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

We read with interest the recent studies by Skuladottir and colleagues and Cogan and colleagues, which report a novel association between monoallelic loss-of-function (LoF) variants in the ITSN1 gene and risk for Parkinson’s disease (PD)^1,2. The first study observed that heterozygous LoF variants in ITSN1 confer a large risk for developing PD in analyses combining the deCODE, UK Biobank, and Accelerating Medicines Partnership Parkinson’s Disease (AMP-PD) datasets¹. The second study identified five PD patients from three families who harbored distinct heterozygous frameshift ITSN1 variants and reported their clinical phenotype². Yet another recent study reported similar results in an analysis of the UK Biobank, the AMP-PD, and the All of US cohorts, where they observed an increased risk of PD in individuals with LoF variants in ITSN1 and provided functional evidence in a Drosophila model, showing an interaction between ITSN1– encoded Intersectin-1 and α-synuclein³. Still, the clinical manifestation of PD in patients with LoF variants in ITSN1 remains largely unexplored. We aimed to investigate the reported association of ITSN1 LoF variants with PD in a large independent patient group and to determine whether patients with ITSN1 LoF variants present with a clinically distinct PD form.

We investigated our data from the Rostock International Parkinson’s Disease (ROPAD) study (NCT03866603), an observational clinical study that enrolled and genetically characterized PD patients from Europe, the Americas, and Israel^4,5. Our analysis included data from 8660 reportedly unrelated PD patients recruited in Europe (68%), North America (16%), Israel (10%), Turkey (4%), and South America (2%) between April 2019 and August 2024 who had exome sequencing (ES) or genome sequencing (GS) data available and no relevant variants in known PD genes. Specifically, the patients analyzed in the present study did not have (i) a pathogenic (P) or likely pathogenic variant (LP) in an autosomal dominant PD-related gene (LRRK2, SNCA, VPS35, RAB32, and CHCHD2), (ii) two heterozygous or one homozygous P/LP variant in an autosomal recessive PD-related gene (PARK7, PINK1, and PRKN), or (iii) a P, LP, or risk factor variant in GBA1. In addition, we examined the presence of ITSN1 LoF variants in 210,657 ES- or GS-tested individuals of diverse ancestries from the CENTOGENE Biodatabank who were referred in various types of settings (routine exome/genome-based diagnostic genetic testing, healthy parents from trio analyses, or participants of other observational clinical studies) and who were not referred for or diagnosed with PD. ITSN1 LoF variants were defined as start-loss, stop-gain, stop-loss, splice-donor, splice-acceptor, or frameshift variants, predicted by CENTOGENE’s in-house bioinformatic pipeline^6,7. We extracted ROPAD study electronic case report form (eCRF) information from the Centogene Biodatabank, including year of birth, sex, age at disease onset (AAO), age at PD diagnosis (AAD), age at enrollment (AAE), family history of PD, and motor clinical evaluation (partial UPDRS part III), and compared all available parameters between ITSN1 LoF variant-positive and ITSN1 LoF variant-negative study participants^4,8,9. The disease duration (DD) was calculated as the difference between the AAO and the AAE. Statistical analyses (linear regression, Mann–Whitney, Fisher’s exact, and Pearson chi-square tests) were performed with GraphPad Software (version 11).

We identified 13 PD patients with heterozygous ITSN1 LoF variants. This corresponds to 0.15% of the ROPAD patient group and represents a significant enrichment over the 0.025% in a group of 210,657 ES- or GS-tested individuals from the CENTOGENE Biodatabank (odds ratio (OR) = 6.09, 95% confidence interval (CI) = 3.31–11.19, p = 1.34 × 10⁻⁶). Additionally, the proportion of ITSN1 LoF alleles in ROPAD PD patients (0.075%) is approximately six times higher than the proportion of ITSN1 LoF alleles found in non-Finnish European individuals in the Genome Aggregation Database (0.012%; gnomAD v4.1.1) and is also enriched relative to each other gnomAD population (Supplementary Table 1)¹⁰. The overrepresentation of ITSN1 LoF variants in ROPAD largely overlaps with the CIs of the ORs observed in both prior association studies (OR = 7.3, 95% CI = 3.5–15.2 and OR = 10.5, 95% CI = 5.2–21.3)^1,2. Furthermore, the frequency of ITSN1 LoF variant carriers in our study corresponds to those reported by Skuladottir and colleagues (0.15% in meta-analysis)¹ and Cogan et al. (0.14%)². Among ROPAD participants, ITSN1 LoF variants are the sixth most common disease-relevant positive genetic finding following variants in the GBA1, LRRK2, PRKN, SNCA, and GCH1 genes⁵.

We identified 12 distinct ITSN1 LoF variants among the aforementioned 13 PD patients. The spectrum of variants included nonsense (n = 4), frameshift (n = 6), and splice site changes (n = 2) (Table 1). One variant was observed in two patients, and the remaining 11 were observed once each. Our analyses of potential relatedness between the two German patients harboring the same ITSN1 variant (c.2842_2843del, Table 1) did not reveal evidence of close relatedness or a shared extended founder haplotype (Supplementary Material), indicating that in these two instances the variant likely arose independently.

Table 1.

ITSN1 variants detected in the present study

Variant				Patients
cDNA level (NM_003024.3)	Protein level (NP_003015.2)	Predicted effect	Previously reported	Number (sex, country)	Age at onset (years)
c.328del	p.(Ser110Leufs*37)	frameshift	No	1 (female, Germany)	53
c.765dup	p.(Gln256Thrfs*12)	frameshift	No	1 (male, Spain)	22
c.934C>T	p.(Arg312*)	nonsense	Yes3,13	1 (female, Germany)	57a
c.1504C>T	p.(Arg502*)	nonsense	Yes13	1 (male, United Kingdom)	72
c.1825-2A>G	p.(?)	splice site	No	1 (male, Germany)	47a
c.2842_2843del	p.(Met948Valfs*41)	frameshift	Yes1–3,13	2 (male, Germany, female, Germany)	42
					56
c.3038dup	p.(Tyr1013*)	nonsense	No	1 (male, Germany)	52b
c.3158_3161del	p.(Asn1053Metfs*2)	frameshift	No	1 (female, Germany)	57
c.3310C>T	p.(Arg1104*)	nonsense	Yes1,3,13	1 (male, Belgium)	54a
c.3469+1G>A	p.(?)	splice site	No	1 (male, Israel)	45
c.4180_4192del	p.(Glu1394Lysfs*9)	frameshift	No	1 (female, Italy)	47
c.4584del	p.(Lys1528Asnfs*29)	frameshift	No	1 (male, Israel)	60

^aPatient with a positive family history,

^bPatient with an unknown family history. In all other patients, family history was negative. Individuals reported in refs. ^1–3 are either PD patients or controls harboring the respective heterozygous ITSN1 variants, while individuals reported in refs. ¹³ are patients with autism harboring the respective heterozygous ITSN1 variants.

As in both previous studies, our data show that the spectrum of disease-associated ITSN1 variants is largely private, which contrasts with the observed spectrum of variants in other dominant PD genes, where few recurrent variants predominate in PD patients^11,12. Interestingly, over 40 distinct ITSN1 LoF variants have been reported in individuals with autism and neurodevelopmental disorders^13,14, including four variants identified in our patient group (Table 1), potentially representing a remarkable example of pleiotropy. Given that Intersectin-1 is involved in clathrin-mediated endocytosis, synaptic vesicle recycling, and actin/CDC42-related signaling, its haploinsufficiency may disturb either neurodevelopmental circuitry or later-life neuronal maintenance. The ultimate outcome may depend on genetic modifiers, expression dosage (in case of splice‑altering variants), environmental stressors, and other factors.

