Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 https://doi.org/10.1186/s13024-021-00455-2 RESEARCH ARTICLE Open Access Identification of sixteen novel candidate genes for late onset Parkinson’s disease Alessandro Gialluisi1†, Mafalda Giovanna Reccia1†, Nicola Modugno1, Teresa Nutile2, Alessia Lombardi1, Luca Giovanni Di Giovannantonio2, Sara Pietracupa1, Daniela Ruggiero1,2, Simona Scala1, Stefano Gambardella1,3, International Parkinson’s Disease Genomics Consortium (IPDGC), Licia Iacoviello1,4, Fernando Gianfrancesco2, Dario Acampora2, Maurizio D’Esposito2, Antonio Simeone2, Marina Ciullo1,2 and Teresa Esposito1,2* Abstract Background: Parkinson’s disease (PD) is a neurodegenerative movement disorder affecting 1–5% of the general population for which neither effective cure nor early diagnostic tools are available that could tackle the pathology in the early phase. Here we report a multi-stage procedure to identify candidate genes likely involved in the etiopathogenesis of PD. Methods: The study includes a discovery stage based on the analysis of whole exome data from 26 dominant late onset PD families, a validation analysis performed on 1542 independent PD patients and 706 controls from different cohorts and the assessment of polygenic variants load in the Italian cohort (394 unrelated patients and 203 controls). Results: Family-based approach identified 28 disrupting variants in 26 candidate genes for PD including PARK2, PINK1, DJ-1(PARK7), LRRK2, HTRA2, FBXO7, EIF4G1, DNAJC6, DNAJC13, SNCAIP, AIMP2, CHMP1A, GIPC1, HMOX2, HSPA8, IMMT, KIF21B, KIF24, MAN2C1, RHOT2, SLC25A39, SPTBN1, TMEM175, TOMM22, TVP23A and ZSCAN21. Sixteen of them have not been associated to PD before, were expressed in mesencephalon and were involved in pathways potentially deregulated in PD. Mutation analysis in independent cohorts disclosed a significant excess of highly deleterious variants in cases (p = 0.0001), supporting their role in PD. Moreover, we demonstrated that the co-inheritance of multiple rare variants (≥ 2) in the 26 genes may predict PD occurrence in about 20% of patients, both familial and sporadic cases, with high specificity (> 93%; p = 4.4 × 10− 5). Moreover, our data highlight the fact that the genetic landmarks of late onset PD does not systematically differ between sporadic and familial forms, especially in the case of small nuclear families and underline the importance of rare variants in the genetics of sporadic PD. Furthermore, patients carrying multiple rare variants showed higher risk of manifesting dyskinesia induced by levodopa treatment. * Correspondence: teresa.esposito@igb.cnr.it † Alessandro Gialluisi and Mafalda Giovanna Reccia contributed equally to this work. 1 IRCCS Istituto Neurologico Mediterraneo Neuromed, Pozzilli, Isernia, Italy 2 National Research Council, Institute of Genetics and Biophysics “Adriano Buzzati-Traverso”, Naples, Italy Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 2 of 18 Conclusions: Besides confirming the extreme genetic heterogeneity of PD, these data provide novel insights into the genetic of the disease and may be relevant for its prediction, diagnosis and treatment. Keywords: Late onset Parkinson’s disease, Whole exome sequencing, Novel candidate genes for Parkinson’s disease, Rare variant burden analysis Background were recruited at the “Parkinson Institute Biobank,” Parkinson’s disease (PD) is a neurodegenerative move- member of the Telethon Network of Genetic Biobank ment disorder characterized by the loss of mesodience- (biobanknetwork.telethon.it/) (hereafter-called PIB co- phalic dopaminergic (mdDA) neurons of the substantia hort). Forty-seven PD patients (siblings or parent-offspring nigra pars compacta (SNpc), in association with the pairs) from these families underwent whole-exome sequen- presence of Lewy bodies in some surviving neurons [1, cing (WES) analysis to identify novel PD candidate genes. 2]. mdDA neurons play a crucial role in the control of Forty PD patients (of which only a small amount of DNA motor, sensory-motor and motivated behavior [3]. Their was available), from the same 23 families, were used for degeneration causes the characteristic symptoms of PD, segregation analysis to confirm the variants identified which include resting tremor, bradykinesia, rigidity, through WES approach. All PD subjects (age at onset 61.7; postural instability and a variety of other motor and Standard Deviation (SD) ± 8.57) were evaluated by neurolo- non–motor symptoms [2]. It has been demonstrated that gists of the Parkinson Study Group, according to published disease’s symptoms are evident when patients have diagnostic criteria (Table S1) [7]. Written informed consent already lost 50–60% of their DA neurons, suggesting the was obtained from all participants. need for early diagnosis of the disease [2]. Most PD cases are sporadic, with unknown etiology. Approxi- MNI cohort mately 5–10% of PD cases appear to have monogenic Three PD families with supposedly dominant transmis- forms of inheritance including autosomal dominant sion were recruited at the IRCCS Mediterranean Neuro- (SNCA, LRRK2, and VPS35) and recessive (PARK2, logical Institute (MNI) in Pozzilli (hereafter called MNI PINK1, DJ1) forms [4–6]. However, many additional loci cohort) (Fig. 1; families numbered 24 through 26). Two associated to PD have recently been identified, even PD cases for each family underwent WES analysis to though for most of them the molecular function is still identify novel PD candidate genes. Segregation analysis unknown [5]. Moreover, there is a high genetic hetero- to confirm WES results was performed on PD cases geneity at the basis of PD, with many different rare mu- from the same families (Fig. 1). tations usually detected only in a single family or in Three hundred ninety-four independent and unrelated small populations [5]. Despite extensive ongoing studies, PD patients (243 males; 172 familiar and 222 sporadic the pathogenic mechanisms underpinning PD remain cases), were used as validation cohort, as well as to largely elusive [6]. Moreover, although the Next Gener- assess the polygenic inheritance of rare variants as risk ation Sequencing (NGS) methods are widespread also in factor for PD occurrence and for endophenotypes mani- the diagnostic field, their application to PD is still lim- festation (Table S1) [8]. The project was approved by the ited to a small number of patients with a clear family ethical committees of IRCCS Neuromed and written in- history and to a small number of causative mutations, formed consent was signed by all the participants. All due to the high genetic heterogeneity associated with the subjects were of European ancestry and were evaluated disease and to the difficulty in interpreting test results. by qualified neurologist of the Parkinson Centre from In this study, we performed a multi-stage procedure June 2015 to December 2017, with a thorough protocol aimed at the identification of potential PD-causative var- comprising neurological examination and evaluation of iants and then at demonstrating that the co-inheritance non-motor domains. Information about family history, of multiple rare variants in Mendelian genes may in- demographic characteristics, anamnesis, and pharmaco- crease the risk of PD in a non-Mendelian fashion, and logical therapy was also collected [8, 9]. The mean age at that this may influence the occurrence of the disease diagnosis was 58.26 years (SD 9.74) [8, 9]. and the manifestation of clinical signs. The Movement Disorder Society revised version of the Unified Parkinson’s Disease Rating Scale Part III (18 Patients and methods items, maximum score 72; hereafter called UPDRS) [10] Study participants was used to assess clinical motor symptoms. These in- PIB cohort cluded language, facial expressions, tremor, rigidity, agil- Twenty-three PD families with supposedly dominant ity in movements, stability, gait and bradykinesia. transmission (Fig. 1, families numbered 1 through 23) Cognitive abilities were tested through an Italian Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 3 of 18 Fig. 1 Graphical representation of the 26 families used as discovery cohort. Pedigrees of the 26 Italian families with supposedly dominant PD forms. Affected individuals are indicated with dark symbols. Patients who underwent WES analysis are shown with dark arrows. Additional family members for which DNA was available are indicated with an asterisk. Only affected subjects were used for segregation analysis. For each family, the rare deleterious variants reported in Table 1 were carried by the affected family members validated version of the Montreal Cognitive Assessment (MoCA) [11]. Cognitive domains assessed include short- term memory (5 points); visuospatial abilities via clock drawing (3 points), and a cube copy task (1 point); ex- ecutive functioning via an adaptation of Trail Making Test Part B (1 point), phonemic fluency (1 point), and verbal abstraction (2 points); attention, concentration, and working memory via target detection (1 point), serial subtraction (2 points), digits forward and back- ward (1 point each); language via confrontation naming with low-familiarity animals (3 points), and repetition of complex sentences (2 points); and orien- tation to time and place (6 points). The total score was given by the sum of these domains, and then di- vided by the maximum score that could be obtained (30 points). Where one or more domains could not be tested (e.g. visuospatial tasks, due to unavailability of optical devices), the corresponding score was sub- tracted from the total score obtainable. Non-motor symptoms were assessed through an Italian validated version of Non Motor Symptoms Scale (NMMS) for Parkinson Disease [12]. This scale tests 9 items, including cardiovascular domain, sleep/ fatigue, mood/cognition, perceptual problems/halluci- nations, attention/memory, gastrointestinal, urinary, sexual function, and ability to taste or smell. For each item, both severity and frequency of symptoms is measured, so that the scale accounts for both as- pects. Here, the sleep domain was slightly modified by adding a further question on the occurrence of vivid dreams. This question was treated as all the others, i.e. the severity of impairment was scored from 0 (no symptoms) to 3 (severe impairment), and the frequency of impairment was scored from 0 (less than once a week) to 4 (daily impairment), then the total score of the sub-item was computed as the product of severity by frequency, and added to the scores of the other sub-items. In the endophenotypes