rs17482078 — ERAP1 R725Q

ERAP1 R725Q — When a Broken Molecular Scissors Cuts Both Ways

Every cell in your body runs a continuous self-inventory. Proteins are broken down, and peptide fragments are trimmed to the exact 8–9 amino acid length required for MHC class I¹¹ MHC class I
The molecular display platform on every nucleated cell that presents peptide samples to patrolling CD8+ T cells, allowing the immune system to distinguish healthy "self" cells from infected or cancerous ones loading. ERAP1 is the trimming enzyme — an endoplasmic reticulum aminopeptidase that shaves amino acids off the N-terminus of longer precursor peptides until they are the right size. The rs17482078 variant changes arginine to glutamine at position 725, in the peptide-binding domain of the enzyme. That single substitution is one of the five amino acid changes that together define Haplotype 10 (Hap10)²² Haplotype 10 (Hap10)
One of ten major ERAP1 haplotypes. Hap10 is the low-activity ERAP1 variant, distinguished by five non-ancestral amino acids at positions 349, 528, 575, 725, and 730 — the ERAP1 variant found in roughly 22% of humans that has dramatically lower enzymatic activity than the ancestral form.

What makes this variant biologically remarkable is the direction-of-effect paradox. The same molecular defect — reduced peptide trimming — protects against one class of autoimmune disease (ankylosing spondylitis and psoriasis) while predisposing to another (Behçet's disease). The reason is that the protective or harmful consequence depends entirely on which HLA class I allele the improperly trimmed peptides are loaded onto.

The Mechanism

ERAP1's peptide-binding domain (Domain IV) holds the incoming peptide substrate while the zinc-active site in Domain II cleaves its N-terminal residue. Residue 725 sits in Domain IV and helps position the substrate. The R725Q substitution — replacing the positively charged arginine with the uncharged glutamine — alters the electrostatic environment of the binding pocket, contributing to the dramatically reduced activity that characterises Hap10³³ Hap10
Biochemically characterised in Frontiers Immunology 2024: Hap10 is not simply inactive but has distinct substrate specificity, being up to 60-fold less active on small peptides but competent at trimming larger ones.

In HLA-B27-positive individuals, oversized (undertrimmed) peptides cannot stably bind HLA-B27, reducing the availability of the aberrant HLA-B27 peptide complexes⁴⁴ HLA-B27 peptide complexes
HLA-B27 is thought to promote ankylosing spondylitis partly through misfolded free heavy chains and partly through presentation of arthritogenic peptides to CD8+ T cells; undertrimmed peptides reduce both mechanisms hypothesised to drive AS inflammation. The T allele therefore protects against AS — but only in HLA-B27 carriers.

In HLA-B*51-positive individuals, the opposite happens. HLA-B*51 appears to require a different peptide repertoire for stable surface expression, and the undertrimmed peptides generated by Hap10 fit HLA-B*51's binding groove unusually well. The result is an altered immunodominance hierarchy for CD8+ T cells, which reacts against a different set of "self" antigens⁵⁵ reacts against a different set of "self" antigens
Kirino et al. 2022: Hap10/HLA-B51 combination generates undertrimmed peptides that shift the CD8 T cell response toward a BD-associated immunodominance pattern; the response is qualitatively different, not simply amplified and contributes to Behçet's disease pathogenesis.

The Evidence

The ankylosing spondylitis finding was established by Evans et al. in Nature Genetics (2011)⁶⁶ Evans et al. in Nature Genetics (2011) in a study that analysed the ERAP1 locus in 3,023 AS cases and 8,779 controls. The key result was that the ERAP1 association completely disappeared in HLA-B27-negative individuals (interaction P=7.3×10⁻⁶), demonstrating that ERAP1 variants influence AS risk exclusively through their effect on HLA-B27 biology. The rs17482078 T allele showed a protective OR of approximately 0.73 in the HLA-B27+ subset. A subsequent meta-analysis of six studies (Lee et al. 2011)⁷⁷ meta-analysis of six studies (Lee et al. 2011) aggregating 4,594 AS cases and 3,971 controls confirmed the T allele's protective effect in European populations (OR=0.726, 95% CI 0.655–0.805, P<1×10⁻¹⁰).

The Behçet's disease story was resolved by a landmark GWAS by Kirino et al. in Nature Genetics (2013)⁸⁸ GWAS by Kirino et al. in Nature Genetics (2013) that genotyped 779,465 SNPs in 1,209 Turkish BD patients and 1,278 controls, with independent replication in Turkish and Japanese samples. Two ERAP1 coding variants — rs10050860 (D575N) and rs17482078 (R725Q) — recessively conferred BD risk. In the combined cohort, the rs17482078 TT genotype had an OR of approximately 1.6 overall, but in HLA-B*51-positive individuals the OR rose to 3.78 (95% CI 1.94–7.35), with the HLA-B*51/ERAP1 interaction reaching P=9×10⁻⁴. A follow-up study Kirino et al. 2016⁹⁹ Kirino et al. 2016 narrowed the ERAP1 signal to a specific protein allotype — essentially Hap10 — as the single dominant risk factor for BD in HLA-B*51 carriers.

Practical Implications

For CC homozygotes (ancestral, higher-activity ERAP1), the clinical picture is neutral to mildly elevated AS risk in HLA-B27 carriers and no elevated Behçet's risk.

For CT heterozygotes, the enzyme activity is intermediate — one Hap10 allele contributes to the peptide pool but the ancestral allele's output predominates. Risk modulation is present but attenuated relative to TT homozygotes.

For TT homozygotes, the fully Hap10 state means the entire cellular ERAP1 output runs through the low-activity, altered-specificity enzyme. This is the genotype most worth acting on: substantially reduced AS risk in HLA-B27 carriers, but meaningful BD risk elevation, especially in HLA-B*51 carriers. Behçet's disease is rare in populations of non-Mediterranean/Middle Eastern ancestry (~1:10,000–1:300,000 globally), but among individuals of Turkish, Iranian, or Japanese ancestry it reaches prevalence of 1:1,000 or higher.

Interactions

The most important interaction is with HLA-B*51 for Behçet's disease risk and with HLA-B27 for ankylosing spondylitis risk. In both cases the ERAP1 rs17482078 effect is absent (or reversed) in the HLA-negative population and fully expressed only in the HLA-positive population — among the clearest examples of genetic epistasis in human immunogenetics.

Within ERAP1 itself, rs17482078 acts in concert with rs30187 (K528R), rs10050860 (D575N), rs27044 (Q730E), and rs2287987 (M349V) to define the Hap10 haplotype. The disease risk attributable to rs17482078 alone is largely mediated through its position within this haplotype combination: the five-variant Hap10 confers a fundamentally different enzyme than any single change would.