Accumulation of the neurotoxic proteolytic derivative of amyloid beta precursor protein (APP), amyloid-beta (Aβ) peptide, is proposed to be key to the pathogenesis of Alzheimer’s disease (AD) (Hardy and Selkoe, 2002). Sortilin-related receptor 1 (SORL1) encodes a mosaic protein (SORLA) consisting of several distinct domains, including the vacuolar protein sorting 10 protein (VPS10P) domain for Aβ binding and the low-density lipoprotein receptor domain for APP binding and lipoprotein binding (Barthelson et al., 2020). SORLA acts as a sorting receptor for retrograde trafficking of APP to the trans-Golgi-network to prevent APP processing to Aβ and anterograde movement of Aβ for lysosomal degradation (Andersen et al., 2016). SORL1 mutations in human neurons lead to reduced levels of SORLA, resulting in defects of the neuronal endolysosome function and autophagy (Hung et al., 2021). In addition, being a low-density lipoprotein receptor, SORLA mediates neuronal uptake of apolipoprotein E (ApoE)-rich lipoproteins (Yajima et al., 2015), the misfolding of which contributes significantly to AD pathogenesis (Barthelson et al., 2020). Ablation of SORLA expression increases Aβ in the brain of knockout mice (Andersen et al., 2005), and SORLA-overexpressing cells have remarkably reduced levels of extracellular Aβ and lower levels of intracellular APP (Offe et al., 2006). In the brains of AD patients, SORL1 expression is reduced (Scherzer et al., 2004). Moreover, both common and rare variants of SORL1 have been associated with late-onset and early-onset AD, respectively (Campion et al., 2019).

SORL1 variants were first identified, among several genes belonging to endocytic pathways, to be associated with sporadic AD by a pioneering study in Caucasians (Rogaeva et al., 2007). Targeted single-nucleotide polymorphism (SNP) analyses and genome-wide association studies have validated the association not only in populations of Caucasian origin but also in Asian populations (Reitz et al., 2011; Lambert et al., 2013; Miyashita et al., 2013). A number of studies have shown significant associations between specific SORL1 polymorphisms and various phenotypes in AD patients, including lower Aβ levels in cerebrospinal fluid (Alexopoulos et al., 2011) and serum (Chou et al., 2016), increased tau protein in cerebrospinal fluid (Louwersheimer et al., 2015), hippocampal atrophy (Cuenco et al., 2008; Louwersheimer et al., 2015; Xiromerisiou et al., 2021), white matter hyperintensity (Cuenco et al., 2008), frontal symptoms (Huang et al., 2020), rate of cognitive decline (Hsieh et al., 2021), and Parkinsonian features (Cuccaro et al., 2016; Xiromerisiou et al., 2021).

We have previously reported that in the Han Chinese population in Taiwan, two common variants of SORL1, rs2298813, and rs1784933, were associated with late-onset AD (Chou et al., 2016). In the elderly population in Australia and the United States, rs2298813 has also been identified in individuals with late-onset AD (Assareh et al., 2014; Cuccaro et al., 2016). The association of rs1784933 with the risk of late-onset AD has also been reported in the Han Chinese population in China (Feng et al., 2015; Zhang et al., 2017) and Mexicans (Toral-Rios et al., 2022). The SNPrs2298813 is located at the VPS10P domain, and the nonsynonymous substitution of alanine to threonine at the 528th residue (A528T) of SORLA has been shown to increase the secretion of Aβ42, soluble APPα, and APPβ in vitro (Vardarajan et al., 2015). The SNPrs1784933 is located in the 3’ region of SORL1, and minor allele carriers with late-onset AD had lower plasma concentrations of Aβ42 (Chou et al., 2016). The endophenotypic effects of various SORL1 polymorphisms on the brain have been revealed in nondemented individuals (Liang et al., 2015; Huang et al., 2016; Yin et al., 2016; Li et al., 2017). For example, among the non-demented elders, rs1699102 was associated with gray matter volume of the right middle temporal pole (Li et al., 2017), and rs1784933 and rs753780 was associated with right parahippocampal volume (Yin et al., 2016). However, the neuroimaging correlates of rs2298813 and rs1784933 have not been established in AD patients. We, therefore, aimed to examine the associations between the two SORL1 SNPs and gray matter volume and cortical thickness of different brain regions in late-onset AD patients.