analysis we used UPDRS score < 30 (corresponding to a mild phenotype: 1–2 of Hoehn and Yahr (HY) scale) and ≥ 30 (corresponding to a severe phenotype: 3–5 HY scale) [13], while in non-motor symptoms we used as cutoff ≤54 and > 54 (9 items × 2 (mild im- pairment) × 3 (weekly impairment)). Statistical Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 4 of 18 significance was set to α = 0.01, correcting for five Quality control and variant annotation phenotypes tested. The alignments of the 100-bp paired-end reads to the human reference genome was performed by using the Italian control cohort Burrows Wheeler Aligner (BWA) MEM v0.7.5 [17]. As general Italian population, we downloaded whole After removal of duplicate reads through Picard Mark- genome data of 107 samples of the Tuscan Italians (TSI) Duplicates command (with standard options), we called population of the 1000 Genome Project (phase 3 release) the single nucleotide variants (SNVs) and insertions/de- [14] and extracted the exome regions covered by the letions (indels) for all samples using HaplotypeCaller same capture kit used in our sequencing experiments (BP_RESOULTION option) and GenotypeGVCFs in (SureSelect All Exome kit v6; Agilent® Technologies, Genome Analysis Toolkit (GATK) v3.5–0-g36282e4, Santa Clara, CA, USA). following the manufacturer best practice guidelines Moreover, 96 neurological controls, including 38 from (available at https://software.broadinstitute.org/gatk/ MNI and 58 from the Moli-sani genetic biobank (mean best-practices/) [18]. Variants with Minor Allele Count age 77 years; SD 5.4; 45 women) [15], underwent tar- (MAC) = 0, number of alternative alleles ≠ 2 and call geted resequencing (NGS-TR) of the panel of 26 PD rate < 95% were also filtered out, as well as samples with genes identified in this study. identical-by-descent (IBD) sharing and sex mismatches, and samples with call rate < 90%. Similarly, samples were International cohort checked for absence of outliers in terms of genome-wide WES data of 1148 young-onset unrelated PD cases (aver- homozygosity, of number of singleton variants and of age age at onset 40.6 years; range 35–56 years) and 503 genetic ancestry (through Multidimensional Scaling control participants of European ancestry from the Inter- Analysis in PLINK) [19]. Variants passing quality control national Parkinson’s Disease Genomics Consortium were annotated to genes (within 10 kb from transcrip- (IPDGC) were used to replicate the observations made in tion start/stop site) through ANNOVAR [20]. Variant the Italian cohort. Details on this dataset and on the QC annotation contained information concerning variant carried out on these samples are reported elsewhere [16]. type, Minor Allele Frequency (MAF) in the general population, and predictions of the variant’s effect on Next generation sequencing (NGS) gene function. MAF was annotated in NHLBI GO Genomic DNA was isolated from peripheral blood lym- Exome Sequencing Project ESP6500si-v2 (European phocytes by PAX gene Blood DNA Midi Kit (QIAGEN, American and African American population), 1000 Hilden, Germany). Genomes Project (AFR [African], AMR [Admixed American], EAS [East Asian], EUR [European], SAS Whole-exome sequencing (WES) [South Asian], Exome Aggregation Consortium (ExAC) WES was performed on 47 affected individuals (23 fam- (EUR, non-Finish European population [NFE], AFR, ilies) from PIB cohort, 6 affected individuals (3 families) SAS, EAS and AMR). SIFT, PolyPhen2 and Combined and 106 independent and unrelated PD cases from MNI Annotation Dependent Depletion (CADD) were used to cohort. Exonic regions were enriched using the SureSe- assess the deleterious effects of the identified variants. lect All Exome kit v6 (Agilent® Technologies, Santa To discover novel PD causative variants we analyzed Clara, CA, USA) based on DNA fragmentation and cap- WES data from 53 PD patients (mean age at onset ture. Exomes were barcoded and sequenced at Helm- 61.71, SD 8.57) belonging to 26 late onset PD families holtz Zentrum, München, Germany, using the Illumina® with supposedly dominant transmission (Fig. 1), by using HiSeq2000 platform. a multi-step filtering approach as reported in our previ- ous publications [21, 22]. Prioritization of candidate Targeted resequencing (NGS-TR) genes was made through segregation analysis. STRING NGS-TR was performed on 288 independent and unre- database analysis was used to identify network of PD- lated PD cases from MNI cohort and on 96 Italian interacting genes products [23]. healthy individuals. Probes specific for 100 genes includ- ing the 26-targeted genes, identified in this study, and GBA were designed with Nimble Design software (Roche Multiple ligation dependent probe amplification (MLPA) Diagnostics, Mannheim Germany). Targeting regions The commercially available kit P051-P052 (MRC-Hol- were enriched using the SeqCap kit based on DNA land, Amsterdam, Netherlands) was used for the multi- fragmentation (Kapa Hyper plus kit) and capture (Roche plex dosage of exons for the following genes: DJ1 (4 Diagnostics, Mannheim Germany). Targeted regions probes in P051), SNCA (5 probes in P051, 1 probe in were barcoded and sequenced on MiSeq platform (Illu- P052), PARKIN (12 probes in P051, 12 in P052), LRRK2 mina, San Diego, CA, US). (8 probes in P052), PINK1 (8 probes in P051). The Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 5 of 18 MLPA was performed on DNA from the 26 families expression of the glyceraldehydes-3-phosphate de- who underwent WES analyses. hydrogenase (Gapdh) gene. Pitx3 was used as an ex- ample of mdDA neurons restricted gene expression Polygenic variant load in unrelated Italian individuals marker. Standard deviation was calculated by using We identified 26 PD candidate genes in this study data of three different experiments. (AIMP2, CHMP1A, GIPC1, HMOX2, HSPA8, IMMT, KIF21B, KIF24, MAN2C1, RHOT2, SLC25A39, SPTBN1, Immunohistochemistry assays TMEM175, TOMM22, TVP23A, ZSCAN21, PARK2, Immunohistochemistry was performed on PINK1, DJ-1, LRRK2, HTRA2, FBXO7, EIF4G1, DNAJC6, paraformaldehyde-fixed, wax-included brains as de- DNAJC13, SNCAIP). To assess the co-existence of scribed previously [25]. Mouse and rat adult brains were multiple rare variants in these 26 genes we counted the obtained from healthy animals sacrificed in accordance number of the rare exonic variants in each individual. with the recommendations of the European Commis- The protocol used a MAF ≤ 0.001 which was chosen to sion. All the procedures related to animal treatments avoid the loss of heterozygous pathogenic mutations in were approved by Ethic-Scientific Committee for Animal recessive genes such as PARK2, PINK1 and DJ1. The Experiments and Italian Health Ministry. Formalin-fixed pipeline included: and paraffin-embedded (FFPE) Substantia Nigra tissue sections from human adult normal brain were purchased – Annotation and selection of all the exonic variants; at BioChain (Newark, CA, USA). Rabbit antibodies were – Exclusion of synonymous variants; directed against SLC25A39 (Abcam, Cambridge, UK, – Exclusion of variants with alternative allele Ab-105,683; dilution 1:80) and Tyrosine Hydroxylase frequency > 0.001; (TH) (Chemicon International, Temecula, CA, USA, – Count of the variants detected per subject in the set ab152; dilution 1:600). Mouse antibodies were directed of 26 candidate genes. against GIPC1 (Santa Cruz Biotechnology, Dallas, TX, – Statistical analysis USA, Sc-271,822; dilution 1:80), TOMM22 (Abcam, Cambridge, UK, Ab-57,523; dilution 1:80), ZSCAN21 The co-inheritance of rare variants (MAF ≤ 0.001) was (Novus Biologicals, Centennial, CO, USA, NBP2–45443; investigated by counting the number of variants laying dilution 1:80), HSPA8 (Santa Cruz Biotechnology, Dallas, in the 26 genes, for each sample analyzed, both cases TX, USA, Sc-59,570; 1:100), TH (Chemicon Inter- and controls. Fisher’s Exact tests were performed to test national, Temecula, CA, USA, MAB318; dilution 1:500). differences in the number of cases and controls carrying 0, 1 or ≥ 2 variants. Odds Ratios (ORs) and their corre- URLs sponding 95% Confidence Intervals (95% CIs) were also STRING (https://string-db.org/); 1000 Genomes Project estimated in R [24]. Bonferroni correction for two vari- phase 3 release (ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/ ant load contrasts was applied, resulting in a corrected release/20130502/); Human Gene Mutation Database α = 0.025. (HGMD, http://www.hgmd.cf.ac.uk/ac/index.php; Jun- Then an analysis of Receiver Operating Characteristic 2017 version); Parkinson Disease Mutation Database (ROC) curve was performed to establish the accuracy of (PDmutDB, https://www.molgen.vib-ua.be/PDMutDB/; the test. Jun-2017 version); NHLBI GO Exome Sequencing Pro- ject ESP6500si-v2 (http://evs.gs.washington.edu/EVS/); Expression studies 1000 Genomes Project (http://www.1000genomes.org/); The expression profile of the 16 novel PD candidate genes Exome Aggregation Consortium (ExAC) (http://exac. AIMP2, CHMP1A, GIPC1, HMOX2, HSPA8, IMMT, broadinstitute.org/); SIFT (http://sift.bii.a-star.edu.sg/ KIF21B, KIF24, MAN2C1, RHOT2, SLC25A39, SPTBN1, www/SIFT_seq_submit2.html/); PolyPhen2 (http:// TMEM175, TOMM22, TVP23A, ZSCAN21 was evaluated genetics.bwh.harvard.edu/pph2/); Combined Annotation on total RNAs from the mesencephalon of adult mice Dependent Depletion (CADD) (cadd.gs.washington.edu/ (post-natal day (P) 45). RNA was isolated by using TRIrea- ); R (http://www.R-project.org/). gent® (Sigma-Aldrich, Saint Louis, MO, USA) protocol. 