Association of Butyrylcholinesterase-K Allele and Apolipoprotein E ɛ 4 Allele with Cognitive Decline in Dementia with Lewy Bodies and Alzheimer’s Disease

INTRODUCTION

Cholinergic disturbances play an important role in several types of dementia, including Alzheimer’sdisease (AD) and dementia with Lewy bodies (DLB) [1]. The cholinergic enzymes acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) catalyze the hydrolysis of the neurotransmitter, acetylcholine (ACh) [2]. Synaptic membrane-bound AChE is more important in regulating cholinergic neurotransmission, while BuChE plays a more important role in the control of extracellular ACh levels that modulate the functional properties of glial cells [3–5]. In the cerebrospinal fluid (CSF) of patients with mild AD, BuChE activity is associated with CSF markers of astroglial reactivity and function such as glial fibrillary acidic protein (GFAP) and S100B [3]. Higher BuChE activity is associated with higher levels of these astroglial reactivity markers, with lower levels of proinflammatory cytokines and higher complement factor levels [3].

The butyrylcholinesterase gene (BCHE) is located in the long q-arm of chromosome 3 [6, 7] and over 40 mutations have been identified. The BCHE-K variant is the most common polymorphism of BCHE, which is resulted from substitution of an alanine residue at codon 539 to threonine (A539T). Other mutations of BCHE gene includes the atypical, J, H, and silent variants. Studies have shown that the distribution and catalytic properties of this enzyme in humans depend on the type of BCHE variant present [8–10]. For instance, BCHE-K produces a BuChE protein variant with approximately 30% lower enzymatic activity compared to the wild type [11].

The activity of BuChE in CSF is reduced in APOE ɛ4 carriers with mild AD [12]. This reduction in APOE ɛ4 carriers is more marked in those who are also BCHE-K carriers, and this group of patients in particular exhibited a BCHE-K allele-dose dependent reduction in BuChE activity. In contrast, APOE ɛ4 non-carriers show higher and similar CSF BuChE enzymatic activity regardless of BCHE-K carrier status [12]. Interestingly, there is a BCHE-K allele-dose dependent decrease in the markers of glial reactivity (GFAP and S100B), which are found mainly in APOE ɛ4 carriers [3].

The BCHE-K variant has also been associated with reduced phosphorylation of tau in patients with dementia [13]. In confirmed late-onset (>65 years of age) AD, a greater load, both of overall neuritic plaques and of cholinesterase-positive neuritic plaques, has been demonstrated in the temporal cortex of BCHE-Kcarriers, who were <80 years of age than of all other patients [14]. Interestingly, a mutation in the 5’untranslated region of the BCHE gene (rs509208) in a genome-wide association studies investigation of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) mild cognitive impairment (MCI)/mild AD cohort was shown to be an independent and substantial contributor to cortical fibrillar Aβ burden in humans [15]. Together, APOE and BCHE loci explained 15% of the variance in cortical Aβ levels (APOE 10.7% ,BCHE 4.3%) [15].

Nonetheless, the results of studies regarding the frequency of BCHE-K carriers in dementia are inconsistent, likely due to variation by geographical region and small sample sizes. The allele is generally found in up to one third of Caucasians and Asians [16, 17]. Although some studies have found evidence of a protective effect of the BCHE-K allele on the likelihood of developing AD [18–20], other studies have suggested that the BCHE-K allele increases the risk of developing AD in the presence of an APOE ɛ4 allele [21–24] while decreasing the risk in non- APOE ɛ4 carriers [25, 26]. A high frequency of BCHE-K has been observed in DLB [27]. In addition, significantly higher frequencies of BCHE-K, APOE ɛ4 alone and both alleles together have been seen in mild to moderate DLB compared to patients with mild to moderate Parkinson’s disease with dementia (PDD) [28], particularly in patients aged <75 years.

The APOE ɛ4 allele is associated with a more rapid rate of cognitive decline and hippocampal atrophy before the clinically symptomatic phase of dementia, in amnestic MCI and in mild AD [29, 30]. The influence of APOE ɛ4 on the rate of progression of the disease diminishes with advanced age and progression of AD [28, 31]. Studies suggest that the BCHE-K variant [32, 33] is associated with slower cognitive decline. In longitudinal studies in amnestic MCI, mild to moderate PDD and mild AD individuals with BCHE-K without an APOE ɛ4 allele generally had the slowest cognitive decline, in contrast to the most rapid decline in those with both BCHE-K and APOE ɛ4 alleles [21, 34]. However, in moderate AD the rates of decline in these genotype groups seems to be reversed [34].

