1.1 Size matters; one mutation, two different forms of HD

HD is generally conceived as an adult-onset neurodegenerative disorder. However, another less common (5%–10%) juvenile form (JHD, known as rigid or Westphal variant) of the disease also exists, typically when the CAG triplet repeat expansion is >65 (Hunnicutt et al., 2016). Studies have shown that the sex of the transmitting parent, usually the father, exerts a major influence on CAG repeat expansion leading to earlier symptom onset (Telenius et al., 1993). The symptoms of JHD differ from those typically seen in adult-onset HD. Children with HD display mental retardation, hyperactivity, and aggressive behavior. Some of these children also have microcephaly, suggesting that this form of HD may represent a developmental rather than a neurodegenerative disorder (Hunnicutt et al., 2016; Letort & Gonzalez-Alegre, 2013). With disease progression, dystonia, rigidity, and chorea also occur. A fundamental difference between JHD and adult-onset HD is the high prevalence of epileptic seizures in the juvenile form (Gambardella et al., 2001; Rasmussen et al., 2000; Seneca et al., 2004). The cause of epileptic seizures remains unknown (Cummings et al., 2009). However, we have hypothesized that seizures could be the result of faulty development of cortical circuits, similar to those observed in malformations of cortical development (MCD), specifically focal cortical dysplasia (FCD) of Taylor (Estrada-Sánchez, Levine, & Cepeda, 2016; Taylor, Falconer, Bruton, & Corsellis, 1971). This was based on evidence that in some animal models of HD, and presumably also in human cases, the cerebral cortex progressively becomes hyperexcitable (Cummings et al., 2009).

1.2 HTT is essential for normal brain development

The HTT protein has multiple functions that include, among many others, vesicle trafficking, spindle orientation during cell division, endocytosis, transcriptional regulation, maintenance of cell morphology and survival (Saudou & Humbert, 2016). It is also known that normal HTT plays a crucial role during development, as lack of this protein is lethal (Duyao et al., 1995; Nasir et al., 1995; Saudou & Humbert, 2016; Zeitlin, Liu, Chapman, Papaioannou, & Efstratiadis, 1995). Other studies have shown that embryonic, conditional deletion of HTT from cortical pyramidal neurons (CPNs) reduced cortical volume and neuron abundance (Dragatsis et al., 2017). Similarly, loss of HTT function in subpallial lineages disrupted forebrain interneuron species early in life and also led to a number of neurological deficits reminiscent of HD (Mehler et al., 2019).

In the past few years, the idea that HD in general, and JHD in particular, is not solely a neurodegenerative but also a neurodevelopmental disease, has gained momentum (Durieux, Schiffmann, & de Kerchove d'Exaerde, 2011; Godin et al., 2010; Wiatr, Szlachcic, Trzeciak, Figlerowicz, & Figiel, 2017). Indeed, the HTT protein may alter different aspects of chromatin regulation and transcription during neural development (Durieux et al., 2011). For example, in vivo inactivation of HTT by RNA interference or deletion of the gene affects spindle orientation and cell fate of cortical progenitors in the ventricular zone of mouse embryos, altering the thickness of the developing cortex as well as the polarization and migration of newly generated neurons (Godin et al., 2010; Molina-Calavita et al., 2014). Furthermore, depletion of HTT in post-mitotic projection neurons leads to the mislocalization of layer-specific neuronal populations in the mouse neocortex, suggesting that HTT, via regulation of RAB11-dependent N-Cadherin trafficking, is critical for neuronal migration (Barnat, Le Friec, Benstaali, & Humbert, 2017). Importantly, the authors also evinced that mutant HTT (mHTT) loses its capacity to promote neuronal migration. Normal HTT also is required for the correct establishment of cortical and striatal excitatory circuits and this function is lost when the mHTT is present (McKinstry et al., 2014). When cortical HTT function is conditionally silenced from CPNs, cortical and striatal excitatory synapses form and mature at an accelerated rate through postnatal day (P)21 but exuberant synaptic connectivity is lost over time in the cortex, resulting in the deterioration of synapses by 5 weeks of age (McKinstry et al., 2014). It can thus be postulated that mHTT impairs neurodevelopmental pathways (Blockx et al., 2012; HD iPSC Consortium, 2017). Indeed, expression of mHTT during early development is sufficient to produce a permanent HD phenotype even if expression is terminated at P21. Furthermore, developmental deficits associated with HTT function render cells more susceptible to degeneration (Arteaga-Bracho et al., 2016; Molero et al., 2016).