2 μgs of total RNA were reverse transcribed with the Super- Results script III-First strand kit (Thermo Fisher Scientific, Identification of novel candidate genes for Parkinson’s Waltham, MA, USA). Quantitative PCR (qPCR) reactions disease were performed in triplicate, using gene specific primers The study aimed at the identification of genetic risk pro- (Table S2) and ITaq Universal Sybr Green Supermix (Bio- files for PD. The discovery cohort included 53 PD patients Rad, Hercules, CA, USA) following the manufacturer’s belonging to 26 families showing a dominant mode of in- directions. Results were normalized versus the heritance of the disease (Fig. 1). As validation cohort, we Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 6 of 18 used 1542 independent PD cases and 706 controls from model of inheritance. Unfortunately, in the remaining different data sets of European ancestry (see Patients and eight families we were not able to identify potential Methods for details) (Table S1) [8, 9, 16, 26]. causative variants, probably due to the highly stringent We first analysed the WES data (without applying any criteria used in the pipeline of analysis. Seven (PARK2, filter) of the 53 PD patients belonging to 26 families to PINK1, DJ-1(PARK7), LRRK2, FBXO7, DNAJC6, search for mutations in genes already known to be DNAJC13) out of the 26 identified genes were already associated to dominant form of PD such as, SNCA, known as causing PD, including familial recessive forms, VPS35 and LRRK2. Subsequently, we searched for large while three (HTRA2, EIF4G1, SNCAIP) genes were PD insertion/deletions in SNCA and LRRK2 genes by using candidate genes/at risk factors for which a clear link MLPA approach. All annotated variants were surveyed with the disease is not completely established [5, 6]. into the Parkinson’s disease Mutation Database (https:// Moreover, it is interesting to note that in several fam- grenada.lumc.nl/LOVD2/TPI/home.php) to confirm ilies, single mutations in recessive genes such as PARK2, their pathogenicity. One out of the 26 analyzed families PINK1, PARK7, were co-inherited with variants in other carried a pathogenic mutation in LRRK2 gene candidate genes, suggesting that they might play a role (c.G4322A, p.R1441H). as risk factors in the heterozygous status. The presence To search for novel candidate-disease genes we used a of a second mutation in these recessive genes including multistep filtering approach. We selected splicing vari- large deletions/insertions was excluded trough MLPA ants (regardless of their exonic or intronic position; approach. Sixteen out of the 26 genes analyzed were novel minor allele frequency (MAF) ≤ 0.001) and exonic dis- PD candidate genes involved in pathways potentially ruptive variants (MAF ≤ 0.001; CADD phred score ≥ 20; deregulated in PD such as mitochondrial metabolism and excluding synonymous changes), shared between af- oxidative stress (AIMP2, HMOX2, IMMT, MAN2C1, fected relatives within each of the 26 families (siblings or RHOT2, SLC25A39, TOMM22) [27–35], vesicular traffick- parent-offspring pairs), assuming a dominant mode of ing, microtubule dynamics, autophagy (CHMP1A, GIPC1, transmission of the disease. The variants were selected HSPA8, KIF21B, KIF24, SPTBN1, TMEM175, TVP23A) assuming a MAF ≤ 0.001 in public exome databases in- [36–42] or in SNCA gene expression (ZSCAN21) (Fig. 2a) cluding ESP6500si-v2, 1000 Genomes Project and ExAC. [43, 44]. STRING database analysis showed that nine out The cut off was set to 0.001, taking in account the preva- of the 16 novel genes (AIMP2, GIPC1, HSPA8, IMMT, lence of rare dominant diseases (1–2/10,000) and the RHOT2, SPTBN1, TMEM175, TOMM22, ZSCAN21) MAF of the most frequent mutation associated with late encoded for proteins interacting with known PD genes onset PD p.(G2019S) in the LRRK2 gene, also known as (Fig. 2a and b) [39, 45]. rs34637584, which shows a MAF ranging from 0.001 to Only very recently, IMMT, ZSCAN21 and TMEM175 0.0002 in different databases. Retained variants (6488) have been associated to PD [43, 46, 47] further support- were filtered for CADD phred score ≥ 20 (3712 ing the discovery pipeline employed in this study. remaining variants), corresponding to those variants To further support the genetic involvement of the which are predicted to be amongst the 1% most deleteri- 16 novel identified genes in PD pathogenesis, we ex- ous variants in the genome. Subsequently we compared tended the mutational analysis to 1542 independent data of the affected relatives of each family to select PD patients (including MNI (394 cases) and IPDGC shared variants. This analysis identified an average num- (1148 cases)) and 706 controls from different cohorts ber of 10 variants shared between PD relatives per fam- of European ancestry, to detect further disruptive var- ily. Segregation analysis was performed in additional iants within these genes and to compare the total affected relatives belonging to the same 26 PD families load of these variants between cases and controls. (1–3 individuals per family based on DNA availability) Overall data identified 256 different variants (MAF ≤ (Fig. 1). Considering the late age at onset of the disease 0.001; CADD phred score ≥ 20), of which 170 were unaffected family members were not investigated. This present only in cases, 61 only in controls and 25 were analysis disclosed 28 rare disruptive variants (23 non- shared between cases and controls. Remarkably, the synonymous, 2 stop-gain, 1 frameshift, 2 non-frameshift 170 different deleterious variants (153 non- deletions) laying in 26 genes, which were shared among synonymous, 5 stop-gain, 6 frameshifts, 3 non- familial PD cases in 18 out of the 26 analyzed families frameshift deletions and 3 splicing) were identified in (Table 1). Among these, in 10 families we found single 90 Italian patients (MNI cohort) and 153 patients of heterozygous deleterious variants in a single gene segre- the IPDGC cohort, confirming the extreme genetic gating with PD phenotype, supporting a dominant model heterogeneity at the basis of PD (Table S3). None of of inheritance. Instead, we identified 2 variants in 6 these variants was found in 706 healthy control sub- families and 3 variants in 2 families in different genes jects, including both IPDGC and Italian population segregating with PD phenotype suggesting a polygenic controls (namely MNI healthy subjects and Tuscans Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 7 of 18 Table 1 Rare disruptive variants (MAF ≤ 0.001, CADD phred≥20) in the 26 PD candidate genes identified in the discovery families CHR Genomic dbSNP Gene RefSEq Nucleotide AA Exonic Cl MAF Max in CADD Family position (hg19) Change Change Function public data sets phred 16 723,544 rs143816083 RHOT2 NM_138769 c.G1795T p.G599W NSV N 0.0001 25 1 3 184,039,223 rs746291399 EIF4G1 NM_198241 c.C851G p.S284C NSV N 0.00004 25 2 1 8,022,928 rs142405016 PARK7 NM_007262 c.G83A p.R28Q NSV 0.0002 35 5 7 6,062,997 rs749728733 AIMP2 NM_006303 c.T638C p.I213T NSV N 0.00002 27 7 2 86,408,450 rs201861204 IMMT NM_001100169 c.C91T p.R31C NSV N 0.0002 26 7 1 20,971,042 rs74315358 PINK1 NM_032409 c.G836A p.R279H NSV P 0.00004 26 7 16 89,717,990 CHMP1A NM_002768 c.C92T p.A31V NSV N NA 23 9 2 74,757,823 HTRA2 NM_013247 c.A586G p.N196D NSV N NA 23 9 3 132,222,165 DNAJC13 NM_015268 c.C4824G p.S1608R NSV N NA 20 10 17 42,398,033 SLC25A39 NM_001143780 c.C758A p.T253N NSV N NA 21 10 2 54,895,594 rs200093475 SPTBN1 NM_003128 c.C6983A p.T2328N NSV N 0.00004 27 12 22 39,078,021 rs200476832 TOMM22 NM_020243 c.C38T p.P13L NSV N 0.0002 25 12 19 14,589,354 GIPC1 NM_202470 c.875dupA p.K292fs Fs_ins N NA NA 13 6 162,394,354 rs114974496 PARK2 NM_004562 c.C714G p.C238W NSV P 0.00004 25 13 5 121,787,252 rs751412863 SNCAIP NM_005460 c.A2710G p.K904E NSV N 0.00004 24 14 12 40,704,237 rs34995376 LRRK2 NM_198578 c.G4322A p.R1441H NSV P 0.00001 27 14 7 99,654,828 ZSCAN21 NM_145914 c.C199T p.R67W NSV N NA 24 16 9 34,290,328 rs372797268 KIF24 NM_194313 c.T971C p.I324T NSV N 0.0001 26 17 16 10,911,991 TVP23A NM_001079512 c.G58T p.E20X stopgain N NA 38 18 1 65,858,474 DNAJC6 NM_001256864 c.C1829T p.P610L NSV N NA 23 19 15 75,652,021 MAN2C1 NM_001256494 c.1882_1887del p.628_629del NFs_del N NA NA 19 15 75,652,016 MAN2C1 NM_001256494 c.1890_1892del p.630_631del NFs_del N NA NA 19 22 32,870,999 FBXO7 NM_012179 c.C10T p.R4W NSV N NA 33 20 16 4,557,822 rs772245870 HMOX2 NM_001286271 c.C226T p.R76W NSV N 0.00001 34 20 1 200,957,957 rs751952433 KIF21B NM_001252100 c.G3235A p.A1079T NSV N 0.00002 29 21 11 122,931,465 HSPA8 NM_006597 c.G247A p.V83I NSV N NA 21 24 4 947,062 TMEM175 NM_032326 c.C547T p.R183X stopgain N NA 34 25 1 200,974,537 KIF21B NM_001252100 c.C631T p.R211C NSV N NA 34 26 CHR Chromosome, hg19 human genome build to which these variants are annotated, dbSNP reference number in SNP database, ref. seq reference number of the gene transcript, AA Change Amino acid change, CI Clinical interpretation, P Pathogenic, UP Uncertain pathogenicity, N Novel, PD Parkinson’s disease, IPDGC International Parkinson’s Disease Genetics Consortium, CADD phred Combined Annotation Dependent Depletion, Fs_ins Frame shift insertion, NFs_del Non Frame shift deletion, NSV Non-synonymous variant, MAF Minor Allele Frequency, NA Not Annotated, MAF max in public datasets: highest allelic frequency annotated in public databases including 1000 Genomes Project (AFR. AMR. EAS. EUR. SAS), ExAC browser (NFE. AFR. SAS. EAS and AMR), ESP6500si-v2 (European American and African American population) (TSI) “pseudo-controls “of the 1000 Genomes Project) Expression analysis in human, mouse and rat DA neurons [14]. Interestingly, significant enrichment of variants Expression analysis through quantitative PCR (qPCR) as- in these 16 genes was observed in patients compared says showed that the 16 novel PD genes were all tran- to controls (243 patients (15.7%) vs 69 controls scribed in the mesencephalon of adult mice at post-natal (9.7%); OR = 1.73 [1.3–2.29]; p = 0.0001 χ 2 = 14.01). day (P) 45 (Fig. 4). Then we studied through Noteworthy, all the affected residues identified in PD immunohistochemistry experiments whether five repre- patients occupy functionally important amino acid posi- sentative genes (TOMM22, GIPC1, ZSCAN21, tions, which are highly conserved amongst vertebrates SLC25A39 and HSPA8) were co-expressed with TH in (Fig. 3; Fig. S1). In summary, although functional data adult DA neurons of the substantia nigra (SN) and ven- are necessary to substantiate the causative role of the tral tegmental area (VTA). We first analyzed adult hu- variants identified, these results suggest a potential in- man SN neurons and we found that TH+ neurons co- volvement of the candidate genes in PD etiopathogenesis expressed all the five genes (Fig. 5a and Fig. S2a). This and indicate an extreme genetic heterogeneity of muta- analysis was repeated in P45 mouse (Fig. 5b-g and Fig. tions in PD patients. S2b-d) and P60 rat (Fig. S3) brains. In mouse mdDA Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 8 of 18 a b Fig. 2 Venn diagram and STRING analysis representing the logical relations between the 16 novel gene products and the network of PD- interacting genes products. a Venn diagram showing the 16 novel identified genes involved across different interconnected pathways: interactors of PD related genes, genes involved in mitochondrial metabolism and those involved in vesicular trafficking and autophagy. b Schematic representation of the network of interactions of the 16 novel PD gene products with known PD-associated genes. STRING database analysis identified 9 genes interacting with PD-associated