Huntington disease (HD [MIM 143100]) is a neurodegenerative disorder caused by an expanded CAG trinucleotide repeat that lengthens a polyglutamine tract near the amino terminus of the huntingtin protein [1]. The mutation involves a "gain-of-function" that leads to the selective loss of vulnerable neurons, most notably medium spiny neurons in the caudate nucleus, and results in characteristic progressive writhing movements, in addition to psychological changes and cognitive decline [2]. The age at onset of HD is variable, though the motor disturbance typically begins in mid-life, and is followed by an inexorable decline that ends in death after a course of 10–20 years. The length of the expanded CAG trinucleotide tract is the primary determinant of disease onset, although the mechanism whereby mutant huntingtin triggers the cascade of HD pathogenesis that eventually produces these symptoms is not yet clear [3, 4]. Like several other polyglutamine neurodegenerative disorders, there is a strong inverse correlation between the length of the polyglutamine tract and the age at neurologic onset, with the longest mutations leading to juvenile HD. Indeed, the HD CAG repeat length alone accounts for more than half of the overall variance in age at neurologic onset. We and others have showed that the remaining variation in age at neurologic onset of HD is highly heritable and that other genetic factors act to modify the pathogenic process [5–16]. Identification and characterization of such modifiers is of critical importance, since their capacity to alter pathogenesis would make them potential targets for development of therapeutics for this currently untreatable disorder.

To date, one candidate genetic modifier of HD has been confirmed in multiple independent studies. An association between genetic variation at GRIK2 (GenBank: mRNA NM_021956), encoding the GluR6 subunit of the kainate receptor, and age at neurologic onset in HD was first reported by Rubinsztein et al., based on 293 HD patients from the United Kingdom [13]. A TAA repeat polymorphism (D6S1028; CHLC.ATA22H10) in the 3' untranslated region (UTR) of GRIK2 accounted for 13% of the residual variance in onset age after accounting for the effect of the HD CAG mutation. The 16 repeat allele (designated '155' based on the PCR product length) was found in those with younger than expected HD onset ages, leading to the suggestion that the TAA repeat was a neutral polymorphism in linkage disequilibrium with a functional variant in GRIK2, or in a nearby gene. The GRIK2 association was confirmed in 258 unrelated U.S. HD patients, where the relatively rare 16 TAA allele was associated with an approximate 5-year-younger age at onset [11]. Subsequently, a modifier role for this locus has been demonstrated in HD populations from France, India, and Italy [6, 7, 12]. To address the issue of a functional variant at or near the GRIK2 locus, we have tested whether a coding sequence variant explains the modifier effect, and subsequently, we have used analysis of polymorphisms in the GRIK2 region to implicate the TAA repeat itself as the source of the genetic modification of onset age.

GRIK2, encoding the GluR6 subunit of the kainate receptor, is abundantly expressed in the caudate nucleus, cortex, hippocampus and cerebellum in human. Kainate receptors are one of three subtypes of ionotropic receptors for the excitatory amino acid transmitter L-glutamate. They both mediate excitatory synaptic transmission at primary afferent synapses and modulate presynaptic neurotransmitter release [20]. Interaction of excitatory amino acids with their receptors can potentially cause toxicity that upsets Ca⁺⁺ homeostasis and damages energy metabolism, and such 'excitotoxicity' due to glutamate has long been postulated to be involved in HD pathogenesis [21, 22]. However, a recent analysis of systemic 3-nitropropionic acid induced striatal lesions in Grik2^-/- mice showed no difference in final histopathology or motor impairment compared with similarly treated wild-type mice, suggesting that striatal GluR6 kainate receptors do not play a critical role in determining neuronal pathology [23]. In this experiment, Grik2^-/- mice did display motor symptoms earlier than wild-type mice, despite the lack of difference in neuropathology, suggesting that the GluR6 kainate receptors may have played some modulatory role in synaptic transmission rather than mediating excitotoxicity.

There is abundant evidence that genetic modifiers of HD pathogenesis exist, from studies performed before the HD gene was cloned through studies performed more recently with the benefit of molecular definition of the HD mutation [5–16, 24, 25]. GRIK2 is currently the only genetic modifier of HD that has been specifically identified and confirmed in more than two independent studies [6, 7, 11–13]. Our findings indicate that 'modifier' chromosomes did not have a functional coding sequence difference that directly alters GluR6. Rather, they shared the 16 TAA allele but had different haplotypes for SNPs within a 2.3 kb region encompassing the polymorphic repeat, suggesting that the TAA allele itself, rather than a linked polymorphism, is responsible for the modifier effect. Though the effect may be peculiar to 16 TAA alleles, we cannot exclude that it is a length-dependent effect encompassing both 16 TAA and longer alleles, as there are too few 17 TAA alleles to test for a statistically significant modifier effect. However, the very early onset of GUS17843 suggests that 17 TAA alleles may also have the potential to modify neurologic onset in HD.

It is likely that the 3' UTR TAA repeat modifier effect acts at the level of the GRIK2 mRNA, with multiple opportunities for allelic alterations in regulation beyond considerations of translational efficiency and stability. Indeed, with the exception of the immediate vicinity of the TAA repeat, the 3'UTR of GRIK2 is remarkably conserved across species, suggesting that it has an important regulatory role (Figure 2). GRIK2 displays alternative splicing of exon 16 that leads to two major splice variants encoding proteins with different carboxyl-termini. Two other minor alternative splice forms, one skipping exon 12 (predicting a truncated protein) and one in which a cryptic splice site in exon 11 joins to exon 13 (predicting a protein lacking the first two of four transmembrane domains), have been described. Moreover, three major A to I RNA editing sites that can affect ion permeability of the receptor due to resulting amino acid substitutions I>V, Y>C and Q>R, respectively are located in exons 11 (2) and 12. Fully unedited receptors at the Q/R site exhibit higher relative calcium permeability and higher channel conductance as compared to edited receptors [26–28]. Interestingly, two alternative transcripts differing in the length of the 3'UTR have been reported, only one of which includes the TAA repeat [19, 29]. The shorter mRNA uses a polyadenylation signal located 92 bases after the stop codon, upstream of the TAA repeat while the longer mRNA extends beyond a putative polyadenylation signal 1547 bases downstream of the stop codon, 3' to the TAA repeat [29]. Thus, numerous opportunities exist for the TAA repeat polymorphism to potentially modify HD onset by virtue of an allelic regulatory effect on the production, polyadenylation status, alternative splicing, subcellular targeting, editing or stability of the GRIK2 mRNA. The aforementioned Grik2^-/- mouse experiments also indicate that an effect on age at neurologic onset may not necessarily be mediated by excitotoxicity, but could occur due to an independent effect on motor function. The fact that GRIK2 is capable of modifying the pathogenic process in HD sufficiently to have a noticeable impact on onset age indicates that it is a potential target for development of an effective therapeutic intervention. However, it will be necessary to delineate of the precise basis for the GRIK2 modifier effect in order to effectively undertake a search for chemical compounds that delay, rather than hasten HD onset.

GRIK2 acts as a modifier of HD pathogenesis, not through a coding sequence polymorphism that changes the structure of the GluR6 protein, but rather through a TAA repeat polymorphism in the 3'UTR. Translating this knowledge into a treatment to delay HD pathogenesis will first require definition of the functional consequences of the TAA repeat, possibly involving altered regulation of GRIK2 mRNA.