HTRA2/Omi (PARK13) did not hold up

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

K. M. Strauss, Hélène Plun-Favreau, Olaf Riess, Thomas Gasser, Zbigniew Wszolek, Jörg Schulz, Rejko Krüger and colleagues (15 authors, Human Molecular Genetics 2005). Later tests by Bogaerts, Van Broeckhoven and colleagues (2008), Krüger and the GEO-PD consortium (2011), and Saini, Rouleau, Fon, Gan-Or and colleagues (2021). No funder is recorded here.

What happened

HTRA2 codes for a protease (a protein-cutting enzyme) that sits in mitochondria, the cell's energy units. Mice lacking it lose neurons and show a parkinsonian picture. Strauss et al. screened German patients and found a rare change, G399S, in four patients and no healthy controls. They also found a common change, A141S, that was linked to the disease. Both reduced the protease's activation, and G399S made cells more prone to stress-induced death. The gene was given the label PARK13 (listed as OMIM 610297 in the LOVD disease index).

What followed:

• 2008: Bogaerts et al. sequenced 266 Belgian patients and 273 controls. They found a new change, R404W, and supported a role for the gene.
• 2011: the GEO-PD consortium (20 sites) typed five common HTRA2 variants in 6,378 patients and 8,880 controls. There was no overall association; odds ratios ranged from 0.98 to 1.08. The authors called it the largest association study of any single Parkinson's gene up to then.
• 2014–15 (online 24 Nov 2014): a six-generation Turkish family carried G399S with essential tremor (a shaking disorder); people with two copies developed parkinsonism in middle age (Unal Gulsuner et al., PNAS). A 2015 PNAS letter (Tzoulis et al., titled "HTRA2 p.G399S in Parkinson disease, essential tremor, and tremulous cervical dystonia") commented on that paper; its findings were not read.
• 2021: Saini et al. fully sequenced HTRA2 and four other genes in 2,408 patients and 3,444 controls. They found no association and said the PARK label should be reconsidered.

Why it moved us toward a cure

It is a lesson in evidence, alongside landmark 192: a candidate-gene hit with strong cell biology is not enough. The mitochondrial story it supported has stronger genetic footing elsewhere. Genes such as PINK1 and parkin are shown in many families.

Outcome

Dead end as a general Parkinson's gene. Whether G399S is a rare cause of tremor or parkinsonism in a few families is not settled.

Sources

• Strauss et al., Hum Mol Genet 2005;14:2099–2111, DOI 10.1093/hmg/ddi215
• Bogaerts et al., Hum Mutat 2008, DOI 10.1002/humu.20713
• Krüger et al., Neurobiol Aging 2011;32:548.e9, DOI 10.1016/j.neurobiolaging.2009.11.021
• Unal Gulsuner et al., PNAS 2014, DOI 10.1073/pnas.1419581111 · Tzoulis et al., PNAS 2015, DOI 10.1073/pnas.1503105112
• Saini et al., Neurobiol Aging 2021;100:119.e7, DOI 10.1016/j.neurobiolaging.2020.10.019
• LOVD: PARK13, OMIM 610297
• No correction, retraction or expression of concern found for these papers (Crossref and Europe PMC comment links checked 20 Oct 2026; the Gulsuner and Tzoulis papers have only linked letters).
• Not yet verified: abstract-level reading only (full texts not read); the Tzoulis letter's findings; the year "PARK13" was assigned; whether a later study re-supported HTRA2.