genes neurons belonging to the anterior SN located in the pre- survival and/or DA neurotransmission. Therefore, to- tectum (Fig. 5b, c and Fig. S2b) or to the SN and VTA gether with previous data, this expression analysis fur- of mesencephalic origin (Fig. 5d-g and Fig. S2c, d) ther supports the possibility that mutations identified in showed that the expression of TOMM22, GIPC1, the 16 novel PD candidate genes may potentially predis- ZSCAN21, SLC25A39 and HSPA8 colocalized with most pose for PD occurrence. of the TH+ neurons (Fig. 5b-g and Fig. S2b-d). A similar result was observed when this expression analysis was Polygenic load analysis of rare exonic variants in familial performed in rat SN and VTA neurons of mesencephalic and sporadic cases origin (Fig. S3). These data indicate that at least five out Based on the observation that 8 out of the 18 families of of the 16 novel PD genes were expressed in mdDA neu- the discovery cohort carried two or more variants in the rons supporting the possibility that they may be cell- 26 PD genes (Table 1), we hypothesized that the genetic autonomously involved to provide mdDA neurons with combination of multiple rare variants in these genes may functional features potentially controlling identity and/or represent a specific genetic hallmark of PD risk. To test Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 9 of 18 Fig. 3 Graphical representation of the protein domains and mapped mutations. For those genes whose protein structure was available (AIMP2, GIPC1, KIF24, KIF21B, RHOT2, SLC25A39, SPTBN1, TMEN175), we report the protein domains and mapped amino acid changes identified in the cohort of 1542 patients. Amino acid changes are referred to by their single letter code this hypothesis, we analyzed in 394 unrelated and inde- corroborates the polygenic view of PD risk for both pendent PD patients (including 172 familial and 222 sporadic and familial cases. To measure the accuracy of sporadic cases from MNI data set) and in 203 controls the test, we applied the ROC curve and found a max- (MNI, Moli-sani, and TSI data sets) [8, 9, 14, 15] the co- imum sensitivity of 0.50 (based on the presence of at inheritance (per individual) of rare (MAF < 0.001) exonic least one variant (specificity 0.66)) and a maximum spe- (excluding synonymous) and splicing variants detected cificity > 0.93 (based on the co-occurrence of two or in the panel of the 26 genes (details in Patients and more variants (sensitivity of 0.17)) (AUC = 0.59, SE = ± Methods). For this analysis, we selected rare variants 0.020, p = 4.42 × 10− 6) (Fig. 6b). Overall, the presence of based on a frequency threshold (MAF ≤ 0.001), without two or more disruptive variants in the 26 candidate considering their degree of deleteriousness. This analysis genes represented the best tradeoff between sensitivity identified 220 different variants among cases and 65 in (0.17) and specificity (0.93). Moreover, considering that the controls; only 18 of them were shared between case is well established that rare variants in LRRK2, PINK1, and controls (Table S4). The co-inheritance of these var- PARK2 and PARK7 genes have a higher prevalence in iants is reported in Table 2 and Fig. 6a. We observed PD patients compared to healthy individuals [48–52], we that, approximately 17% of the PD patients carried two analyzed the presence of at least one rare variant or more variants (cases 17.3% vs controls 6.8%; OR = 3.3 (MAF < 0.001) in patients and controls taking into ac- [1.8–6.7]; p = 4.4 × 10− 5) (Table 2). Remarkably, sporadic count solely the 16 novel candidate genes. This analysis cases showed a significant distribution within the same confirmed an excess of patients carrying rare variants in class (sporadic cases 13.9% vs controls 6.8%, OR = 2.6 the 16 novel genes (cases 31.6% vs controls 23.1%; p = [1.3–5.1]; p = 0.005) (Table 2). These differences 0.014), as also reported in the gene identification para- remained statistically significant after Bonferroni correc- graph for rare deleterious variants (MAF ≤ 0.001; CADD tion for multiple testing of two contrasts (0 vs 1 and 0 phred score ≥ 20) selected in the 16 novel genes in cases vs ≥2 variants load classes) (α = 0.025). This finding and controls (243 (15.7%) vs 69 (9.7%); OR = 1.73 [1.3– Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 10 of 18 Fig. 4 Expression analysis of the novel identified Parkinson’s disease genes. qPCR experiment showed that the 16 novel genes were expressed in mesencephalon of P45 adult mice. Pitx3 was used as an example of mdDA neurons restricted gene expression. Gapdh was used to normalize the results. Data were shown as absolute values with standard deviation calculated from three different experiments. Genes were grouped on the basis of their expression level in three different scales 2.29]; p = 0.0001 χ 2 = 14.01), further supporting their variants in AIMP2, IMMT and PINK1 genes (Table 1), role into PD etiopathogenesis. while the GBA variant c.C1223T p.T408M (T369M, Moreover, the pool of rare variants identified in the previous nomenclature) was present in one of the two PD cases was different from those detected in controls siblings of family 16 who also carried the c.G4322A (Table S4). This suggests that, besides the load of rare p.R1441H mutation in LRRK2 gene and the deleterious variants (≥2), also the nature of the mutation is import- variant c.A2710G p.K904E in SNCAIP. In the independ- ant in predisposing/causing PD. ent cohort of PD cases and controls we found a signifi- Thus, the availability of a detailed database including cant distribution of GBA variants (42 cases (10.6%) vs 8 variants detected only in PD cases or only in controls controls (3.9%); p = 0.002, OR = 2.91 [1.34–6.32]). might be relevant to improve the operational character- Polygenic load analysis including multiple rare variants istics of our predictive protocol in the general in the 26 genes as well as rare pathogenic variants in GBA population. gene showed that, approximately 20% of the PD patients carried two or more variants (cases 20.5% vs controls Analysis of GBA gene 7.2%; OR = 3.59 [1.97–6.90]; p = 3.4 × 10− 6) (Table 2). Re- To test the involvement of GBA gene, that is the most markably, the increase in significance was mainly attribut- important genetic risk factor for PD, alone and in com- able to sporadic cases (sporadic cases increase from 13.9 bination with the 26 PD genes, we analyzed all GBA ex- to 18.5% with GBA mutations) (Table 2). onic variants annotated in the 26 late onset PD families and in the Italian validation cohort (394 independent PD Analysis of PD endophenotypes cases and 203 controls) (Table S5). In the discovery To investigate the relation of the polygenic variant load cohort the GBA variant c.G1093A p.E365K (E326K, pre- in the selected 26 candidate genes with PD endopheno- vious nomenclature) was present in the two affected sib- types, including both motor and non-motor symptoms, lings of family 7, co-inherited with single heterozygous we analyzed the cohort of 394 PD-MNI unrelated Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 11 of 18 Fig. 5 Expression analysis of TOMM22, GIPC1, ZSCAN21, SLC25A39 and HSPA8 in human and mouse DA neurons. a-g Immunohistochemistry experiments performed on human (a) and P45 mouse (b-g) adult brain sections showed that TOMM22, GIPC1, ZSCAN21, SLC25A39 and HSPA8 were co-expressed with TH in DA neurons of the human SN (a) and in mouse DA neurons of the SN (b-e) and VTA (f, g). Images in (c, e, g) correspond to magnification of the boxed area reported in (b, d, f). Scale bars correspond to 100 μm. Abbreviations: SN stands for substantia nigra and VTA for ventral tegmental area patients mentioned before, for which clinical assess- scales) [12] and presence of L-dopa induced Dyskinesia ments were available (see Patients and Methods for de- (LID), a common motor side effect of levodopa therapy tails on the scales used). More specifically, we tested in PD patients [8, 9]. Overall data show that the selected how the presence of 0, 1 ≥ 2 rare variants in the 26 PD genes might influence preferentially LID occurrence, genes affected PD age at onset (AAO), motor symptoms although the contrast would not survive correction for (UPDRS scale) [10, 13], non-motor symptoms (NMS multiple testing of five phenotypes (p 0.038; Fig. 6c; Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 12 of 18 Table 2 Polygenic variant load in the Italian cohort Series N variants PD cases FPD (n = 172) SPD Cnt. (n = 203) p OR 95%CI (n = 394) (n = 222) MAF ≤ 0.001 0 197 (50.0%) 82 (47.6%) 115 (51.8%) 135 (66.5%) baseline 1 129 (32.7%) 53 (30.8%) 76 (34.2%) 54 (26.6%) 0.014 1.6 1.1–2.4 ≥2 68 (17.3%) 37 (21.6%) 31 (13.9%) 14 (6.8%) 4.4 × 10−5 3.3 1.8–6.7 MAF ≤ 0.001 0 182 (46.2%) 79 (45.9%) 103 (46.4%) 128 (63.0%) baseline including variants in GBA gene 1 130 (33.2%) 53 (30.8%) 77 (35.1%) 59 (29.0%) 0.028 1.55 1.04–2.32 −6 ≥2 82 (20.5%) 40 (23.2%) 42 (18.5%) 16 (7.2%) 3.4 × 10 3.59 1.97–6.90 MAF Minor allele frequency, PD Parkinson’s disease, FPD Familial PD cases, SPD Sporadic PD cases, Cnt Controls fromTSI, Tuscan Italian population, MNI cohort and Moli-sani genetic biobank, p p-value calculated with Fisher Exact Probability Test, OR Odds Ratio, CI Confidence interval Fig. 6 Distribution of variants load, ROC curve and analysis of PD endophenotypes. a, b Very rare (MAF ≤ 0.001), exonic variants (exclusion of synonymous changes) were annotated in 26 PD genes in 394 unrelated Italian cases and in 203 controls. a The histogram shows the percentage of samples (cases in blue and controls in green) carrying 0, 1, 2, 3, 4 variants in the selected genes; b ROC curve and analysis of sensitivity and specificity. The test shows that the distribution is high significant and the test may predict the disease in about 17% of at risk individuals in the general population, carrying at least 2 variants, with specificity > 93%. c Percentage of PD cases manifesting L-dopa-induced dyskinesia (LID), and earlier age at onset (AAO) of Parkinson’s disease. d Polygenic variant load, including GBA variants, was inversely associated with age at PD onset at the nominal significance level (p 0.044). For each comparison, we set statistical significance threshold at p < 0.05. Statistical significance was reported with asterisk Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 13 of 18 Table S6A). When we took into account also GBA alters neuronal dendritic tree branching and spine for- variants, this contrast was not significant anymore, while mation [37]. Noteworthy, similarly to Kif21b knockout variant load was inversely associated with age at PD mice, which exhibit learning and memory deficits [37], onset at the nominal significance level (p 0.044; Table our patients carrying KIF21B mutations (Table S3) S6B; Fig. 6d). showed mild cognitive impairment assessed with Mon- treal Cognitive Assessment (MoCA) test (mean 23 SD Discussion 2). KIF24 is a centriolar kinesin that specifically regulates In this study, we discovered a set of 16 genes worthy of ciliogenesis by dynamically controlling microtubule future studies to confirm their role in onset or progres- polymerization [38]. Single Nucleotide Polymorphisms sion in PD. We identified 170 rare and deleterious (SNPs) located in the KIF24 gene have been associated variants, absent in healthy individuals, in 15.7% of PD to Frontotemporal Lobar Degeneration (FTLD) [59]. patients (243 PD cases out of 1542) supporting the ex- Interestingly, the patients of the MNI cohort carrying treme genetic heterogeneity of mutations in PD. mutations affecting the motor domain of KIF24 protein Although additional studies are needed to confirm the (Fig. 3) showed mild-severe cognitive impairment functional role of the novel identified genes in PD etio- (MoCA score ranging from 13 to 23). pathogenesis, a number of published studies support this HMOX exerts both pro- and anti-oxidant effects, hypothesis. which may influence the pathogenesis or progression of In particular, the transcription factor ZSCAN21 may Parkinson’s disease [32, 60]. α-Mannosidase (MAN2C1) directly control SNCA gene expression [28, 37], and the is the enzyme responsible for the partial demannosyla- c.C199T p.(R67W) mutation identified in this study tion occurring in the cytosol. Man2c1-deficient mice re- affects a highly conserved arginine residue necessary for vealed neuronal and glial degeneration [61]. Charged the formation of the ZSCAN21 functional homodimer multivesicular body protein 1A (CHMP1A) is a member [53]. TMEM175 is a lysosomal K+ channel that is re- of the ESCRT-III (endosomal sorting complex required quired to maintain membrane potential and pH stability for transport-III) complex which localizes to the nuclear in lysosomes as well as to control α-synuclein aggrega- matrix and regulates chromatin structure [36, 62]. tion [42, 54]. These functions may be affected by knock CHMP1A regulates proliferation of neuronal progenitor down of the gene TMEM175 [42]. Although TMEM175 cells [36, 62]. TVP23A is a novel gene reported in gene was previously identified in a genome wide association database (Gene ID: 780776) as involved in intracellular study of PD [47], detrimental mutations have not been vesicular transport. identified before this study. Previous studies identified the p.G2019S mutation in RHOT2 (MIRO2), IMMT and TOMM22 directly LRRK2 as a major causative event in 1–4% of PD pa- interact with PINK-1 and Parkin and influence both tients, depending on the analyzed population [63]. mitochondrial homeostasis and vulnerability of DA cells Within our cohort of patients, the p.G2019S mutation in to toxins [27–31, 55]. The aminoacyl-tRNA synthetase LRRK2 was detected in 8 out of 394 PD patients (2%). complex interacting multifunctional protein-2 (AIMP2) Moreover, other studies also correlated the severity but is a parkin substrate and its overexpression leads to a se- not the occurrence of the PD to the cumulative effect of lective, age-dependent, progressive loss of DA neurons LRRK2 mutations and rare variants in other PD related via activation of poly (ADP-ribose) polymerase-1 genes (e.g. GBA) [26, 64–66]. (PARP1) [35]. GIPC1 interacts with LRRK2 in DA cells In this study, we highlight the polygenic nature of both in vitro and in the mouse ventral mesencephalon in vivo sporadic and familial forms of PD, especially when the [45]. The Drosophila homolog of GIPC1 is expressed in familial forms come from small nuclear families. We the central brain of adult flies and its reduced expression identified a panel of 26 candidate genes carrying likely significantly affects locomotive activity and longevity pathogenic variants and observed the presence of mul- [56]. SLC25A39 is a mitochondrial carrier of manganese tiple rare variants (≥ 2) in these 26 genes in about 17% that promotes neuronal survival in Drosophila [34]. βII- of PD patients, including familial and sporadic cases. spectrin, encoded by SPTBN1, is required for axon This percentage increased to 20% of PD patients when growth and axonal transport of synaptic cargo [57]. we included rare pathogenic variants in GBA. Remark- Moreover, it specifically interacts with α-synuclein and ably, this increase was mainly attributable to sporadic its overexpression induces neurite formation [39]. The cases. chaperone-mediated autophagy (CMA) protein HSPA8/ This means that the molecular diagnosis of PD should HSC70 (heat shock protein 8) is essential in post-mitotic always take in account the analysis of multiple candidate neurons for diluting toxic intracellular components in- genes. However, although the evidence reported here cluding the disassembling of α-synuclein amyloid fibrils looks very promising, at this stage of advancement of [58]. KIF21B is a processive kinesin whose depletion our study we cannot offer any improvement in Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 14 of 18 practicing medicine, until further confirmatory studies Supplementary Information and functional validations are warranted. Undoubtedly, The online version contains supplementary material available at https://doi. org/10.1186/s13024-021-00455-2. the extreme genetic heterogeneity observed in PD pa- tients, everyone genetically and biologically unique, car- Additional file 1: Figure 1S. Evolutionary conservation study of the rying different genetic combinations of rare variants not likely pathogenic variants identified in the 16 novel Parkinson’s disease shared by the majority of PD patients, could cause great candidate genes. The affected residues are reported in bold underlined. Figure 2S. Single channel images of human and mouse uncertainty in the interpretation of the results. Never- immunostainings shown in Fig. 5. (a-d) Single channel images showing theless, at a molecular level, these findings offer a ra- expression of TH in combination with TOMM22, GIPC1, ZSCAN21, tional support to address additional genetic studies SLC25A39 and HSPA8 in human SN neurons (a) and in pretectal SN (b) mesencephalic SN (c) and VTA (d) mouse neurons. Images in (a) testing this protocol in larger cohort of PD patients and correspond to those reported in Fig. 5a; images in (b-d) correspond to controls. If these data were confirmed we could consider those reported in Fig. 5c, e, g, respectively. Figure 3S. Expression analysis the possibility of using this protocol to identify, with of TOMM22, GIPC1, ZSCAN21, SLC25A39 and HSPA8 in rat DA neurons. a- e Immunohistochemistry experiments performed on P60 rat brain high specificity, subjects potentially at risk of developing sections showed that TOMM22, GIPC1, ZSCAN21, SLC25A39 and HSPA8 PD by counting per individual the number of rare vari- were co-expressed with TH in mesencephalic DA neurons of the SN (b, ants in the panel of candidate genes. c) and VTA (d, e). Images in (c, e) correspond to magnification of the boxed area reported in (b, d). Sections in (a) are low magnification im- Moreover, our data also underline the importance of ages showing only TH immunostaining and corresponding to the same rare variants in the genetics of sporadic PD and corrob- sections reported in (b-e). Scale bars correspond to 100 μm. Abbrevia- orate the proposed approach as complementary to gen- tions: SN stands for substantia nigra and VTA for ventral tegmental area. In (c, e) are shown merge and single channel images. Table S1. Cohort ome wide association studies (GWAS) and to classical description. Table S2. List of primers used for gene expression analysis co-segregation analyses of NGS data to identify variants by quantitative PCR (qPCR). Table S3. Rare disruptive variants (MAF ≤ implicated in PD. 0.001, CADD phred ≥ 20) identified in the 16 novel PD candidate genes in the validation cohorts. Table S4. Rare variants (MAF ≤ 0.001) in the 26 The present study also reports, for the first time, stat- PD candidate genes in affected and healthy individuals of the Italian co- istical evidence that the polygenic variant load in the PD hort. Table S5. GBA variants annotated in Italian cohort of patients and candidate genes analyzed is associated with the occur- controls. rence of L-dopa induced Dyskinesia, an important motor side effect of PD treatment. Although it is tempting to Acknowledgements speculate about the translational relevance of this find- We thank the participating patients and their families. We are greteful to Dr. ing, caution is suggested in the interpretation since the Stefano Goldwurm and the “Parkinson Institute Biobank,” member of the Telethon Network of Genetic Biobank (biobanknetwork.telethon.it/) and the association was only nominally significant and disap- Moli-sani Biobank at Neuromed Biobanking Centre, member of BBMRI-Italy. peared when GBA variants were included in the analysis. IPDGC consortium members and affiliations: Interestingly, in the latter setting a high variant burden United Kingdom: Alastair J Noyce (Preventive Neurology Unit, Wolfson Institute of Preventive Medicine, QMUL, London, UK and Department of was associated with earlier PD onset, in line with previ- Molecular Neuroscience, UCL, London, UK), Rauan Kaiyrzhanov (Department ous evidence [67, 68]. of Molecular Neuroscience, UCL Institute of Neurology, London, UK), Ben Middlehurst (Institute of Translational Medicine, University of Liverpool, Liverpool, UK), Demis A Kia (UCL Genetics Institute; and Department of Molecular Neuroscience, UCL Institute of Neurology, London, UK), Manuela Conclusions Tan (Department of Clinical Neuroscience, University College London, Our findings encourage further studies to improve the London, UK), Henry Houlden (Department of Molecular Neuroscience, UCL Institute of Neurology, London, UK), Huw R Morris (Department of Clinical predictive/diagnostic power of this protocol for PD and Neuroscience, University College London, London, UK), Helene Plun-Favreau lead us to surmise that this strategy may hypothetically (Department of Molecular Neuroscience, UCL Institute of Neurology, London, be extended to other polygenic disorders with largely UK), Peter Holmans (Biostatistics & Bioinformatics Unit, Institute of Psycho- logical Medicine and Clinical Neuroscience, MRC