Not only is cortical brain development compromised by the presence of mHTT, but CPN function itself is altered by the early formation of mHTT aggregates. In R6/2 transgenic mice (Mangiarini et al., 1996), intranuclear inclusions can be detected as early as 3 weeks of age (Cummings et al., 2012; Meade et al., 2002; Morton, Lagan, Skepper, & Dunnett, 2000). At this age, intranuclear inclusions are most abundant in CA1 hippocampal region and cortical layers III–V, whereas in striatum inclusions are rare (Morton et al., 2000). Using more sensitive immunohistochemical methods, diffuse proto-aggregates have been observed in developing axonal tracts during embryonic development and early postnatal brains in juvenile and adult-onset mouse models of HD (Osmand, Bichell, Bowman, & Bates, 2016). These axonal aggregates could alter synaptic physiology during early postnatal development. In support, using a corticostriatal coculture from YAC128 mice, a model of adult-onset HD (Slow et al., 2003), synaptic transmission was impaired as early as three weeks in vitro (Buren, Parsons, Smith-Dijak, & Raymond, 2016). In addition, in an in vitro model of HD based on the generation of induced pluripotent stem cells from HD patients and controls, it was observed that HD-derived cells displayed a greater number of neuronal progenitors compared with controls. This cell population showed enhanced vulnerability to brain-derived neurotropic factor (BDNF) withdrawal in the JHD lines (Mattis et al., 2015). Interestingly, increased vulnerability was due to N-methyl-D-aspartate (NMDA) glutamate receptor-mediated toxicity, suggesting that aberrant Ca²⁺ signaling could be involved.

The observation of aberrant migration and polarization of cortical progenitors is reminiscent of the cortical maldevelopment observed in humans with FCD, a disorder characterized by dyslamination, CPN misorientation, and the presence of dysmorphic pyramidal neurons, all of which contribute to epileptogenesis (Cepeda, Hurst, Calvert, et al., 2003; Dautan et al., 2016; Taylor et al., 1971). Thus, we could hypothesize that the presence of mHTT in neuronal progenitors affects cortical organization and induces diffuse architectural and cellular abnormalities, similar to those observed in FCD, which results in cortical hyperexcitability (Blumcke et al., 2011; Cummings et al., 2009; Estrada-Sánchez et al., 2016).

1.3 The cerebral cortex is hyperexcitable in HD brains

The central role of the cerebral cortex in the genesis of the HD phenotype has long been recognized in clinical and experimental studies (Estrada-Sanchez & Rebec, 2013; Laforet et al., 2001; Paulsen et al., 2006; Virlogeux et al., 2018). In this section we review clinical and experimental evidence.

1.4 New observations in developing R6/2 mice point to aberrant cortical neuron development and early hyperexcitability

1.4.1 Morphological and electrophysiological findings

Recently we extended our studies on CPN excitability to developing R6/2 and control littermates (P7 and P14). Mice (n = 13) of either sex were used at P7 (6 WT and 7 R6/2). Experimenters were blind to the genotype, which was only identified a posteriori from DNA tail samples, thus ensuring unbiased results. CPNs in layers II/III and V from the motor cortex were visualized with infrared differential interference contrast (IR-DIC) optics (see detailed methods in Supplementary Material). In order to increase cortical excitability, the GABA_A receptor antagonist bicuculline (BIC, 10 μM), alone or in conjunction with the type “A” K⁺ channel blocker 4-aminopyridine (4-AP, 100 μM), were used. Both compounds are proconvulsant and we have used them in the past to examine cortical excitability and seizure susceptibility in HD mice (Cummings et al., 2009). The patch pipette also contained biocytin (0.2%) to label the recorded cells and, following histological processing, the fine morphology of CPNs was examined.