Our analyses showed that, when comparing ITSN1-LoF-positive and ITSN1-LoF-negative study participants, the male-to-female ratio (1.6 vs. 1.7, respectively, p = 1.0), positive family history of PD (23% vs. 34%, p = 0.5), median AAO (53 vs. 55 years, p = 0.3, Fig. 1), median AAD (54 vs. 56 years, p = 0.2), median AAE (61 vs. 63 years, p = 0.1), and median DD (5 vs. 5 years, p = 0.7) did not differ significantly. Our multiple linear regression analysis of partial UPDRS III scores, adjusted for DD, showed that, on average, ITSN1-LoF-positive patients did not differ significantly from ITSN1-LoF-negative individuals (p = 0.075). Of note, the previous two studies of large cohorts of PD patients observed significantly younger AAOs among ITSN1-LoF-positive than among ITSN1-LoF-negative individuals (mean AAOs reported in ref. ¹: 62 vs 70 years, respectively; median AAOs reported in ref. ³: 64 vs 70 years, respectively). The younger AAO in the analyzed subset of ROPAD participants (mean and median: 55 years) likely accounts for the absence of a significant AAO difference between ITSN1-LoF-positive and ITSN1-LoF-negative patients in our study. Additionally, the low median AAO in our ITSN1-LoF-positive participants is comparable to that of the five PD patients with heterozygous LoF ITSN1 variants reported by Cogan et al. (47 years)². Interestingly, two of the latter patients also had a GBA1 variant classified as mild². Similarly, we identified a female patient (AAO: 70 years, no family history of PD) with the c.2842_2843del ITSN1 variant and another mild GBA1 variant (c.1226A>G; p.(Asn409Ser)). We, however, excluded this woman from our ITSN1-related analyses, as we suspected that the GBA1 variant she harbors in addition to the ITSN1 change may have modified her disease risk, onset, or severity as documented in carriers of dual LRRK2 and GBA1 variants^5,15.

Fig. 1. Age at onset in ITSN1 loss-of-function (LoF) variant-positive and variant-negative ROPAD participants.

Comparison of age at onset between the 13 participants carrying an ITSN1 LoF variant and the 8647 variant-negative participants. Violin plots show the distribution of age at onset; individual dots represent participants, and horizontal lines indicate the median (bold) and interquartile range (dashed).

In conclusion, our findings corroborate the association between LoF variants in ITSN1 and an increased risk of PD. They also show that PD patients with heterozygous ITSN1 LoF variants are indistinguishable from idiopathic PD patients, indicating that ITSN1 is a novel gene related to classical PD. Thus, further research is needed to understand the natural history of ITSN1-related PD (PARK-ITSN1) and treatment responses in patients with ITSN1-LoF-associated PD. Finally, our study illustrates the utility of well-curated proprietary clinical and genetic data in providing valuable insights for diagnostic labs, research institutions, and pharmaceutical companies.

Author contributions

C.A.G., P.B. and C.B. conceptualized the research project. C.A.G., J.R., P.B. and C.B. organized the research project. C.A.G., A.W., E.N.S. and C.B. executed the research project. J.J.P., F.C. and P.B. designed the observational study. J.J.P., F.C. and J.R. executed the observational study. C.G., A.W., E.N.S. and C.B. designed the data analysis. C.G., A.W., E.N.S., J.R. and C.B. executed the data analysis. C.G., A.W. and C.B. wrote the first draft of the manuscript. All authors reviewed the manuscript.

Data availability

Due to privacy considerations and General Data Protection Regulation (GDPR) compliance, the data from ROPAD study participants are not publicly available. However, the data that support this study's findings are available from the corresponding author upon reasonable request.

Code availability

The code used for the haplotype relatedness analysis is publicly available in the project’s GitHub repository at: https://github.com/AnaWestenberger/ITSN1project/blob/49fb76f14ef9e1cd01b932a4fbf8835da09f8d0c/ITSN1_relatedness_haplotypecheck.sh.

Declarations

Competing interests

C.A.G., J.J.P., F.C., E.N.S., J.R., P.B., and C.B. are current or former employees of CENTOGENE GmbH. A.W. is an advisor for medical writing and data analysis to CENTOGENE GmbH. The authors declare no competing financial or non-financial interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41531-026-01562-x.