We found that the volume of the right lentiform nucleus differed between the homozygous major allele carriers of rs2298813 and the minor allele carriers in late-onset AD patients. With respect to rs1784933, there were no neuroimaging correlates of the genotype. Among the homozygous major allele carriers of rs2298813, there was an association between right putaminal volume and verbal fluency performance, and the right putaminal and right pallidal volumes were predictive of clinical progression. On the contrary, the minor allele carriers of rs2298813 had a smaller volume of the right lentiform nucleus and a higher cognitive fluctuation score. However, the right putaminal and right pallidal volumes were not related to neurocognitive test performance nor predictive of clinical progression among these patients. It appeared that the major and minor alleles of rs2298813 had differential effects on the volume of the right lentiform nucleus and differentially modulated the association between the right lentiform nucleus and cognitive function.

The earliest associations between rs2298813 and rs1784933 and brain atrophy were delineated in a collection of autopsied AD brains (Cuenco et al., 2008). The 3-SNP haplotypes containing the rs1784933 region were related to pathological scoring and MRI traits of hippocampal atrophy in white AD patients, whereas no significant associations were identified for rs2298813 (Cuenco et al., 2008). In an Australian cohort, a 3-SNP haplotype containing rs2298813 was associated with whole brain atrophy in both males and females, but rs2298813 was not individually associated with brain atrophy (Assareh et al., 2014). A study using the Alzheimer’s Disease Neuroimaging Initiative database investigated the effect of eight SORL1 SNPs, including rs2298813 and rs1784933, on AD-related brain atrophy in subjects with normal cognition or mild cognitive impairment (MCI) (Yin et al., 2016). The A allele of rs2298813 tended to be associated with a lower volume in the right parahippocampal gyrus, and the G allele of rs1784933 was found to be associated with a higher rate of atrophy in the right parahippocampal gyrus across a 2-year span (Yin et al., 2016). In a young healthy Caucasian population, 117SNPs in and surrounding SORL1 were examined to determine any association with hippocampal volume, and the majority of significant associations occurred at the 3’ region, where rs1784933 is located (Bralten et al., 2011). The discrepancy in the relationship between rs2298813/rs1784933 and brain atrophy may lie in ethnic differences (Jin et al., 2013) and the small effect of a single SNP that may be better demonstrated in quantitative measurement of brain volume or in haplotype analysis.

The putamen and pallidum (forming the lentiform nucleus) are essential elements of the extrapyramidal system and are usually involved in motor disturbances, such as Parkinsonism or Huntington’s disease (Albin et al., 1989). Although not included among the typical AD-related brain structures, numerous studies have disclosed an association of these regions with AD. The volume of the putamen has been shown to be reduced in AD brains relative to subjects with MCI or normal aging (De Jong et al., 2008; Roh et al., 2011; Cho et al., 2014; Tang et al., 2014; Eustache et al., 2016; Pini et al., 2016). Such reductions can occur as early as in the prodromal stage (Eustache et al., 2016). Atrophy in the pallidum has not been consistently observed in previous studies (Cho et al., 2014; Pini et al., 2016). The pallidum is relatively resistant to degeneration even in moderate stages of the disease (Roh et al., 2011) although mild atrophy may be observed (Li et al., 2013; Tang et al., 2014; Wang et al., 2018). AD pathology has been demonstrated to deposit heavily in the putamen and less in the pallidum (Braak and Braak, 1990). Similarly, iron detection using MR techniques (quantitative susceptibility mapping or phase imaging) has revealed iron accumulation in the putamen and pallidum (Bartzokis et al., 2000; Cogswell et al., 2021), which was associated with higher amyloid PET standardized uptake value ratios (Cogswell et al., 2021). The association of putaminal or pallidal atrophy with common variants of SORL1 has not been previously specified in AD patients. The most relevant study was conducted by Huang et al. investigating the effect of rs3824968 on gray matter volume in a nondemented Chinese population across a wide age span (Huang et al., 2016). Participants carrying the A allele had accelerated atrophy with age in the right putamen (Huang et al., 2016). Thus, we are the first to identify a link between putaminal/pallidal atrophy and SORL1 polymorphisms in AD.

