GIGYF2 and EIF4G1: disputed genes

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

GIGYF2: Christian Lautier, Stefano Goldwurm, Alexandra Dürr and colleagues (Am J Hum Genet 2008). EIF4G1: Marie-Christine Chartier-Harlin, Matthew Farrer and colleagues (Am J Hum Genet 2011).

What happened

GIGYF2. A region of chromosome 2 (PARK11) had been linked to familial Parkinson's. Lautier sequenced the GIGYF2 gene in 123 Italian and 126 French familial patients. Seven different missense changes were found in 12 unrelated patients (4.8%), and none in controls; three insertions or deletions were found in four more. Three of the changes segregated with disease in four small families, with two unaffected carriers in one. Bras et al. (Hum Mol Genet 2009) sequenced Portuguese and North American samples and found two of the published "mutations" in neurologically normal controls, concluding GIGYF2 mutations are not strongly related to the disease.

EIF4G1. The gene codes for a protein that starts translation (building proteins from RNA). A large multi-case family with a p.Arg1205His change was found, plus four other substitutions in patients with familial parkinsonism or Lewy body disease and not in controls. Replication then went badly:

• Lesage 2012 sequenced EIF4G1 in 251 dominant index cases and 236 controls. One previously reported variant (R1197W) was in a control, and their data did not support pathogenicity for several variants.
• Nuytemans 2013 (exome sequencing, 213 patients and 272 controls) found R1205H in the original family, nonpenetrant in one unaffected 86-year-old, and called it a strong risk factor. There was no overall contribution of EIF4G1 variants in their data.
• Saini 2021 sequenced GIGYF2, EIF4G1 and three other genes in 2,408 patients and 3,444 controls. No gene was associated with Parkinson's, and the authors said their "PARK" designations should be reconsidered.
• Huttenlocher 2015 screened 2,146 European patients plus an Icelandic data set of 93,698 people. R1205H appeared in one sporadic patient; it was also present in 76 older Icelanders, only 5 of whom reported Parkinson's symptoms. Conclusion: not a strong or common risk factor.

Why it moved us toward a cure

It helps the project avoid false trails. Both genes were built into mechanism stories (the GIGYF2 protein binds Grb10, a signalling protein; EIF4G1 controls the start of protein-making). Neither has produced a drug or a validated pathway. The standard now is to see a variant in several unrelated families, with controls screened, before calling a gene causal, as for SNCA (landmark 4). See also the earlier UCHL1 claim (landmark 191).

Outcome

Dead end (the gene claim did not hold up). Both are generally listed as disputed or unconfirmed rather than established Parkinson's genes. Whether the EIF4G1 R1205H founder family is a rare true cause with low penetrance remains open.

Sources

• Lautier et al., Am J Hum Genet 2008;82:822–833, DOI 10.1016/j.ajhg.2008.01.015
• Bras et al., Hum Mol Genet 2009, DOI 10.1093/hmg/ddn340
• Chartier-Harlin et al., Am J Hum Genet 2011;89:398–406, DOI 10.1016/j.ajhg.2011.08.009
• Lesage et al., Neurobiol Aging 2012, DOI 10.1016/j.neurobiolaging.2012.05.006 · Nuytemans et al., Neurology 2013, DOI 10.1212/WNL.0b013e31828727d4 · Huttenlocher et al., J Med Genet 2015, DOI 10.1136/jmedgenet-2014-102570
• Saini et al., Neurobiol Aging 2021, DOI 10.1016/j.neurobiolaging.2020.10.019 (abstract read)
• Not yet verified: the PARK11/PARK18 labels in current databases; whether GIGYF2 was formally retracted (no retraction found, but this was not searched in a retraction database).