Few studies have examined the role of APOE ɛ4 on cognitive decline in DLB and none in a routine clinical setting, where patients are receiving standard of care treatment. Similarly, the combined effect of APOE ɛ4 and BCHE-K genotypes on cognitive decline in dementia, again in a routine clinical setting where patients may be receiving cholinesterase inhibitors and other treatment has not been examined.

In the present study, we wished to examine the distribution of BCHE-K in a relatively large DLB cohort, and compare the findings with AD and healthy controls. We also explored the association of BCHE-K and APOE ɛ4 genotypes with the rate of cognitive decline in a subgroup of patients in a routine clinical setting, who were followed for five years.

MATERIALS AND METHODS

RESULTS

DISCUSSION

In this study we present the frequency of theBCHE-K genotype and its clinical correlates in routineclinical practice patients with a diagnosis of DLB or AD. One of the main findings was that the BCHE-K genotype was less common among patients with DLB than in control subjects. However, the frequency of BCHE-K genotype in the control group was very high (51.2%) compared to the albeit variable frequency of 33% found in previous studies [17]. We checked this further with another Norwegian cohort (DemGene) showing 41% , which still seems higher than expected frequency of 33% in the control population of other studies [41]. This might suggest that a high BCHE-K frequency among Norwegian population, which warrants further studies.

In these patients who were receiving standard ofcare treatment, including cholinesterase treatment in most cases, no statistically significant associations between BCHE-K genotype and the rate of cognitive decline were observed. In contrast, we found that the APOE ɛ4 genotype was associated with more rapid cognitive decline among the patients, which is in agreement with similar longitudinal associations observed in cognitively normal elderly subjects [42, 43].

Insufficient studies regarding BCHE-K genotype frequency among patients with DLB exist. In one of the largest studies, the frequency of BCHE-K genotype in DLB and PDD was 47.4 and 34.4 % , respectively [28]. The study also showed that the proportion of patients carrying both BCHE-K and APOE ɛ4 in these two Lewy body disorders was significantly different in association with APOE ɛ4, at 28.1% and 11.5% , respectively. However, the study had no control group for comparison. In this study, BCHE-K genotype frequency was 33.9% in the DLB group, while the frequency of the combined APOE ɛ4/BCHE-K genotype was 17% in the DLB and 14% in the control group. Also considering the high frequency of the BCHE-K carriers without an APOE ɛ4 allele in the control group, a plausible explanation for the lower proportion of BCHE-K carriers without an APOE ɛ4allele among DLB patients could be due to expected differences in the phenotypic activity of BuChE-K variant protein. This protein variant is associated with 30–50% reduction in acetylcholine-degrading capacity in CSF [1], which could result in more efficient central cholinergic signaling, masking ongoing pathology in the cholinergic neural network. This might prolong the time for subjects to reach the symptom threshold for dementia. This might explain both the reported association in DLB between BCHE-K and better cognitive performance on measure of attention that implicates subcortical structures [33]. Thus, a reduced likelihood of developing dementia and/or a delay in the onset of symptoms of dementia in those with a BCHE-K allele in the absence of an APOE ɛ4 allele could be the reason for the higher BCHE-K frequency observed in healthy controls compared to DLB in our study.

In another study in patients with mild to moderate PDD, carriers of both APOE ɛ4 and BCHE-K exhibited a significantly more rapid cognitive decline than the patients carrying only one of these genotypes [28], indicating a potential interaction between the two genotypes similar to that seen in amnestic MCI [21]. In the present study, although the overall analysis did not indicate an interaction between the two genotypes, the explorative repeated-measures analysis performed on the subpopulation of the patients who completed the full 5-year follow-up period, supported the possibility that, the combined APOE ɛ4 and BCHE-K genotypes may be related to a higher rate of cognitive decline. On the other hand, in the absence of an APOE ɛ4 allele, the BCHE-K genotype may reduce the rate of cognitive decline (Fig. 2). Since this finding was based on analysis of small groups, it should currently be interpreted with caution. However, this study suggests— that over a 5-year period— standard of care therapy may have little effect on modulating the interaction of genotype on cognitive decline in patients with DLB and AD.