Centre for Neuropsychiatric unknown genetic etiology, such as dyslexia [69]. Genetics & Genomics, Cardiff, UK), John Hardy (Department of Molecular Therefore, in the future studies on PD genetic aeti- Neuroscience, UCL Institute of Neurology, London, UK), Daniah Trabzuni (De- ology should be directed towards analyzing the polygenic partment of Molecular Neuroscience, UCL Institute of Neurology, London, UK; Department of Genetics, King Faisal Specialist Hospital and Research burden of rare variants with presumably large effect size, Centre, Riyadh, 11211 Saudi Arabia), John Quinn (Institute of Translational and validating them at the functional level. Medicine, University of Liverpool, Liverpool, UK), Vivien Bubb (Institute of Translational Medicine, University of Liverpool, Liverpool, UK), Kin Y Mok (De- partment of Molecular Neuroscience, UCL Institute of Neurology, London, Abbreviations UK), Kerri J. Kinghorn (Institute of Healthy Ageing, University College London, PD: Parkinson’s disease; LOPD: Late onset Parkinson’s disease; London, UK), Kimberley Billingsley (Institute of Translational Medicine, Univer- GWAS: Genome wide association studies; mdDA: Mesencephalic- sity of Liverpool, Liverpool, UK), Nicholas W Wood (UCL Genetics Institute; diencephalic DA; SNpc: Substantia nigra pars compacta; PIB: Parkinson and Department of Molecular Neuroscience, UCL Institute of Neurology, Institute Biobank; WES: Whole exome sequencing; MNI: Mediterranean London, UK), Patrick Lewis (University of Reading, Reading, UK), Sebastian Neurological Institute; NGS: Next Generation Sequencing; TR: Targeted Re- Schreglmann (Department of Molecular Neuroscience, UCL Institute of Neur- sequencing; SNVs: Single nucleotide variants; MAF: Minor Allele Frequency; ology, London, UK), Ruth Lovering (University College London, London, UK), CADD: Combined Annotation Dependent Depletion; TSI: Tuscan Italians; Lea R’Bibo (Department of Molecular Neuroscience, UCL Institute of Neur- qPCR: Quantitative PCR; ROC: Receiver Operating Characteristic; FPD: Familial ology, London, UK), Claudia Manzoni (University of Reading, Reading, UK), PD; SPD: Sporadic PD Mie Rizig (Department of Molecular Neuroscience, UCL Institute of Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 15 of 18 Neurology, London, UK), Mina Ryten (Department of Molecular Neuroscience, Neurological Disease; Taube/Koret Center for Neurodegenerative Disease Re- UCL Institute of Neurology, London, UK), Sebastian Guelfi (Department of search, San Francisco, CA, USA), Niccolo E. Mencacci (Northwestern University Molecular Neuroscience, UCL Institute of Neurology, London, UK), Valentina Feinberg School of Medicine, Chicago, IL, USA), Codrin Lungu (National Insti- Escott-Price (MRC Centre for Neuropsychiatric Genetics and Genomics, Cardiff tutes of Health Division of Clinical Research, NINDS, National Institutes of University School of Medicine, Cardiff, UK), Viorica Chelban (Department of Health, Bethesda, MD, USA), Andrew B Singleton (Laboratory of Neuroge- Molecular Neuroscience, UCL Institute of Neurology, London, UK), Thomas netics, National Institute on Aging, Bethesda, MD, USA), Sonja W. Scholz Foltynie (UCL Institute of Neurology, London, UK), Nigel Williams (MRC (Neurodegenerative Diseases Research Unit, National Institute of Neurological Centre for Neuropsychiatric Genetics and Genomics, Cardiff, UK), Karen E. Disorders and Stroke, Bethesda, MD, USA), Xylena Reed (Laboratory of Neuro- Morrison (Faculty of Medicine, University of Southampton, UK), Carl Clarke genetics, National Institute on Aging, Bethesda, MD, USA). Roy N. Alcalay (De- (University of Birmingham, Birmingham, UK and Sandwell and West Birming- partment of Neurology, College of Physicians and Surgeons, Columbia ham Hospitals NHS Trust, Birmingham, UK). University Medical Center, New York, NY, USA, Taub Institute for Research on France: Alexis Brice (Institut du Cerveau et de la Moelle épinière, ICM, Inserm Alzheimer’s Disease and the Aging Brain, College of Physicians and Surgeons, U 1127, CNRS, UMR 7225, Sorbonne Universités, UPMC University Paris 06, Columbia University Medical Center, New York, NY, USA). UMR S 1127, AP-HP, Pitié-Salpêtrière Hospital, Paris, France), Fabrice Danjou Canada: Ziv Gan-Or (Montreal Neurological Institute and Hospital, Department (Institut du Cerveau et de la Moelle épinière, ICM, Inserm U 1127, CNRS, UMR of Neurology & Neurosurgery, Department of Human Genetics, McGill Univer- 7225, Sorbonne Universités, UPMC University Paris 06, UMR S 1127, AP-HP, sity, Montréal, QC, H3A 0G4, Canada), Guy A. Rouleau (Montreal Neurological In- Pitié-Salpêtrière Hospital, Paris, France), Suzanne Lesage (Institut du Cerveau stitute and Hospital, Department of Neurology & Neurosurgery, Department of et de la Moelle épinière, ICM, Inserm U 1127, CNRS, UMR 7225, Sorbonne Human Genetics, McGill University, Montréal, QC, H3A 0G4, Canada), Lynne Universités, UPMC University Paris 06, UMR S 1127, AP-HP, Pitié-Salpêtrière Krohn (Montreal Neurological Institute and Hospital, Department of Neurology Hospital, Paris, France), Jean-Christophe Corvol (Institut du Cerveau et de la & Neurosurgery, Department of Human Genetics, McGill University, Montréal, Moelle épinière, ICM, Inserm U 1127, CNRS, UMR 7225, Sorbonne Universités, QC, H3A 0G4, Canada), Lynne Krohn (Montreal Neurological Institute and Hos- UPMC University Paris 06, UMR S 1127, Centre d’Investigation Clinique Pitié pital, Department of Neurology & Neurosurgery, Department of Human Genet- Neurosciences CIC-1422, AP-HP, Pitié-Salpêtrière Hospital, Paris, France), Maria ics, McGill University, Montréal, QC, H3A 0G4, Canada), Martinez (INSERM UMR 1220; and Paul Sabatier University, Toulouse, France), The Netherlands: Jacobus J van Hilten (Department of Neurology, Leiden Germany: Claudia Schulte (Department for Neurodegenerative Diseases, University Medical Center, Leiden, Netherlands), Johan Marinus (Department of Hertie Institute for Clinical Brain Research, University of Tübingen, and DZNE, Neurology, Leiden University Medical Center, Leiden, Netherlands). German Center for Neurodegenerative Diseases, Tübingen, Germany), Kathrin Spain: Astrid D. Adarmes-Gómez (Instituto de Biomedicina de Sevilla (IBiS), Hos- Brockmann (Department for Neurodegenerative Diseases, Hertie Institute for pital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Seville), Miquel Clinical Brain Research, University of Tübingen, and DZNE, German Center for Aguilar (Fundació Docència i Recerca Mútua de Terrassa and Movement Disorders Neurodegenerative Diseases, Tübingen, Germany), Javier Simón-Sánchez Unit, Department of Neurology, University Hospital Mutua de Terrassa, Terrassa, (Department for Neurodegenerative Diseases, Hertie Institute for Clinical Barcelona.), Ignacio Alvarez (Fundació Docència i Recerca Mútua de Terrassa and Brain Research, University of Tübingen, and DZNE, German Center for Movement Disorders Unit, Department of Neurology, University Hospital Mutua Neurodegenerative Diseases, Tübingen, Germany), Peter Heutink (DZNE, de Terrassa, Terrassa, Barcelona.), Victoria Alvarez (Hospital Universitario Central de German Center for Neurodegenerative Diseases and Department for Asturias, Oviedo), Francisco Javier Barrero (Hospital Universitario San Cecilio de Neurodegenerative Diseases, Hertie Institute for Clinical Brain Research, Granada, Universidad de Granada), Jesús Alberto Bergareche Yarza (Instituto de University of Tübingen, Tübingen, Germany), Patrizia Rizzu (DZNE, German Investigación Sanitaria Biodonostia, San Sebastián), Inmaculada Bernal-Bernal (Insti- Center for Neurodegenerative Diseases), Manu Sharma (Centre for Genetic tuto de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen del Rocío/CSIC/ Epidemiology, Institute for Clinical Epidemiology and Applied Biometry, Universidad de Sevilla, Seville), Marta Blazquez (Hospital Universitario Central de University of Tubingen, Germany), Thomas Gasser (Department for Asturias, Oviedo), Marta Bonilla-Toribio (Instituto de Biomedicina de Sevilla (IBiS), Neurodegenerative Diseases, Hertie Institute for Clinical Brain Research, and Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Seville), Juan A. DZNE, German Center for Neurodegenerative Diseases, Tübingen, Germany), Botía (Universidad de Murcia, Murcia), María Teresa Boungiorno (Fundació Docèn- United States of America: Mark R Cookson (Laboratory of Neurogenetics, cia i Recerca Mútua de Terrassa and Movement Disorders Unit, Department of National Institute on Aging, Bethesda, USA), Sara Bandres-Ciga (Laboratory of Neurology, University Hospital Mutua de Terrassa, Terrassa, Barcelona.) Dolores Neurogenetics, National Institute on Aging, Bethesda, MD, USA), Cornelis Buiza-Rueda (Instituto de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen Blauwendraat (National Institute on Aging and National Institute of Neuro- del Rocío/CSIC/Universidad de Sevilla, Seville), Ana Cámara (Hospital Clinic de Bar- logical Disorders and Stroke, USA), David W. Craig (Department of Transla- celona), Fátima Carrillo (Instituto de Biomedicina de Sevilla (IBiS), Hospital Universi- tional Genomics, Keck School of Medicine, University of Southern California, tario Virgen del Rocío/CSIC/Universidad de Sevilla, Seville), Mario Carrión-Claro Los Angeles, USA), Derek Narendra (Inherited Movement Disorders Unit, Na- (Instituto de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen del Rocío/ tional Institute of Neurological Disorders and Stroke, Bethesda, MD, USA), CSIC/Universidad de Sevilla, Seville), Debora Cerdan (Hospital General de Segovia, Faraz Faghri (Laboratory of Neurogenetics, National Institute on Aging, Be- Segovia), Jordi Clarimón (Memory Unit, Department of Neurology, IIB Sant Pau, thesda, USA; Department of Computer Science, University of Illinois at Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona and Cen- Urbana-Champaign, Urbana, IL, USA), J Raphael Gibbs (Laboratory of Neuro- tro de Investigación Biomédica en Red en Enfermedades Neurodegenerativas genetics, National Institute on Aging, National Institutes of Health, Bethesda, (CIBERNED), Madrid), Yaroslau Compta (Hospital Clinic de Barcelona), Monica Diez- MD, USA), Dena G. Hernandez (Laboratory of Neurogenetics, National Insti- Fairen (Fundació Docència i Recerca Mútua de Terrassa and Movement Disorders tute on Aging, Bethesda, MD, USA), Kendall Van Keuren-Jensen (Neuroge- Unit, Department of Neurology, University Hospital Mutua de Terrassa, Terrassa, nomics Division, TGen, Phoenix, AZ USA), Joshua M. Shulman (Departments