rs56038477 — DPYD HapB3 tag (c.1236G>A, E412E)

DPYD c.1236G>A — The Synonymous Tag SNP for the HapB3 Haplotype

DPYD encodes dihydropyrimidine dehydrogenase (DPD), the rate-limiting enzyme¹¹ rate-limiting enzyme
DPD catabolizes 80-90% of administered 5-fluorouracil into inactive metabolites, preventing toxic accumulation in the bone marrow and gastrointestinal tract that breaks down fluoropyrimidine chemotherapy drugs. rs56038477 is a synonymous variant in exon 11 of DPYD (c.1236G>A, p.Glu412=) that, on its own, has no effect on DPD enzyme activity — it changes the DNA sequence but not the protein. Despite this, it is one of the most clinically important DPYD variants because it serves as the standard tag SNP²² tag SNP
A nearby variant in linkage disequilibrium with a functional variant, used as a cheap, easily-genotyped proxy when the true causal variant is harder to detect for the DPYD HapB3 haplotype, whose functional driver is the deep intronic splice variant c.1129-5923C>G (rs75017182).

The Mechanism — Why a Silent Variant Matters

c.1236G>A substitutes a glutamic-acid-coding codon (GAG) with another glutamic-acid-coding codon (GAA) at position 412 of the DPD protein. Because both codons specify glutamic acid (Glu, E), the substitution is synonymous — the resulting DPD enzyme is structurally and functionally identical to the wild-type protein. Synonymous variants typically have no clinical impact (occasional exceptions involve codon usage or mRNA stability effects, neither of which applies here for DPD).

What makes rs56038477 clinically important is its linkage disequilibrium³³ linkage disequilibrium
Two variants are in LD when they co-occur on the same chromosome more often than expected by chance, usually because they arose on the same ancestral haplotype and haven't been separated by recombination with rs75017182, the deep intronic c.1129-5923C>G variant that actually causes decreased DPD activity. The two variants sit within a few kilobases of each other in DPYD's intron-10/exon-11 region and arose together on a single European ancestral haplotype, travelling together ever since. When van Kuilenburg et al. characterised the HapB3 haplotype in 2010⁴⁴ van Kuilenburg et al. characterised the HapB3 haplotype in 2010
Intragenic deletions and a deep intronic mutation affecting pre-mRNA splicing in the dihydropyrimidine dehydrogenase gene. Hum Genet 2010;128(5):529-38, they found that every HapB3 patient carried both variants on the same chromosome, and functional studies isolated the splicing defect to the intronic variant alone.

This matters because the intronic c.1129-5923C>G is technically harder to genotype⁵⁵ technically harder to genotype
Deep intronic positions are rarely included on consumer SNP arrays and are sometimes not called reliably by short-read sequencing pipelines that don't target them explicitly; the exonic c.1236G>A is covered by virtually all clinical and consumer genotyping platforms than the exonic c.1236G>A, which falls in a well-covered coding region and is included on virtually all clinical pharmacogenomic panels and most consumer genotyping chips. For most of the 2010s, c.1236G>A was used as a cheap proxy for HapB3, and the clinical evidence base — including the CPIC guideline and the Alpe-DPD trial — was built on c.1236G>A genotyping as the HapB3 readout. CPIC continues to recommend it as the primary HapB3 tag in the 2018 guideline update.

Because DPYD sits on the minus strand of chromosome 1, the coding-strand G>A change that papers describe corresponds to a C>T change on the GRCh38 plus strand — which is how consumer genotyping chips and whole-genome sequencing files report it. The plus-strand reference allele at chr1:97573863 is C (the wild-type, non-HapB3 allele), and the T is the tag allele associated with HapB3.

The Evidence

rs56038477 has been used as the standard HapB3 tag SNP in the major clinical evidence supporting HapB3 screening:

• Amstutz et al. 2009 Pharmacogenomics⁶⁶ Amstutz et al. 2009 Pharmacogenomics
  The original DPYD haplotype assessment in 111 Swiss cancer patients, identifying HapB3 as one of several haplotypes significantly associated with severe 5-FU toxicity defined HapB3 as a multi-variant haplotype including c.1236G>A and identified it as a toxicity predictor.
• van Kuilenburg et al. 2010 Hum Genet⁷⁷ van Kuilenburg et al. 2010 Hum Genet
  Demonstrated complete co-occurrence of c.1236G>A and c.1129-5923C>G in HapB3 patients and identified the intronic variant as the functional driver via RT-PCR splicing studies confirmed the linkage and isolated the splicing mechanism to the intronic variant.
• Meulendijks et al. 2015 Lancet Oncol⁸⁸ Meulendijks et al. 2015 Lancet Oncol
  Individual patient data meta-analysis of 7,365 cancer patients across eight studies, genotyping HapB3 via c.1236G>A found that HapB3 carriers (c.1236G>A heterozygotes) had a 1.59-fold relative risk of severe fluoropyrimidine toxicity (95% CI 1.29-1.97, p<0.0001), independent of DPYD*2A and c.2846A>T effects.
• The CPIC 2018 guideline update⁹⁹ The CPIC 2018 guideline update
  Amstutz U, Henricks LM, Offer SM, et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) Guideline for Dihydropyrimidine Dehydrogenase Genotype and Fluoropyrimidine Dosing: 2017 Update. Clin Pharmacol Ther 2018 classified HapB3 as a decreased-function allele (activity score 0.5) and recommended 50% dose reduction, using c.1236G>A as the primary HapB3 tag.
• The Alpe-DPD prospective trial by Henricks et al.¹⁰¹⁰ The Alpe-DPD prospective trial by Henricks et al.
  1,103-patient Dutch multicentre trial prospectively genotyping for all four CPIC DPYD variants before fluoropyrimidine chemotherapy and applying 50% dose reductions to carriers implemented HapB3 screening via c.1236G>A genotyping and demonstrated that 50% dose reduction prevents severe toxicity without compromising overall or progression-free survival.
• European implementation cohorts such as PhotoDPYD 2023¹¹¹¹ PhotoDPYD 2023
  Spanish nationwide screen of 8,054 cancer patients using the four-variant CPIC panel including c.1236G>A as HapB3 proxy confirmed HapB3 allele frequencies of 2-4% in Spanish/European populations, consistent with gnomAD v4.

