Epistatic effects of immunoglobulin KM (κ Marker) allotypes and APOE ϵ4 allele on the risk of Alzheimer’s disease

Abstract

We investigated whether immunoglobulin KM (κ marker) allotypes were associated with Alzheimer’s disease (AD) individually and/or epistatically with APOE ϵ4, the strongest known genetic susceptibility factor for the disease. Results showed a significant (p = 0.01) interaction between KM and APOE ϵ4. In KM 3/3 homozygotes, APOE ϵ4 was strongly associated with AD (OR = 8.3); however, in non-KM 3/3 subjects, the association between APOE ϵ4 and AD was markedly lower (OR = 0.9) and non-significant. If confirmed, these results may identify a subgroup of the population with a markedly higher risk of developing AD, who might require a different preventive/treatment strategy.

Introduction

Late-onset Alzheimer’s disease (AD) is the most common cause of dementia in older adults, and its prevalence is growing in our aging population. Both genetic and environmental factors play major roles in the etiopathogenesis of this disease. Numerous AD risk genes have been identified by genome-wide association studies (GWAS) and meta-analyses of AD, but most of the disease heritability remains unaccounted. At least a part of this “missing” heritability may be attributable to as-yet-unidentified susceptibility genes and to epistasis (gene-gene interaction), which is usually not measured in GWAS and other genomic studies. To our knowledge, possible role of immunoglobulin KM (κ marker) allotypes—genetic markers of κ light chains—in the etiopathogenesis of AD has not been investigated. There is a good rationale for the involvement of KM allotypes in AD pathogenesis, as these determinants influence immunity to hepatitis C virus, Epstein-Barr virus, and to Porphyromonas gingivalis,^1–4 all reported to be risk factors for AD and dementia.^5–8 Additional rationale is provided by the observation that the ratio of κ/λ light chains is higher in AD patients compared to controls,⁹ which could potentially be due to the influence of KM allotypes on the level of κ light chains. (No allotypes have been described for the λ light chains.)

The aim of the present investigation was to determine whether KM allotypes—individually and/or epistatically with APOE ϵ4—contributed to the risk of AD. APOE ϵ4 is the strongest known genetic susceptibility factor for AD.¹⁰

Methods

DNA samples

Coded DNA samples from patients and controls were obtained from the Emory University Alzheimer’s Disease Research Center Biorepository. All participants providing blood samples gave their informed consent following the protocols approved by the Institutional Review Board at Emory University. All subjects were White, and 40% were male. The mean age of both cases and controls was 74.2 years. Comprehensive cognitive assessments, including the Montreal Cognitive Assessment (MoCA), were administered to all patients as part of their evaluation at the Emory Cognitive Neurology Clinic and the Emory Goizueta Alzheimer’s Disease Research Center (ADRC). These samples were processed and stored in accordance with the 2014 ADC/NIA best practices guidelines.

Genotyping

Genotyping was done blinded to the case/control status of the samples.

APOE genotyping was performed previously using an Affymetrix Precision Medicine Array. The κ chain determinants KM 1 and 3 were characterized by a polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method.¹¹ Three alleles—KM 1, KM 1,2, and KM 3—segregate at the KM locus on chromosome 2. Over 98% of the people positive for KM 1 are also positive for KM 2.¹² KM 1, without KM 2, is extremely rare. In this investigation, positivity for KM 1 includes both KM 1 and KM 1,2 alleles.

Statistical analyses

Demographics and allelic distributions were compared between AD cases and controls using t-tests and chi square tests, as appropriate, depending on whether they were continuous or categorical variables. Due to relatively small sample size, genotypes were binarized based on the presence/absence of the minor allele. A logistic regression model was used to investigate the epistatic effects of KM allotypes and APOE ϵ4 associated with AD case/control status; the model also included the main effects for KM allotypes and APOE ϵ4. A sensitivity analysis was conducted in which all carriers of APOE ϵ2(n= 13) were excluded from the analysis, as APOE ϵ2 has been shown to be protective of AD.¹³

Results

Table 1 presents the descriptive statistics for the study population demographics and allelic distributions, stratified by case/control status. As expected, the frequency of APOE ϵ4 allele was significantly (p < 0.05) higher in AD patients compared to controls (71.6 versus 35.2%). The KM 1 allotype by itself was not significantly associated with the risk of AD; however, results of the logistic regression model (shown in Supplemental Table 1) noted that it showed significant (p = 0.01) interaction with APOE ϵ4 allele. Specifically, this model revealed that in KM 3/3 homozygotes, APOE ϵ4 was strongly associated with AD (OR = 8.3; 95% CI: 4.0–17.1); however, in non-KM 3/3 (KM 1/1 +KM 1/3) subjects, the association between APOE ϵ4 and AD was markedly lower and non-significant (OR = 0.9; 95% CI: 0.2–4.1). Figure 1 shows the proportion of AD cases and controls that have each of the four genotype combinations. The proportion with both KM 1 and APOE ϵ4 absent (light green) was substantially larger in the controls compared with the AD cases (58.2% versus 18.2%), while the proportion with KM 1 absent and APOE ϵ4 present (orange) was much smaller in the controls compared with the AD cases (25.5% versus 65.9%).