Moreover, the interaction between APOE ɛ4 and BuChE may be of particular interest. We have shown that in AD patients, the APOE ɛ4 genotype modulates the phenotypic expression of the BCHE-K variant, which leads to an exaggerated reduction in BuChE activity in BCHE-K carriers [12]. In the presence of APOE ɛ4 therefore, these subjects are expected to accumulate more Aβ pathology and synaptic dysfunction due to hypofunctional glial [5] and thereby reduced Aβ clearance. This is indirectly supported by 1) our finding here that among subjects aged <75, 27% of DLB and 37% of AD patients had combined APOE^ɛ4 +|BCHE^K + genotype compared to 17% among the controls (Table 4), together with 2) immunohistochemical analyses of AD brain showing about six folds accumulation of ChE-positive Aβ deposits in K-carriers patients aged less than 80 years [14, 44], and 3) enzymological analyses of cortical brain homogenate showing 37% significant increase of BuChE level in K homozygotes than wild-type homozygotes in dementia patients [45].

In addition the accelerated debut of clinical symptoms due to synaptic ACh signalin deficits in carriers of both of these alleles may lead to severe cholinergic neurodegeneration and early onset of amnestic symptoms [21], and so they may show an initial robust symptomatic response to cholinesterase inhibitiontherapy in MCI and mild AD, but poor response at later stages of the disease as increasing damage to cholinergic synapses renders them non-functional [5]. This may explain our observation that patients with both the APOE ɛ4 and BCHE-K have a more rapid cognitive decline [28].

As was mentioned, we found that the frequency of BCHE-K is more common in controls than in people with DLB, suggesting that the BCHE-K variant may act as a protective factor against DLB, particularly in homozygotes. However, our finding did not support previous reports of a slower rate of decline in BCHE-K carriers, as this may only be evident in patients with early disease in those without a concomitant APOE ɛ4 allele. However, exploratory analyses suggested a possible association between the BCHE-K variant and the APOE ɛ4 allele and more rapid cognitive decline over a 5-year period despite the use of cholinesterase treatment. Future longitudinal studies on the impact of BCHE-K and other BCHE mutations on the occurrence and course of AD and DLB should control for recruitment biases; the influence of gender, age, and stage of disease; and include larger sample sizes and CSF assessments of BuChE activity, complement, markers of glial activity, and AD biomarkers.

Limitations include that the DLB patients were recruited at two clinical centers, while the AD and healthy controls were recruited at one of the centers, which might have influenced the findings. Even though the frequency of BCHE-K carriers did not show statistical differences between the centers, the frequency of BCHE-K carriers was numerically higher in DLB patients from TrønderBrain compared to the DemVest DLB patients. The rate of cognitive decline in AD is strongly influenced by gender, genotype, diagnosis, stage of dementia, and cholinesterase treatment [5, 46] and due to the small sample size, it is not possible to fully disaggregate the contribution of these factors in patients with DLB in the current study. However, the use of multiple linear mixed effect modeling allowed us to adjust for the effects of age and gender, and nearly all patients in the longitudinal study were treated with a cholinesterase inhibitor from baseline. Another potential limitation is cohort bias effect, whereby slow progressing patients are more likely to enroll in a cohort study and complete follow-up visits than fast progressing patients. This may have influenced the distribution of genotypes in the study and underestimated the proportion of individuals carrying both BCHE-K and APOE ɛ4 alleles who develop DLB and AD. In addition, most DLB patients have some degree of AD-type pathology, and it has been suggested that DLB patients with BCHE-K and APOE ɛ4 alleles may have a more AD-like phenotype [28]. Although the number of cases in the present work is relatively large for a DLB group, it is still small for this kind of analysis. The number of cases in the AD group is also small. The BCHE-K genotyping was performed using two separate SNP preparations, which adds to the strength of this study.

Diagnoses were based on clinical assessment, and thus misdiagnosis cannot be excluded. However, the most common misclassification in DLB is AD, and given the relatively similar genotype distributionmisdiagnosis is unlikely to influence the findings. Misclassification of healthy and diseased patients is unlikely given the large difference in cognition between the groups, the comprehensive and standardized diagnostic procedures, including frequent availability of results from DaTScan and autopsy in a subset from the DemVest study. The study also has the advantage of a longitudinal design for a subset of individuals, up to five years.