One faulty copy of PRKN is probably not a risk factor

The research record

This is the project's file for this landmark, as written and checked. Sources are listed at the end; anything not yet confirmed is marked "Not yet verified".

Who

Yu and Rudakou, with Ziv Gan-Or (McGill) as senior author (Movement Disorders 36:178-187, online 24 September 2020; the Europe PMC record lists NIH funding). Lubbe, Huw Morris (senior author) and the International Parkinson's Disease Genomics Consortium (Human Molecular Genetics 30:78-86, 2021). Zhu, Narendra (senior author) and colleagues at the US National Institutes of Health (Brain 145:2077-2091, 2022).

What happened

Two broken copies of PRKN cause young-onset parkinsonism (landmark 12), and the PINK1 gene works in the same pathway (landmark 15). Many people carry one broken copy of PRKN and are well. Earlier papers had suggested that one copy raises risk (for PINK1, for example, a 2006 paper in Movement Disorders by Djarmati and colleagues was titled "Heterozygous PINK1 mutations: a susceptibility factor for Parkinson disease?", with a question mark). If true, it would matter for the mitophagy ("clearing damaged mitochondria") theory and for who joins trials.

• Yu 2021. PRKN fully sequenced in 2,809 people with Parkinson's and 3,629 controls, including copy-number changes (missing or extra stretches of DNA). Pathogenic or likely pathogenic single copies were not more common in patients: 1.52% versus 1.8% in controls (false-discovery-rate-corrected P = 0.55). No link to age at onset. The authors suggest the sequencing method as a cheap screen for people with two broken copies before trials.
• Lubbe 2021. Three case-control groups (10,858 cases, 8,328 controls) showed a more than 1.5-fold higher risk for single mutations (P = 0.035) after removing people known to have two. A pooled analysis (19,574 cases, 468,488 controls) gave an odds ratio of 1.65. But when studies that did not check for a second mutation were removed, the link vanished (odds ratio 1.23, P = 0.614, four studies). Among the 30% of apparent single carriers that were rechecked for hidden copy-number changes, 44% had a second mutation. Single mutations were also linked to younger onset. The authors still describe the study as supporting an association, and urged caution about the estimates.
• Zhu 2022. In the UK Biobank (200,606 people with exomes and array data) and 4,094 people in NIH and AMP PD cohorts, one pathogenic PRKN variant was found in 1.8% of UK Biobank participants. For Parkinson's the odds ratio was 0.91 (95% CI 0.58–1.38, P = 0.76). In the NIH and AMP PD group it was 1.29 (0.74–2.38, P = 0.43). Neither was significant. Tests on patient skin cells suggested that hidden loss-of-function variants in single carriers are not a big confounder, which speaks to Lubbe's concern.

Why it moved us toward a cure

It corrects a common assumption: being a carrier is not a clear sign of risk. In my reading, a trial of a mitophagy drug (landmark 15) should focus on people with two broken copies, and Yu 2021 offers a cheap way to screen for them. It also shows how hidden second hits can fake a risk signal. PINK1 single copies were not covered by these three papers.

Outcome

Open, leaning to no risk. Two studies are null; the third reports a raised risk but shows it fades in the better-checked studies, and the authors warn that missed second mutations may explain it. Whether very small effects exist is not excluded.

Sources

• Yu et al., Mov Disord 2021;36(1):178–187, DOI 10.1002/mds.28299 (abstract via Europe PMC, PMID 32970363)
• Lubbe et al., Hum Mol Genet 2021;30(1):78-86, DOI 10.1093/hmg/ddaa273, PMID 33448283
• Zhu et al., Brain 2022;145(6):2077–2091, DOI 10.1093/brain/awab456, PMID 35640906 (PMC9423714)
• Djarmati et al., Mov Disord 2006, PMID 16755580 (title confirmed via Europe PMC)
• No corrections or retractions are recorded for the three papers (Crossref, Europe PMC).
• Not yet verified: the PINK1 evidence beyond the Djarmati title; the earlier papers that claimed raised risk for PRKN; the full funding statements. The trial-design sentence is my inference.