The Linkage Caveat — Why rs75017182 Should Be Genotyped Directly When Possible

Until 2024, the assumption was that c.1236G>A and c.1129-5923C>G were in complete linkage disequilibrium — every HapB3 carrier had both variants, and genotyping one was equivalent to genotyping the other. A 2024 paper by Turner and colleagues¹²¹² Turner and colleagues
Turner AJ, Derezinski AD, Gaedigk A, et al. Updated DPYD HapB3 haplotype structure and implications for pharmacogenomic testing. Clin Transl Sci 2024;17(1):e13699 used large-scale population genomics data to show that the LD is very high but not perfect. A small number of individuals carry c.1236G>A (rs56038477) without the causal c.1129-5923C>G (rs75017182) splice variant — these individuals are functionally normal at the HapB3 locus and would be misclassified as high-risk by c.1236G>A genotyping alone. The reverse also occurs, though more rarely: isolated rs75017182 carriers without the c.1236G>A tag exist and would be missed by tag-only genotyping, putting them at risk of severe toxicity if treated at standard doses.

The clinical implication is that rs75017182 itself should be directly genotyped for HapB3 classification whenever possible, rather than relying on c.1236G>A alone. CPIC has not yet updated the guideline to reflect Turner 2024, and the existing prospective evidence (Alpe-DPD, Meulendijks) was built on c.1236G>A screening — so for now, a c.1236G>A positive result remains clinically actionable as a HapB3 indicator, but the field is moving toward direct splice-variant genotyping as the gold standard. GeneOps stores and interprets both variants separately for this reason: a user with rs56038477 (c.1236G>A) but NOT rs75017182 (the causal splice variant) is classified as functionally normal at the HapB3 locus, while a user with both variants is classified as a true HapB3 carrier requiring dose reduction. A compound action at the supervisor level handles the two-variant HapB3 diagnosis.

Practical Implications

If you carry one or two copies of the c.1236G>A variant (rs56038477), the clinical interpretation depends on whether you also carry the causal c.1129-5923C>G splice variant (rs75017182):

• c.1236G>A positive AND c.1129-5923C>G positive (true HapB3 carrier) — 50% initial fluoropyrimidine dose reduction per CPIC, managed identically to HapB3 heterozygotes via the rs75017182 pathway.
• c.1236G>A positive AND c.1129-5923C>G negative (tag-only carrier, per Turner 2024) — functionally normal DPD activity at the HapB3 locus; standard fluoropyrimidine dosing is appropriate at this locus. Other DPYD variants (*2A, *13, c.2846A>T) must still be ruled out separately.
• c.1236G>A positive with unknown c.1129-5923C>G status — treat as presumptive HapB3 carrier and apply 50% dose reduction. The splice variant carrier rate is very high in c.1236G>A positives (>95% LD in Europeans), and the conservative assumption is safer for a cancer patient. Request c.1129-5923C>G genotyping if your lab supports it.

In practice, most pharmacogenomic panels test c.1236G>A as the HapB3 readout, so most users will arrive at GeneOps with only the tag SNP genotyped. The GeneOps pipeline handles this by applying the conservative CPIC interpretation — c.1236G>A carrier = HapB3 presumed positive = 50% dose reduction recommended — unless the user also has a negative rs75017182 result available, in which case the tag-only status is reported with appropriate caveats.

Interactions

rs56038477 interacts directly with rs75017182, the causal splice variant of the HapB3 haplotype. A user carrying both variants (in cis on the same chromosome, which is the overwhelming majority case in European populations) is a true HapB3 carrier and should be managed as a decreased-function DPD metaboliser: 50% initial fluoropyrimidine dose reduction, therapeutic drug monitoring, cautious escalation to 65-85% of standard by cycle 3 if tolerated. The supervisor-level compound action for HapB3 encodes this combined genotype logic — it fires when both rs75017182 and rs56038477 are present and recommends the CPIC-level HapB3 management.

HapB3 also interacts with the other three CPIC-priority DPYD variants via the activity score system¹³¹³ activity score system
Activity score 2.0 = normal metaboliser; 1.0-1.5 = intermediate metaboliser (heterozygous for a no-function or decreased-function allele); 0-0.5 = poor metaboliser. A compound heterozygote carrying HapB3 alongside a no-function allele (DPYD*2A/rs3918290 or DPYD*13/rs55886062) drops to activity score 0.5 and is typically treated with fluoropyrimidine avoidance rather than dose reduction. A HapB3 + c.2846A>T (rs67376798) compound heterozygote scores 1.0 and requires more aggressive dose reduction than HapB3 alone.

First-degree relatives of c.1236G>A carriers have a ~50% prior probability of carrying the same variant and, because of the high linkage with the causal splice variant, a similar probability of being true HapB3 carriers. Cascade DPYD panel testing is appropriate for first-degree relatives who ever face fluoropyrimidine-based cancer treatment.