rs2275780 — APH1A APH1A Promoter -21C/A

APH1A -21C/A — A 5' UTR Variant at the Edge of the Gamma-Secretase Control Room

The gamma-secretase complex¹¹ gamma-secretase complex
a multi-protein intramembrane protease that cleaves over 90 substrates including APP and Notch receptors is assembled from four essential subunits: presenilin (the catalytic core), nicastrin, PEN-2, and APH-1. Of the APH-1 paralogs, APH1A²² APH1A
anterior pharynx-defective 1 homolog A, encoded on chromosome 1q21.2 is the dominant form — expressed broadly across all tissues, essential for embryonic development, and required for full Notch signalling throughout life. Without APH-1A, the gamma-secretase complex cannot fully assemble, and both Notch-dependent cell fate decisions and amyloid precursor protein (APP) cleavage are compromised.

The variant rs2275780 sits in the 5' untranslated region of APH1A, just 21 base pairs upstream of the translation start codon (position c.-20 in coding-strand notation). Because APH1A is transcribed from the minus strand, papers describe the change as -21C→A (coding strand), which corresponds to a G→T substitution on the genomic plus strand at position chr1:150,268,830 (GRCh38). This proximity to the AUG start codon places it in a region capable of influencing mRNA secondary structure, ribosome scanning efficiency, or upstream open reading frame activity — mechanisms that can modestly alter protein output without changing the protein sequence itself.

The Mechanism

rs2275780 is annotated as a 5' UTR variant in APH1A coding transcripts. Its position at c.-20 relative to AUG means it falls within the Kozak-like region immediately flanking the start codon, where single-nucleotide differences can alter translation initiation efficiency. However, no functional reporter assay, ribosome profiling, or quantitative APH-1A expression study has specifically evaluated this variant's effect on APH-1A output. The nearby and biologically more characterised promoter variant rs3754048 (-980C/G) has been shown to alter binding of the YY1 transcription factor, increasing APH-1A transcriptional activity 2.7-fold in reporter assays³³ increasing APH-1A transcriptional activity 2.7-fold in reporter assays
Qin et al. Aging Cell 2011, elevating gamma-secretase activity and Aβ42 generation. Whether rs2275780 contributes any additive translational effect on top of transcriptional variation from rs3754048 has not been studied.

The biological stakes of APH-1A dosage are real: APH1A homozygous deletion is embryonic lethal in mice⁴⁴ APH1A homozygous deletion is embryonic lethal in mice
Serneels et al. PNAS 2005, producing angiogenesis defects, neural tube malformations, and somitogenesis failure — phenotypes characteristic of severe Notch signalling loss. In the adult brain, APH-1A-containing gamma-secretase complexes are the predominant form and are required for Notch-mediated neuronal differentiation and maintenance. By contrast, APH1B-containing complexes disproportionately produce longer Aβ species (Aβ42, Aβ45) compared to APH1A complexes⁵⁵ APH1B-containing complexes disproportionately produce longer Aβ species (Aβ42, Aβ45) compared to APH1A complexes
Serneels et al. Science 2009, which suggests that shifts in the APH-1A/APH-1B subunit balance influence the amyloidogenic output of the complex independent of total gamma-secretase activity.

The Evidence

The only study directly examining rs2275780 in a human disease context is Wang & Jia 2009⁶⁶ Wang & Jia 2009
Neuroscience Letters 2009, which genotyped 256 sporadic Alzheimer's disease (SAD) patients and 276 normal controls from the North Chinese Han population. Two APH-1a promoter variants were examined together: -980C/G (rs3754048) and -21C/A (rs2275780). The result was unambiguous: "there was no statistical difference between the SAD group and controls regarding the frequency of alleles and genotypes of -21C/A," whether or not the analysis was stratified by APOE ε4 status. The -980C/G variant showed a significant association (allele P = 0.01; genotype P = 0.038); rs2275780 did not.

This finding was not replicated or refuted in subsequent literature. The rs2275780 entry in dbSNP has a single publication citation (the Wang & Jia paper), no ClinVar submission, and no GWAS Catalog hits. The variant is absent from the GWAS Catalog for Alzheimer's disease, cognitive decline, and related traits as of 2026.

The East Asian T allele frequency (38%) is strikingly higher than in African populations (<2%), with Europeans at approximately 14%. This large population stratification means that any future association study would need to control carefully for ancestry, as the T allele could appear spuriously protective or harmful depending on population composition.

Practical Actions

For individuals carrying one or two T alleles at rs2275780, the current evidence does not support any specific intervention. The variant has not been shown to increase Alzheimer's disease risk, alter gamma-secretase activity measurably, or affect any other clinical outcome in published research. No clinical guidelines reference this variant, and it is not included in polygenic risk scores for neurodegeneration.

The broader APH1A context — its central role in Notch signalling, amyloid processing, and neuronal maintenance — does provide a biologically coherent rationale for monitoring neurological health over time, particularly given any family history of dementia. However, this applies to all individuals regardless of rs2275780 genotype.

Interactions

rs2275780 was identified in the same sequencing study as rs3754048 (APH1A -980C/G), the functional risk variant. These two variants are in the same gene regulatory region and were studied together in Wang & Jia 2009. Whether they are in linkage disequilibrium in the North Chinese population studied, or independently segregate, was not reported. Any future compound action between these two variants would require data on their haplotype frequencies and combined functional effects, which currently do not exist in the literature.

The APH1A/APH1B subunit balance (rs2275780 in APH1A vs. APH1B-specific variants such as rs34714364) is a biologically plausible interaction axis: shifts in subunit composition alter the Aβ42/Aβ40 ratio independent of total gamma-secretase activity. This interaction has mechanistic support from Serneels et al. 2009 but has not been characterised at the human SNP level.