Table 1.

Descriptive statistics for the study population, stratified by case/control status.

Variable	Statistic	AD Cases (n = 100a)	Controls (n = 100a)
Age (y)	Mean (SD)	74.2 (5.6)	74.2 (5.6)
Sex	n, % male	40/100 (40.0)	40/100 (40.0)
Race
White	n/d (%)	100/100 (100.0)	100/100 (100.0)
Other	n/d (%)	0/100 (0.0)	0/100 (0.0)
KM 1/3 (allele 1 present)	n/d (%)	17/100 (17.0)	16/99 (16.2)
APOE (allele ϵ4 present)	n/d (%)	63/88 (71.6b)	32/99 (35.2b)

n: numerator; d: denominator.

^aOccasionally a subject could not be genotyped, meaning that the denominators for some of the percentages were not always n = 100.

^bIndicates a significant (p < 0.05) difference between cases and controls by chi-square test.

Figure 1.

The pie charts illustrate the proportion of AD cases and controls that have each of the four genotype combinations. Sample sizes for each combination are listed within each of the respective pie “wedges”. The proportion with both KM 1 and APOE ϵ4 absent (light green) was substantially larger in the controls compared with the AD cases (57/98 [58.2%] versus 16/88 [18.2%]), while the proportion with KM 1 absent and APOE ϵ4 present (orange) was much smaller in the controls compared with the AD cases (25/98 [25.5%] versus 58/88 [65.9%]). The proportion with KM 1 present and APOE ϵ4 absent (pink) was similar between controls and AD cases (10/98 [10.2%] versus 9/88 [10.2%]), as was the proportion with KM 1 present and APOE ϵ4 present (light blue) (controls: 6/98 [6.1%] versus AD cases: 5/88 [5.7%]).

Discussion

These results are intriguing and hitherto unreported. They seem to suggest that KM 1 allotype is a potential genetic modifier of APOE ϵ4, and that it somehow attenuates the strong APOE ϵ4-linked risk of AD. Immunoglobulin κ light chains, which express KM allotypes, may be relevant to neurodegeneration in AD, as the ratio of κ/λ light chains is higher in AD patients compared to controls.⁹ Although it has not been investigated yet, it is possible that KM allotypes contribute to the level of κ light chains, just as GM (γ marker) allotypes contribute to the levels of γ chains.¹⁴ Thus, KM alleles may differentially regulate the level of kappa light chains in AD patients and controls, which would influence the κ/λ ratio. Additional rationale for the KM allotype involvement in AD pathogenesis is provided by the RNA sequencing analysis of postmortem brain regions. A highly significant upregulation of immunoglobulin kappa constant (IGKC) gene—which encodes the KM allotypes—in the precuneus of AD brains compared to matched healthy brains has been reported.¹⁵

The mechanisms underlying the interactive effects of KM 1 and APOE ϵ4 on AD risk are not clear, but one can speculate on a possible mechanism based on the known functions of these genes. One mechanism through which APOE ϵ4 enhances the risk of AD is by increasing the deposition of amyloid-β (Aβ), the hallmark of AD.¹⁰ KM allotypes have not yet been investigated for their possible role in the immunobiology of Aβ, but they have been shown to influence autoantibody responses to several other autoantigens,^16–19 including neuroantigens.²⁰ It is possible that KM 1 allotype is associated with higher levels of anti-Aβ antibodies, which would form immune complexes with Aβ. These immune complexes would be phagocytosed by microglia, thus attenuating the APOE ϵ4-linked risk of AD.

The results may also shed light on the racial differences in the magnitude of APOE ϵ4-related risk of AD, which is higher in White people than in Black people and in American Indians.¹⁰ Interestingly, the frequency of KM 1 allotype is significantly higher in the latter two groups than in White people.²¹ Therefore, it is tempting to speculate that the lower APOE ϵ4-linked risk in Black people and in American Indians may partially be due to the attenuating influence of KM 1 allotype. However, this does not explain higher APOE ϵ4-linked risk in East Asians than in White people,¹⁰ as the former have a higher frequency of the KM 1 allotype.²¹ A study involving a large multiethnic study population is needed to confirm and extend the observations reported here.

This investigation has some limitations—a relatively small sample size and only White subjects in the study population. Since there are racial/ethnic differences in gene frequencies, the findings presented here may not be generalizable to other racial groups. Additionally, a thorough investigation of the association between APOE ϵ2, AD, or KM 1 was beyond the scope of this study given that there were only n = 13 APOE ϵ2 carriers (3 AD cases and 10 controls) in this study population. However, in the sensitivity analysis in which the APOE ϵ2 carriers were excluded, the interaction between KM1 and APOE ϵ4 remained statistically significant (p = 0.01).

Supplementary Material

Supplemental material for this article is available online.

Data availability statement

Data are available from the corresponding author after a reasonable request.