Reply to: Data from the ROPAD study corroborate an association between ITSN1 loss-of-function variants and Parkinson’s disease

We have read with great interest the “Matters Arising” text of Ganoza and colleagues. After three articles^1–3 demonstrating an association between Loss-of-Function (LoF) ITSN1 variants and Parkinson’s Disease (PD), the authors screened the Rostock International Parkinson’s Disease cohort, including 8660 PD participants. They globally replicate previous findings. ITSN1 LoF carriers represent 1-2/1000 PD individuals suspected to have a genetic cause. The increased risk of developing PD for an ITSN1 LoF variant carrier compared to a non-carrier ranges between 5 and 15^1,2, and the resulting phenotype does not differ significantly from that of classical PD. Of note, regarding the latter, the authors do not provide information about potential atypical neurological symptoms (cerebellar ataxia, cognitive impairment, etc.). Upon the identification of three pedigrees with ITSN1 LoF variants in our cohort from the Paris Brain Institute, among which two had a positive segregation, we raised the hypothesis that ITSN1 could also act as a Mendelian gene³. Not supporting this hypothesis, Ganoza and colleagues found that the frequency of positive family history among ITSN1-PD individuals did not differ significantly from ITSN1 LoF negative individuals (23% vs. 34%, respectively, p = 0.5). However, two comments should be made. First, we noticed that the frequency of positive family history in non-carriers seems higher (34%) compared to what would be the actual frequency for classical PD^4,5, likely because their cohort is biased towards PD individuals suspected of having a genetic etiology of PD. As acknowledged by the authors, the early age at onset of their cohort (55 years) is in favor of a recruitment bias. Second, they did not provide pedigrees (or familial structure) for the 13 individuals carrying LoF ITSN1 variants, therefore we do not know how many relatives were affected in individuals with a positive family history, whether the pedigrees indicated an autosomal dominant or recessive inheritance and whether segregation was performed. A possibility is that several of these LoF occurred de novo, explaining why family history was not always positive. In addition, there is growing evidence of reduced penetrance for ITSN1. Addressing these elements will be essential to conclude about the Mendelian hypothesis and provide accurate genetic counseling. While genetic counseling for ITSN1 appears to be as challenging as for LRRK2 and GBA1, both also associated with reduced penetrance, another important element that is specific to ITSN1 further complicates it. Indeed, ITSN1 LoF variants were not only associated with PD but also to neurodevelopmental disorders (NDD)⁶. While other recently discovered genes such as PSMF1 and EPG5 are similarly responsible for a clinical spectrum featuring infantile/childhood onset disorder including NDD to later onset PD^7–9, the type of variants are usually not the same; truncating variants having more severe phenotypic consequences than missense variants. For ITSN1, only LoF variants were reported in PD. These variants are all predicted to result in haplo-insufficiency. Why could the same mechanism lead to two distinct phenotypes? Currently, two ITSN1 protein isoforms are known¹⁰. Therefore, a hypothesis is that the location of ITSN1 LoF variants, whether affecting one of the two isoforms or both, could explain the variable expressivity. However, variants identified in both diseases seem to affect both isoforms. Therefore, as hypothesized by Ganoza and colleagues, other factors remain to be discovered.

Different approaches may be envisioned to identify these factors. First, even though the various roles and functions of the protein product of ITSN1, Intersectin-1, have already been deeply investigated (nicely summarized in a recent review)¹¹, we do not precisely know why LoF variants are associated with PD yet. Functional studies using biological materials from ITSN1 LoF carriers are underway to explore these aspects (personal communication). The second approach would leverage large PD datasets to compare genetic variants frequency, environmental exposures and lifestyle habits of unaffected elderly controls and PD individuals carrying ITSN1 LoF variants. For instance, variants in the gene coding for the second protein of the intersectin family, ITSN2, should be thoroughly investigated, as this protein is able to compensate some defects induced by decreased ITSN1 expression¹¹. The fact that the average age at onset of ITSN1-PD is consistent across studies should help to determine an appropriate inclusion age for unaffected control individuals. Finally, beyond ITSN1 LoF variants, certain missense variants located in the missense constrained region (amino-acids 1199–1722) were reported to be associated with NDD. To our knowledge, this has not been explored for PD yet.

Finding the cause(s) of variable expressivity and reduced penetrance is not only important to better understand ITSN1 contribution to NDD and PD, but also bears crucial consequences for clinical practice. Indeed, as genetic testing is becoming more accessible and ITSN1 LoF variants are not so rare, ITSN1 LoF variants will, to a certain extent, not so rarely be detected as an incidental finding. While the challenge for genetic counseling after its identification in a neurologically unaffected adult should be like the identification of a pathogenic LRRK2 or severe GBA1 variant, its identification in prenatal settings will raise more ethical issues. A carrier fetus could either face NDD, PD or even no ITSN1-related neurological disorder at all, probably leaving pregnancy interruption decision difficult for would-be parents. Therefore, after this work from Ganoza and colleagues confirming the association of ITSN1 with PD, we believe that there is an urgent need for the field to address such key questions.

Author contributions

G.C. wrote the main manuscript text. C.T., L.W., F.C., E.L., A.L., A.D., F.C.D., B.D., M.P., G.M., S.L., and A.B. reviewed the manuscript.

Data availability

No datasets were generated or analyzed during the current study.

Competing interests

The authors declare no competing interests.