Novel CD46 Variant Links Recurrent aHUS to Aberrant Splicing and Lost Complement Control
A rare intronic change in CD46 proved capable of driving recurrent atypical hemolytic uremic syndrome even though computer-based prediction tools gave it a low pathogenicity score. In a 27-year-old woman, a homozygous c.287-5T>C variant disrupted normal splicing, reduced cell-surface CD46 expression, and produced the kind of unchecked complement activity that can ignite thrombotic microangiopathy at the bedside. The case is a reminder that a noncanonical splice-site variant may look quiet in silico while still being biologically loud.
Her presentation was abrupt and classic for complement-mediated thrombotic microangiopathy. After exposure to someone with coronavirus disease 2019, she developed fever, vomiting, and diarrhea, then arrived with thrombocytopenia, hemolysis, and acute kidney injury. Hemoglobin was 11.5 g/dL, platelets were 12 × 10³/µL, creatinine was 3.48 mg/dL, and lactate dehydrogenase rose to 2,959 U/L. Schistocytes were present, haptoglobin was low, stool and blood cultures were negative, Shiga toxin testing was negative, and ADAMTS13 activity remained normal, making thrombotic thrombocytopenic purpura and typical Shiga toxin-mediated HUS unlikely.
The complement profile did not exclude aHUS. C3 sat within the reference range, C4 was slightly above the upper limit, and CH50 was elevated, a pattern consistent with the broader diagnostic challenge of complement dysregulation: standard serum complement tests often fail to capture the local cell-surface injury that matters most. In this disease, the critical defect is not necessarily depletion in the bloodstream but failure to restrain complement on host endothelium, where C3b and C4b deposition can accelerate microvascular thrombosis and renal injury.
COVID-19 Triggered a Familiar Disease Pattern
The patient had a history of aHUS severe enough to require plasma exchange 8 years earlier, but no family history of the disorder. That clinical history sharpened the suspicion for a recurrent complement-mediated process when SARS-CoV-2 antigen and PCR both returned positive. COVID-19 likely served as the inflammatory trigger, amplifying innate immune activation and complement cascade engagement in a host already vulnerable because of defective membrane cofactor protein activity.
Treatment began quickly with eculizumab 900 mg and a 3-day pulse of methylprednisolone at 1 g/day. She also received meningococcal vaccination and antibiotic prophylaxis, an essential safety measure whenever terminal complement blockade is introduced. Her platelet count climbed steadily, creatinine and LDH fell, and a second 900 mg eculizumab dose was given one week later. By day 15, the platelet count had recovered to 244 × 10³/µL, creatinine had improved to 1.15 mg/dL, and LDH had dropped to 323 U/L. She was discharged and has remained relapse-free for more than 2 years without maintenance eculizumab.
A brisk hematologic response after complement blockade fits the biology of the syndrome. Once the terminal complement pathway is suppressed, ongoing endothelial attack eases, microthrombi stop propagating, and renal perfusion can recover before permanent parenchymal injury takes hold.
A Variant That Slipped Past Computer Prediction
Genetic analysis identified a homozygous CD46 intronic variant, NM_002389.4:c.287-5T>C, in the patient. Her mother carried the same variant in heterozygous form and was asymptomatic. The change had not been cataloged in HGMD, gnomAD, or ClinVar, leaving its significance uncertain at first glance. The usual algorithms did not raise a strong alarm: SpliceAI produced an acceptor-loss score of 0.11, Pangolin gave a splice-loss score of 0.22, and CADD-PHRED was 7.934, all values that would ordinarily argue against a highly pathogenic effect.
Those predictions were misleading. Flow cytometry showed markedly reduced CD46 expression on the patient’s peripheral blood mononuclear cells, nearly reaching the isotype-control level, while the heterozygous mother had reduced but still detectable expression. The functional readout mattered far more than the in silico estimate because CD46 is the complement cofactor that helps factor I inactivate C3b and C4b on the surface of host cells. Without adequate expression, the host membrane loses a key layer of protection and becomes vulnerable to complement-mediated injury.
Exon 3 Skipping Provided the Mechanistic Answer
Transcript analysis closed the loop. RT-PCR demonstrated exon 3 skipping in the patient’s CD46 transcript, creating a premature termination codon. A minigene assay reproduced the same exon-skipping event when the variant-containing sequence was inserted into the pSPL3B system and expressed in HEK293T cells, confirming that the intronic change itself was sufficient to alter splicing.
The molecular consequence is straightforward and clinically important: aberrant splicing truncates the transcript, reduces functional protein expression, and undermines a central complement-regulatory checkpoint. In a homozygous state, the defect became severe enough to produce recurrent disease; in the heterozygous mother, residual expression appears to have been enough to avoid clinical illness. The dosage effect fits the observed biology, but it also exposes a practical limitation of relying on prediction software alone for rare noncanonical splice-site variants.
A normal-appearing intronic base can still erase a protective protein from the cell surface.
Functional Validation Should Follow Suspicious Variants in aHUS
This case adds a concrete example to the expanding catalog of complement-pathway disorders caused by CD46 variants. More than 90 CD46 changes have already been linked to aHUS, but not all intronic variants are easy to interpret. Noncanonical splice-site positions are especially difficult because standard annotation tools may miss subtle effects on exon recognition, branch-point usage, or splice acceptor strength. When the clinical phenotype is strong and the genotype is rare, expression studies, RT-PCR, and minigene assays can provide the evidence needed to classify pathogenicity.
For clinicians, the bedside lesson is immediate: recurrent thrombotic microangiopathy after an infectious trigger should keep complement-mediated disease high on the differential even when routine complement levels are not dramatically abnormal. For geneticists, this case shows why a low CADD score or a modest SpliceAI result should not end the investigation when the phenotype fits and the variant segregates with reduced protein expression. For patients, the payoff is equally concrete, because correct molecular diagnosis can support targeted complement blockade instead of repeated empiric rescue therapy.
This report remains limited to a single patient, and the same approach will not automatically generalize to every intronic variant in complement genes. Still, the combination of homozygosity, transcript disruption, loss of surface expression, and functional rescue by complement inhibition makes the pathogenic mechanism unusually coherent in this case. The remaining challenge is not whether CD46 matters in aHUS, but how many apparently quiet intronic changes hide similar splice-disrupting effects until they are tested directly.