Preliminary observations indicate that CPN membrane excitability is increased in R6/2 mice as early as P7, the earliest age we examined. Indeed, blockade of GABA_A receptors with BIC induced paroxysmal discharges more often in CPNs (4 out of 8) from HD mice but not in CPNs (1 out of 7) from age-matched WT littermates. Furthermore, concurrent application of BIC and 4-AP generated ictal-like activity in R6/2 CPNs, but not in WT CPNs (Figure 1b). These observations are consistent with our previous study demonstrating the higher occurrence of complex paroxysmal discharges in R6/2 compared with WT mice at P21 (Cummings et al., 2009), and highlight the fact that changes in cortical excitability are present even earlier during postnatal life.

Morphological evidence of CPN maldevelopment also was observed during early development. Under IR-DIC microscopy we observed that some CPNs displayed misorientation, for example, apical dendrites pointing in the wrong direction, as well as the presence of dysmorphic dendritic and axonal processes (Figure 1a,c). While these processes extend with smooth transitions in CPNs from WT mice, in R6/2 mice some CPNs exhibit curvy, meandering dendrites and axons, suggesting that pathfinding mechanisms during fetal development had been disturbed by the presence of mHTT. At P14, signs of hyperexcitability in CPNs from R6/2 mice also were found (Figure 1d). In R6/2 CPNs, bath application of BIC induced more frequent epileptiform discharges compared with age-matched controls.

1.5 FCD-like abnormalities occur in the cerebral cortex of R6/2 mice

The cellular morphological abnormalities observed in young R6/2 mice are reminiscent of those we found in our studies of cortical tissue samples from pediatric epilepsy surgery patients with FCD histopathology (Abdijadid, Mathern, Levine, & Cepeda, 2015; Cepeda, Hurst, Flores-Hernandez, et al., 2003), suggesting that cortical maldevelopment in HD could underlie cortical hyperexcitability and seizure proclivity. According to the International League Against Epilepsy (ILAE) classification of FCD, there are at least 3 different categories (Barkovich, Dobyns, & Guerrini, 2015; Blumcke et al., 2011). FCD type 1 is characterized by cortical dyslamination with aberrant columnar and/or radial architecture. In addition, CPN misorientation and abnormal processes can be found. FCD type 2 is defined by additional features including the presence of dysmorphic, cytomegalic neurons and in some cases also balloon cells. FCD type 3 occurs only in association with other pathologies such as tumors or hippocampal sclerosis.

In a previous report, we speculated that one of the underlying mechanisms of cortical hyperexcitability and seizures in R6/2 mice could be a mild form of FCD, that is, type 1 (Estrada-Sánchez et al., 2016). To test this hypothesis, we recently examined potential histopathological evidence of FCD in symptomatic R6/2 mice (age P60). We chose this age group in a first approximation because spontaneous or evoked convulsive seizures are only observed in fully symptomatic mice. Six pairs of R6/2 and control mice of either sex were perfused, sliced, and stained for the specific neuronal marker, NeuN to examine cortical cytoarchitecture and CPN morphology (see Supplementary Material). Initial observations indicate that FCD-like abnormalities occur in the motor and somatosensory cortices of at least 50% of R6/2 mouse brains examined (Figure 2a). Abnormalities include cortical dyslamination, neuronal crowding in some areas but others devoid of NeuN label, all suggesting cortical maldevelopment. In addition, biocytin staining of CPNs demonstrated the presence of abnormal processes similar to those observed in R6/2 mice at P7, including misoriented neurons, recurving dendrites and extreme meandering of axonal processes (Figure 2b). Interestingly, in another model of HD with 100 CAG repeats we also documented the presence of misoriented CPNs as well as the common occurrence of dysmorphic dendrites (Laforet et al., 2001). These observations beg the question as to whether dysmorphic dendrites are the result of degenerative or abnormal neurodevelopmental processes.

