rs3846663 — HMGCR HMGCR Exon 13 Haplotype Tag

HMGCR rs3846663 — When Your Cholesterol Gene Edits Its Own Instructions

Every statin ever prescribed targets one protein: HMG-CoA reductase (HMGCR), the rate-limiting enzyme of the cholesterol biosynthesis pathway. What most people don't know is that the HMGCR gene produces two versions of its own enzyme — a full-length, fully active form and a shorter variant missing 53 amino acids from the catalytic domain, encoded by a transcript that skips exon 13. How much of each version your cells make is under tight genetic control, and rs3846663 tags the haplotype that governs that balance.

The Mechanism

rs3846663 sits in an intron of HMGCR on chromosome 5 (GRCh38 position 75,359,901). It is not itself the functional variant, but serves as a proxy — in tight linkage disequilibrium¹¹ linkage disequilibrium
LD: two variants so commonly inherited together that one reliably predicts the other (r²=0.82–0.93) with rs3846662, a SNP located 47 base pairs downstream of exon 13 in intron 13.

The rs3846662 variant directly regulates the splicing factor HNRNPA1²² HNRNPA1
a nuclear RNA-binding protein that promotes skipping of entire exons during mRNA processing. The rs3846662 A allele disrupts an SRSF1 binding site while preserving HNRNPA1 binding, causing the spliceosome to skip exon 13 more frequently. The resulting HMGCRΔ13 transcript, lacking 53 amino acids in the catalytic domain, encodes an enzymatically inactive HMG-CoA reductase.

The T allele at rs3846663 travels with the rs3846662 G allele — the allele that promotes exon 13 retention, yielding more full-length, catalytically active HMGCR. More active enzyme means more cholesterol is synthesized endogenously, and circulating LDL-C levels rise correspondingly. In vitro studies in human lymphoblastoid cells confirm that homozygosity for the minor/G haplotype is associated with up to 2.2-fold lower expression of the alternatively spliced HMGCRΔ13 transcript.

The Evidence

The founding GWAS was conducted by Burkhardt et al. (2008)³³ Burkhardt et al. (2008) in the isolated Kosraean population of Micronesia (n=2,346) and validated in the Diabetes Genetics Initiative Caucasian cohort (~2,758 subjects). rs3846663 reached genome-wide significance for LDL-C association (combined P=7×10⁻¹⁰). Three SNPs in tight LD — rs7703051, rs12654264, and rs3846663 — all met the genome-wide threshold, with a functional splicing experiment in lymphoblast minigene systems confirming that the linked rs3846662 directly controls exon 13 inclusion.

The functional consequence for statin response was characterized in the CAP study⁴⁴ CAP study
Medina et al. Circulation 2008: Alternative splicing of HMGCR is associated with plasma LDL-C response to simvastatin (Cholesterol and Pharmacogenetics, n=170 lymphoblastoid cell lines from simvastatin-treated individuals). The alternatively spliced HMGCRΔ13 transcript explained 6–15% of the variance in statin-induced LDL-C reduction — substantially more than genotype alone (<2%). A higher proportion of the Δ13 transcript (driven by the A-haplotype protective alleles) meant the cells already had less active HMGCR before statin exposure, leaving proportionally less enzyme activity to inhibit.

Sex-dependent effects emerged in a study of French-Canadian familial hypercholesterolemia patients by Tremblay et al. (2016)⁵⁵ Tremblay et al. (2016). Women homozygous for the rs3846662 AA genotype (the exon-skipping-promoting haplotype, which corresponds to the CC genotype at rs3846663) showed significantly greater LDL-C reduction on statins (46.2% vs 38.4% in AA women), while no significant sex-specific difference was observed in men. This suggests the exon 13 splicing mechanism interacts with hormonal physiology in ways that modify statin pharmacodynamics.

At the population level, the landmark Ference et al. NEJM study (2016)⁶⁶ Ference et al. NEJM study (2016) used HMGCR variant scores across 112,772 participants (14 studies, 14,120 CVD events) to show that genetically lower LDL-C through HMGCR variants reduces cardiovascular risk with an OR of 0.81 (95% CI 0.72–0.90) per 10 mg/dL reduction — essentially identical to the observed benefit of pharmacological statin therapy — confirming the LDL-causal model.

Practical Actions

Carriers of the T allele (CT or TT genotype at rs3846663) have relatively more full-length HMGCR expression and correspondingly higher baseline LDL-C. This does not mean statins don't work — they do, but the mechanism differs. With less pre-existing Δ13-driven enzyme suppression, T allele carriers may require higher statin doses to achieve the same LDL-C reduction as CC individuals. Monitoring LDL-C response at 6–8 weeks after statin initiation and adjusting dose to target is therefore especially valuable for TT homozygotes.

CC homozygotes tend to have lower baseline LDL-C and may achieve larger proportional LDL-C reductions on statins, particularly women, due to the amplified exon-skipping effect. Standard statin dosing is generally adequate.

Interactions

rs3846663 is a proxy for the rs3846662-defined H7 haplotype (rs17244841 / rs3846662 / rs17238540), which is the core functional haplotype characterized in statin response studies. Compound analysis across HMGCR haplotypes (H2 + H7 together) shows further attenuation of statin response compared to either haplotype alone, suggesting additive effects of multiple HMGCR splicing-regulatory SNPs. A gene-gene interaction between the LIPC locus (rs1532085) and the HMGCR region has been reported in multi-ethnic data, explaining an additional 0.2–1.1% of HDL-C variance beyond either SNP alone.