Barcelona.), Oriol Dols-Icardo (Memory Unit, Department of Neurology, IIB Sant of Neurology, Neuroscience, and Molecular & Human Genetics, Baylor Col- Pau, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Bar- lege of Medicine, Houston, Texas, USA; Jan and Dan Duncan Neurological celona, and Centro de Investigación Biomédica en Red en Enfermedades Neuro- Research Institute, Texas Children’s Hospital, Houston, Texas, USA), Hirotaka degenerativas (CIBERNED), Madrid), Jacinto Duarte (Hospital General de Segovia, Iwaki (Laboratory of Neurogenetics, National Institute on Aging, Bethesda, Segovia), Raquel Duran (Centro de Investigacion Biomedica, Universidad de Gran- MD, USA), Hampton L. Leonard (Laboratory of Neurogenetics, National Insti- ada, Granada), Francisco Escamilla-Sevilla (Hospital Universitario Virgen de las tute on Aging, Bethesda, MD, USA), Mike A. Nalls (Laboratory of Neuroge- Nieves, Instituto de Investigación Biosanitaria de Granada, Granada), Mario netics, National Institute on Aging, Bethesda, USA; CEO/Consultant Data Ezquerra (Hospital Clinic de Barcelona), Cici Feliz (Departmento de Neurologia, Tecnica International, Glen Echo, MD, USA), Laurie Robak (Baylor College of Instituto de Investigación Sanitaria Fundación Jiménez Díaz, Madrid, Spain), Manel Medicine, Houston, Texas, USA), Jose Bras (Center for Neurodegenerative Sci- Fernández (Hospital Clinic de Barcelona), Rubén Fernández-Santiago (Hospital ence, Van Andel Research Institute, Grand Rapids, Michigan, USA), Rita Guer- Clinic de Barcelona), Ciara Garcia (Hospital Universitario Central de Asturias, Ovi- reiro (Center for Neurodegenerative Science, Van Andel Research Institute, edo), Pedro García-Ruiz (Instituto de Investigación Sanitaria Fundación Jiménez Grand Rapids, Michigan, USA), Steven Lubbe (Ken and Ruth Davee Depart- Díaz, Madrid), Pilar Gómez-Garre (Instituto de Biomedicina de Sevilla (IBiS), Hospital ment of Neurology, Northwestern University Feinberg School of Medicine, Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Seville), Maria Jose Chicago, IL, USA), Steven Finkbeiner (Departments of Neurology and Physi- Gomez Heredia (Hospital Universitario Virgen de la Victoria, Malaga), Isabel ology, University of California, San Francisco; Gladstone Institute of Gonzalez-Aramburu (Hospital Universitario Marqués de Valdecilla-IDIVAL, Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 16 of 18 Santander), Ana Gorostidi Pagola (Instituto de Investigación Sanitaria Biodonostia, Medical Research Center, Oulu University Hospital, Oulu, Finland), Ari Siitonen San Sebastián), Janet Hoenicka (Institut de Recerca Sant Joan de Déu, Barcelona), (Institute of Clinical Medicine, Department of Neurology, University of Oulu, Jon Infante (Hospital Universitario Marqués de Valdecilla-IDIVAL and University of Oulu, Finland; Department of Neurology and Medical Research Center, Oulu Cantabria, Santander, and Centro de Investigación Biomédica en Red en Enferme- University Hospital, Oulu, Finland). dades Neurodegenerativas (CIBERNED), Silvia Jesús (Instituto de Biomedicina de Nigeria: Njideka U. Okubadejo (University of Lagos, Lagos State, Nigeria), Sevilla (IBiS), Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Oluwadamilola O. Ojo (University of Lagos, Lagos State, Nigeria), Seville), Adriano Jimenez-Escrig (Hospital Universitario Ramón y Cajal, Madrid), Kazakhstan: Coordinator - Rauan Kaiyrzhanov (Department of Molecular Jaime Kulisevsky (Movement Disorders Unit, Department of Neurology, IIB Sant Neuroscience, UCL Institute of Neurology, London, UK), Chingiz Shashkin Pau, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Bar- (Kazakh National Medical University named after Asfendiyarov, Almaty, celona, and Centro de Investigación Biomédica en Red en Enfermedades Neuro- Kazakhstan), Nazira Zharkynbekova (South Kazakhstan Medical Academy, degenerativas (CIBERNED)), Miguel A. Labrador-Espinosa (Instituto de Biomedicina Shymkent, Kazakhstan), Vadim Akhmetzhanov (Astana Medical University, de Sevilla (IBiS), Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Astana Kazakhstan), Akbota Aitkulova (National Center for Biotechnology, Seville), Jose Luis Lopez-Sendon (Hospital Universitario Ramón y Cajal, Madrid), Astana, Kazakhstan; Al-Farabi Kazakh national university. Almaty Kazakhstan), Adolfo López de Munain Arregui (Instituto de Investigación Sanitaria Biodonostia, Elena Zholdybayeva (National Center for Biotechnology, Astana, Kazakhstan), San Sebastián), Daniel Macias (Instituto de Biomedicina de Sevilla (IBiS), Hospital Zharkyn Zharmukhanov (National Center for Biotechnology, Astana, Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Seville), Irene Martínez Kazakhstan), Gulnaz Kaishybayeva (Scientific and practical center “Institute of Torres (Department of Neurology, Instituto de Investigación Sanitaria La Fe, Hos- neurology named after Smagul Kaishybayev”, Almaty, Kazakhstan), Altynay pital Universitario y Politécnico La Fe, Valencia), Juan Marín (Movement Disorders Karimova (Scientific and practical center “Institute of neurology named after Unit, Department of Neurology, IIB Sant Pau, Hospital de la Santa Creu i Sant Pau, Smagul Kaishybayev”, Almaty, Kazakhstan), Dinara Sadykova (Astana Medical Universitat Autònoma de Barcelona, Barcelona, and Centro de Investigación Bio- University, Astana, Kazakhstan). médica en Red en Enfermedades Neurodegenerativas (CIBERNED)), Maria Jose Marti (Hospital Clinic Barcelona), Juan Carlos Martínez-Castrillo (Instituto Ramón y Cajal de Investigación Sanitaria, Hospital Universitario Ramón y Cajal, Madrid), Car- Authors’ contributions lota Méndez-del-Barrio (Instituto de Biomedicina de Sevilla (IBiS), Hospital Universi- TE, MC, MDE and AS designed the study. NM and SP recruited the patients, tario Virgen del Rocío/CSIC/Universidad de Sevilla, Seville), Manuel Menéndez performed clinical study and collected the data. MGR and AL performed González (Hospital Universitario Central de Asturias, Oviedo), Marina Mata (Depart- sample management and bio-banking. TN realized the clinical database. AG, ment of Neurology, Hospital Universitario Infanta Sofía, Madrid, Spain) Adolfo Mín- MGR and TE performed the analysis of WES data. TE, SS and SG performed guez (Hospital Universitario Virgen de las Nieves, Granada, Instituto de the NGS-TR analysis. DR performed the ROC analysis. The International Parkin- Investigación Biosanitaria de Granada), Pablo Mir (Instituto de Biomedicina de Se- son’s Disease Genetics Consortium (IPDGC) provided validation data set. LI villa (IBiS), Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Sev- and FG provided controls subjects. MGR performed the mutation analysis of ille), Elisabet Mondragon Rezola (Instituto de Investigación Sanitaria Biodonostia, the patients and the RT-qPCR, experiments. LGDG, DA and AS performed im- San Sebastián), Esteban Muñoz (Hospital Clinic Barcelona), Javier Pagonabarraga munohistochemistry experiments. AG, MGR, and TE, performed statistical (Movement Disorders Unit, Department of Neurology, IIB Sant Pau, Hospital de la analysis. TE and AS wrote the manuscript. All authors revised the final manu- Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, and Centro script. The author(s) read and approved the final manuscript. de Investigación Biomédica en Red en Enfermedades Neurodegenerativas (CIBERNED)), Pau Pastor (Fundació Docència i Recerca Mútua de Terrassa and Movement Disorders Unit, Department of Neurology, University Hospital Mutua Funding de Terrassa, Terrassa, Barcelona.), Francisco Perez Errazquin (Hospital Universitario This work was supported by Italian Ministry of Health grant n. RF 2019– Virgen de la Victoria, Malaga), Teresa Periñán-Tocino (Instituto de Biomedicina de 12370224 to TE, Current Research to TE and grant n. GR2016–02362247 to Sevilla (IBiS), Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, DR. This work was also supported by Italian Ministry of Economic Seville), Javier Ruiz-Martínez (Hospital Universitario Donostia, Instituto de Investiga- Development (M.I.S.E.) project n. F/0009/00X26. AG was supported by ción Sanitaria Biodonostia, San Sebastián), Clara Ruz (Centro de Investigacion Bio- Fondazione Umberto Veronesi. The funders had no role in study design, data medica, Universidad de Granada, Granada), Antonio Sanchez Rodriguez (Hospital collection and analysis, decision to publish, or preparation of the manuscript. Universitario Marqués de Valdecilla-IDIVAL, Santander), María Sierra (Hospital Uni- versitario Marqués de Valdecilla-IDIVAL, Santander), Esther Suarez-Sanmartin (Hos- pital Universitario Central de Asturias, Oviedo), Cesar Tabernero (Hospital General Availability of data and materials de Segovia, Segovia), Juan Pablo Tartari (Fundació Docència i Recerca Mútua de The datasets used and/or analyzed during the current study are available Terrassa and Movement Disorders Unit, Department of Neurology, University Hos- from the corresponding author on reasonable request. pital Mutua de Terrassa, Terrassa, Barcelona), Cristina Tejera-Parrado (Instituto de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen del Rocío/CSIC/Universi- dad de Sevilla, Seville), Eduard Tolosa (Hospital Clinic Barcelona), Francesc Valldeor- Declarations iola (Hospital Clinic Barcelona), Laura Vargas-González (Instituto de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Ethics approval and consent to participate Seville), Lydia Vela (Department of Neurology, Hospital Universitario Fundación Written informed consent was obtained from all participants. The ethical Alcorcón, Madrid), Francisco Vives (Centro de Investigacion Biomedica, Universidad board of the IRCCS Neuromed approved the study protocol (N°9/2015). de Granada, Granada). Austria: Alexander Zimprich (Department of Neurology, Medical University of Vienna, Austria). Consent for publication Norway: Lasse Pihlstrom (Department of Neurology, Oslo University Hospital, Not applicable. Oslo, Norway), Mathias Toft (Department of Neurology and Institute of Clinical Medicine, Oslo University Hospital, Oslo, Norway). Estonia: Sulev Koks (Department of Pathophysiology, University of Tartu, Tartu, Estonia; Department of Reproductive Biology, Estonian University of Competing interests Life Sciences, Tartu, Estonia), Pille Taba (Department of Neurology and The authors declare no competing interests. Neurosurgery, University of Tartu, Tartu, Estonia). Israel: Sharon Hassin-Baer (The Movement Disorders Institute, Department of Author details 1 Neurology and Sagol Neuroscience Center, Chaim Sheba Medical Center, IRCCS Istituto Neurologico Mediterraneo Neuromed, Pozzilli, Isernia, Italy. Tel-Hashomer, 5262101, Ramat Gan, Israel, Sackler Faculty of Medicine, Tel 2 National Research Council, Institute of Genetics and Biophysics “Adriano Aviv University, Tel Aviv, Israel). Buzzati-Traverso”, Naples, Italy. 