1.6 Is HD a special case of MCD?

Cortical development is a delicate process that follows precise spatial and temporal rules. If, for any reason, these creodes (Waddington, 1962) are violated, the cytoarchitecture of the cortex crumbles. In particular, when CPNs terminate in ectopic positions FCD and other structural malformations ensue, leading to a growing number of neurological and psychiatric diseases (Ayoub & Rakic, 2015; Rakic, 2006; Wu et al., 2014). MCD represent a group of pathologies associated with aberrant development of the cerebral cortex and are a common etiology of epilepsy (Barkovich et al., 2015). The causes of MCD are multiple and diverse. Genetic as well as extrinsic factors play a role. As HD is a genetic disorder, we can surmise that mHTT is able to alter the rules that underlie cortical development. This is supported by experimental evidence that the presence of mHTT during embryonic development affects neuronal migration and final positioning and orientation (Barnat et al., 2017; Osmand et al., 2016).

Using Golgi impregnation of striatal neurons from HD patients, Graveland et al. described the abnormal presence of recurving terminal dendrites in MSNs (Graveland, Williams, & DiFiglia, 1985). Notably, the authors suggested that the high frequency of recurved dendrites may be a clue to the pathophysiology of HD. Degeneration and regeneration of CPNs from HD patients was also evinced by increases in the length of terminal branches and an overall greater branching complexity of the dendritic trees, somehow recapitulating normal brain development (Sotrel, Williams, Kaufmann, & Myers, 1993). Furthermore, an antibody used to detect the N-terminal region of mHTT was found in neuronal intranuclear inclusions and dystrophic neurites in the HD cortex and striatum (DiFiglia et al., 1997; Sapp et al., 1999). Dystrophic neurites, likely corresponding to distended axon terminals, were more prevalent in deep cortical layers and because they could be seen in a presymptomatic adult patient it was suggested that they precede clinical onset (DiFiglia et al., 1997). Morphological changes in the cerebral cortex include enlargement of gyral crowns and abnormally thin sulci (Nopoulos et al., 2007; Paulsen et al., 2006). Interestingly, enlargement of cortical gyri also is observed in some FCD types (Blumcke et al., 2011). Cortical thinning and white matter loss are common in pre-manifest HD subjects (Aylward, 2007; Reading et al., 2005; Rosas et al., 2005, 2006; Waldvogel et al., 2015). Notably, smaller intracranial volumes in prodromal HD patients indicate that mHTT can cause abnormal brain development (Nopoulos, Aylward, Ross, Mills, & Langbehn, 2011). Interestingly, smaller intracranial volumes can be associated with cerebellar enlargement, which could explain hypokinesia in JHD (Tereshchenko et al., 2019). Thus, HD therapies have to take into account that the goal is not only to prevent neurodegeneration in a susceptible brain but also to correct aberrant development.

1.7 Targeting cortical maldevelopment as a new strategy for the treatment of HD

There is an almost universal consensus that, in order to treat HD symptoms, interventions should start early. The question is how early? Based on the previous review of the literature and recent morphological and electrophysiological data in very young R6/2 mice, it appears that only by targeting early brain development can any treatment be successful. So the question becomes, can cortical maldevelopment be rescued? Using human HD induced pluripotent stem cell cultures it was demonstrated that the presence of mHTT negatively affects striatal and cortical neuronal progenitor specification and commitment leading to abnormal cell organization and acquisition of mature neuronal identities in cerebral organoids (Conforti et al., 2018). Notably, these defects could be rescued by down-regulating mHTT with synthetic Zinc Finger Proteins or pharmacologically by inhibiting the metalloproteinase ADAM10, which is a mHTT effector. Another study showed that very early behavioral, cellular, and molecular changes associated with the presence of mHTT can be reverted through administration of HDAC inhibitors (Siebzehnrübl et al., 2018). Finally, a recent study using a disease-on-a-chip microfluidic platform to examine the corticostriatal network in vitro, provided further evidence that cortical alterations are critical to the progression of the disease (Virlogeux et al., 2018). Further, substitution of HD cortical neurons with wild-type neurons was sufficient to rescue cellular alterations in mutant striatal neurons. Although we are still far from using these experimental approaches in human patients, these results offer a glimmer of hope.