The putamen and pallidum also have functions related to cognition. In patients with Parkinson’s disease, deep brain stimulation of the pallidum has been shown to improve verbal fluency. (Lee et al., 2018) In human immunodeficiency virus-associated neurocognitive impairment, there was an association between cognitive impairment and putaminal volume (Qi et al., 2021). Among patients with behavioral variant frontotemporal dementia, there is a relationship between atrophy in the putamen and the theory of mind impairment (Baez et al., 2019). In support of this, in our study, relationships between both verbal fluency and the annual rate of changes in MMSE scores and right putaminal volume were observed in the homozygous major allele carriers of rs2298813. In contrast, there were no significant associations between right putaminal/right pallidal volumes and cognitive test performance or clinical progression in the minor allele carriers even though the right lentiform nucleus was smaller in these patients. This may suggest that in this subgroup of patients, atrophy in the right lentiform nucleus did not have a deleterious impact on cognitive function. The atrophy may, for example, have more effect on motor function. Further study is needed to elucidate the phenotypic effect of regional atrophy in the right lentiform nucleus in the minor allele carriers of rs2298813.

The link between rs2298813 and the lentiform nucleus volume may provide a pathophysiological basis for the association between AD and parkinsonism. Indeed, Parkinsonism is not uncommon in AD. Extrapyramidal signs can be detected in one-third of AD patients during the course of the disease (Scarmeas et al., 2004), and Parkinsonian features are related to neuronal loss in the substantia nigra and putamen (Horvath et al., 2014). A previous study reported that three out of four patients with late-onset AD with Parkinsonism carried the A allele of rs2298813 (Cuccaro et al., 2016). In a case report, 4 AD patients with Parkinsonism and psychiatric symptoms were found to have novel mutations in SORL1, with two mutations at the VPS10P region, where rs2298813 is located (Qiu et al., 2021). Our minor allele carriers also had higher scores on the Mayo fluctuation scale, implying a link with features of Lewy body dementia. Moreover, the minor allele of rs2298813 has been found to increase the risk of developing dementia in patients with Parkinson’s disease (Maple-Grodem et al., 2018). In the northern Chinese population, rs2298813 was associated with an increased risk of Parkinson’s disease (Wang et al., 2022). In addition to its involvement in the APP pathway, SORLA also mediates the trophic pathway involving glial cell line-derived neurotrophic factor (Glerup et al., 2013), the absence of which could lead to the loss of dopaminergic neurons (Lin et al., 1993).

There were several limitations of the current study. First, we did not systemically qualify and quantify the motor symptoms in these patients. Therefore, the postulated relationship between atrophy in the putamen and pallidum and pyramidal/extrapyramidal symptoms warrants further study. Second, these patients were followed up for an average of 2 years, so a longer follow-up duration may better confirm the cognitive or motor effects of the volumetric changes in the putamen and pallidum. Third, the sample size is much smaller in the minor allele carrier of rs2298813, so the absence of an association with cognitive test performance may be a consequence of low power. The inclusion of more patients with this genotype would help to confirm our findings.

The volume of the right lentiform nucleus was associated with cognitive function and clinical progression in late-onset AD patients with the homozygous major allele of rs2298813. Whereas, among the minor allele carriers, the volume of the right lentiform nucleus was smaller, and was not associated with cognitive function or clinical progression.

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the institutional review boards of Taipei Veterans General Hospital (IRB number 2012-05-033B). The patients/participants provided their written informed consent to participate in this study.

C-YC wrote the manuscript. J-LF, C-YC, Y-CL, and W-JL contributed to the study concept and design. C-YC, Y-SL, W-JL, AY, and J-LF contributed to analysis and interpretation of data. Y-SL, W-JL, Y-SK, and J-LF contributed to acquisition of data. W-JL, Y-CL, and J-LF revised the manuscript. All authors contributed to the article and approved the submitted version.

This study was supported by grants from the National Health Research Institutes, Taiwan (NHRI-11A1-CG-CO-05-2225-1), the Ministry of Science and Technology, Taiwan (MOST 109-2314-B-075-052-MY2, 110-2321-B-001-011-, 110-2321-B-A49A-502-, 110-2634-F-A49-005-), Taipei Veterans General Hospital (V110C-057, VGHUST110-G1-5-1, V111C-216) and the Brain Research Center, National Yang Ming Chiao Tung University from the Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan.