presynaptic and long-term postsynaptic plasticity effects at three different disease stages in both heterozygous and homozygous knock-in HD mice. In addition, we quantified activity-dependent long-term plasticity in response to two commonly used LTP induction paradigms that are known to rely on different mechanisms (Zhu et al., 2015). To the best of our knowledge, the current study represents the most thorough characterization of the onset and progression of hippocampal plasticity deficits in a mouse model of HD to date.

In all, the current study demonstrates clear increases in basal synaptic strength as well as deficits in short- and long-term synaptic plasticity in the hippocampus of Q175FDN mice. Interestingly, the short-term plasticity impairments did not emerge or progress in parallel with the observed impairments in long-term plasticity, and Q175FDN heterozygous mice were more susceptible to LTP impairment following HFS compared to TBS. Our data suggest that the various plasticity deficits in the HD hippocampus are unlikely to be mediated by a common mechanism; rather, independently developing perturbations in a variety of signaling pathways are likely to combine to contribute to HD hippocampal pathogenesis. As plasticity deficits were exacerbated in homozygous mice that lack non-pathogenic huntingtin, it is of interest for future studies to determine whether allele-selective huntingtin-lowering strategies (Southwell et al., 2018) will prove to be more beneficial than non-selective strategies in preventing or restoring HD-associated plasticity impairments.

The datasets generated for this study are available on request to the corresponding author.

All experiments with live animals were approved by the Memorial University’s Animal Care Committee in accordance with the guidelines set by the Canadian Council on Animal Care.

JQ performed the experiments. MP designed and supervised the experiments. Both authors analyzed the data and wrote the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.