3Department of Biomolecular Science, Finland: Kari Majamaa (Institute of Clinical Medicine, Department of University of Urbino Carlo Bò, Urbino, Italy. 4Research Center in Epidemiology Neurology, University of Oulu, Oulu, Finland; Department of Neurology and and Preventive Medicine (EPIMED), Department of Medicine and Surgery, University of Insubria, Varese, Italy. Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 17 of 18 Received: 9 February 2021 Accepted: 6 May 2021 complex multisystem syndrome due to dysfunction of the extracellular matrix. J Med Genet. 2017;54(10):710–20. https://doi.org/10.1136/ jmedgenet-2017-104555. 22. Napolitano F, Terracciano C, Bruno G, De Blasiis P, Lombardi L, Gialluisi A, References et al. Novel autophagic vacuolar myopathies: phenotype and genotype 1. Shulman JM, De Jager PL, Feany MB. Parkinson’s disease: genetics and features. Neuropathol Appl Neurobiol. 2021;00:1–15. https://doi.org/10.1111/ pathogenesis. Annu Rev Pathol Mech Dis. 2011;6(1):193–222. https://doi. nan.12690. org/10.1146/annurev-pathol-011110-130242. 23. Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, et al. STRI 2. Langston JW, Schüle B, Rees L, Nichols RJ, Barlow C. Multisystem Lewy body NG v11: protein-protein association networks with increased coverage, disease and the other parkinsonian disorders. Nat Genet. 2015;47(12):1378– supporting functional discovery in genome-wide experimental datasets. 84. https://doi.org/10.1038/ng.3454. Nucleic Acids Res. 2019 Jan;47(D1):D607–13. https://doi.org/10.1093/nar/ 3. Jellinger KA. The pathology of Parkinson’s disease. Adv Neurol. 2001;86: gky1131. 55–72. 24. R Core Team. R: A language and environment for statistical computing. 4. Kalinderi K, Bostantjopoulou S, Fidani L. The genetic background of Vienna: R Foundation for Statistical Computing; 2014. http://www.R- Parkinson’s disease: current progress and future prospects. Acta Neurol project.org/ Scand. 2016;134(5):314–26. https://doi.org/10.1111/ane.12563. 25. Di Salvio M, Di Giovannantonio LG, Acampora D, Prosperi R, Omodei D, 5. Verstraeten A, Theuns J, Van Broeckhoven C. Progress in unraveling the Prakash N, et al. Otx2 controls neuron subtype identity in ventral tegmental genetic etiology of Parkinson disease in a genomic era. Trends Genet. 2015; area and antagonizes vulnerability to MPTP. Nat Neurosci. 2010;13(12):1481– 31(3):140–9. https://doi.org/10.1016/j.tig.2015.01.004. 8. https://doi.org/10.1038/nn.2661. 6. Martin I, Dawson VL, Dawson TM. Recent advances in the genetics of 26. Lubbe SJ, Escott-Price V, Gibbs JR, Nalls MA, Bras J, Price TR, et al. Additional Parkinson’s disease. Annu Rev Genomics Hum Genet. 2011;12(1):301–25. rare variant analysis in Parkinson’s disease cases with and without known https://doi.org/10.1146/annurev-genom-082410-101440. pathogenic mutations: evidence for oligogenic inheritance. Hum Mol Genet. 7. Postuma RB, Berg D, Stern M, Poewe W, Olanow CW, Oertel W, et al. MDS 2016;25(24):5483–9. https://doi.org/10.1093/hmg/ddw348. clinical diagnostic criteria for Parkinson’s disease. Mov Disord. 2015;30(12): 27. Akabane S, Uno M, Tani N, Shimazaki S, Ebara N, Kato H, et al. PKA regulates 1591–601. https://doi.org/10.1002/mds.26424. PINK1 stability and Parkin recruitment to damaged mitochondria through 8. Gialluisi A, Reccia MG, Tirozzi A, Nutile T, Lombardi A, De Sanctis C, et al. phosphorylation of MIC60. Mol Cell. 2016;62(3):371–84. https://doi.org/10.1 Whole exome sequencing study of Parkinson’s disease and related 016/j.molcel.2016.03.037. endophenotypes in the Italian population. Front Neurol. 2020. https://doi. 28. John GB, Shang Y, Li L, Renken C, Mannella CA, Selker JM, et al. The org/10.3389/fneur.2019.01362. mitochondrial inner membrane protein mitofilin controls cristae morphology. 9. Tirozzi A, Modugno N, Palomba NP, Ferese R, Lombardi A, Olivola E, et al. Mol Biol Cell. 2005;16(3):1543–54. https://doi.org/10.1091/mbc.e04-08-0697. Analysis of genetic and non-genetic predictors of levodopa induced dyskinesia in Parkinson's disease. Front Pharmacol. 2021;12. https://doi.org/1 29. Wang X, Winter D, Ashrafi G, Schlehe J, Wong YL, Selkoe D, et al. PINK1 and 0.3389/fphar.2021.640603. Parkin target Miro for phosphorylation and degradation to arrest 10. Hentz JG, Mehta SH, Shill HA, Driver-Dunckley E, Beach TG, Adler CH. mitochondrial motility. Cell. 2011;147(4):893–906. https://doi.org/10.1016/j. Simplified conversion method for unified Parkinson’s disease rating scale cell.2011.10.018. motor examinations. Mov Disord. 2015;30(14):1967–70. https://doi.org/10.1 30. López-Doménech G, Covill-Cooke C, Ivankovic D, Halff EF, Sheehan DF, 002/mds.26435. Norkett R, et al. Miro proteins coordinate microtubule- and actin-dependent 11. Conti S, Bonazzi S, Laiacona M, Masina M, Coralli MV. Montreal cognitive mitochondrial transport and distribution. EMBO J. 2018;37(3):321–36. https:// assessment (MoCA)-Italian version: regression based norms and equivalent scores. doi.org/10.15252/embj.201696380. Neurol Sci. 2015;36(2):209–14. https://doi.org/10.1007/s10072-014-1921-3. 31. Bertolin G, Ferrando-Miguel R, Jacoupy M, Traver S, Grenier K, Greene AW, 12. Cova I, Di Battista ME, Vanacore N, Papi CP, Alampi G, Rubino A, et al. et al. The TOMM machinery is a molecular switch in PINK1 and PARK2/ Validation of the Italian version of the non motor symptoms scale for PARKIN-dependent mitochondrial clearance. Autophagy. 2013;9(11):1801–17. Parkinson’s disease. Park Relat Disord. 2017;34:38–42. https://doi.org/10.101 https://doi.org/10.4161/auto.25884. 6/j.parkreldis.2016.10.020. 32. Ayuso P, Martínez C, Lorenzo-Betancor O, Pastor P, Luengo A, Jiménez- 13. Shulman LM, Gruber-Baldini AL, Anderson KE, Vaughan CG, Reich SG, Jiménez FJ, et al. A polymorphism located at an ATG transcription start site Fishman PS, et al. The evolution of disability in Parkinson disease. Mov of the heme oxygenase-2 gene is associated with classical Parkinson's Disord. 2008;23(6):790–6. https://doi.org/10.1002/mds.21879. disease. Pharmacogenet Genomics. 2011;21(9):565–71. https://doi.org/10.1 14. The 1000 Genomes Project Consortium. A global reference for human 097/FPC.0b013e328348f729. genetic variation. Nature. 2015;526(7571):68–74. https://doi.org/10.1038/na 33. Wang L, Suzuki T. Dual functions for cytosolic α-mannosidase (Man2C1): its ture15393. down-regulation causes mitochondria-dependent apoptosis independently 15. Iacoviello L, Bonanni A, Costanzo S, De Curtis A, Di Castelnuovo A, Olivieri of its α-mannosidase activity. J Biol Chem. 2013;288(17):11887–96. https:// M, et al. The Moli-Sani project, a randomized, prospective cohort study in doi.org/10.1074/jbc.M112.425702. the Molise region in Italy; design, rationale and objectives. IJPH. 2007;4(2): 34. Slabbaert JR, Kuenen S, Swerts J, Maes I, Uytterhoeven V, Kasprowicz J, et al. 110–8. Shawn, the Drosophila homolog of SLC25A39/40, is a mitochondrial carrier 16. Jansen IE, Ye H, Heetveld S, Lechler MC, Michels H, Seinstra RI, et al. that promotes neuronal survival. J Neurosci. 2016;36(6):1914–29. https://doi. Discovery and functional prioritization of Parkinson's disease candidate org/10.1523/JNEUROSCI.3432-15.2016. genes from large-scale whole exome sequencing. Genome Biol. 2017;30:22. 35. Lee Y, Karuppagounder SS, Shin JH, Lee YI, Ko HS, Swing D, et al. Parthanatos 17. Li H, Durbin R. Fast and accurate short read alignment with burrows mediates AIMP2-activated age-dependent dopaminergic neuronal loss. Nat wheeler transform. Bioinformatics. 2009;25(14):1754–60. https://doi.org/10.1 Neurosci. 2013;16(10):1392–400. https://doi.org/10.1038/nn.3500. 093/bioinformatics/btp324. 36. Howard TL, Stauffer DR, Degnin CR, Hollenberg SM. CHMP1 functions as a 18. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. member of a newly defined family of vesicle trafficking proteins. J Cell Sci. A framework for variation discovery and genotyping using next- 2001;114(13):2395–404. https://doi.org/10.1242/jcs.114.13.2395. generation DNA sequencing data. Nat Genet. 2011;43(5):491–8. https:// 37. Muhia M, Thies E, Labonté D, Ghiretti AE, Gromova KV, Xompero F, et al. doi.org/10.1038/ng.806. The Kinesin KIF21B regulates microtubule dynamics and is essential for 19. Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second- neuronal morphology, synapse function, and learning and memory. Cell generation PLINK: rising to the challenge of larger and richer datasets. Rep. 2016;15(5):968–77. https://doi.org/10.1016/j.celrep.2016.03.086. GigaScience. 2015;4(1):7. https://doi.org/10.1186/s13742-015-0047-8. 38. Kobayashi T, Tsang WY, Li J, Lane W, Dynlacht BD. Centriolar kinesin Kif24 20. Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic interacts with CP110 to remodel microtubules and regulate ciliogenesis. variants from high-throughput sequencing data. Nucleic Acids Res. 2010; Cell. 2011;145(6):914–25. https://doi.org/10.1016/j.cell.2011.04.028. 38(16):e164. https://doi.org/10.1093/nar/gkq603. 39. Lee HJ, Lee K, Im H. α-Synuclein modulates neurite outgrowth by 21. Sampaolo S, Napolitano F, Tirozzi A, Reccia MG, Lombardi L, Farina O, et al. interacting with SPTBN1. Biochem Biophys Res Commun. 2012;424(3):497– Identification of the first dominant mutation of LAMA5 gene causing a 502. https://doi.org/10.1016/j.bbrc.2012.06.143. Gialluisi et al. Molecular Neurodegeneration (2021) 16:35 Page 18 of 18 40. Shang G, Brautigam CA, Chen R, Lu D, Torres-Vázquez J, Zhang X. Structure 60. Ryter SW, Tyrrell RM. The heme synthesis and degradation pathways: role in analyses reveal a regulated oligomerization mechanism of the PlexinD1/ oxidant sensitivity. Heme oxygenase has both pro and antioxidant GIPC/myosin VI complex. Elife. 2017;24:6. properties. Free Radic Biol Med. 2000;28(2):289–309. https://doi.org/10.1016/ 41. Fontaine SN, Zheng D, Sabbagh JJ, Martin MD, Chaput D, Darling A, et al. S0891-5849(99)00223-3. DnaJ/Hsc70 chaperone complexes control the extracellular release of 61. Paciotti S, Persichetti E, Klein K, Tasegian A, Duvet S, Hartmann D, et al. neurodegenerative-associated proteins. EMBO J. 2016;35(14):1537–49. Accumulation of free oligosaccharides and tissue damage in cytosolic α- https://doi.org/10.15252/embj.201593489. mannosidase (Man2c1)-deficient mice. J Biol Chem. 2014;289(14):9611–22. 42. Jinn S, Drolet RE, Cramer PE, Wong AH, Toolan DM, Gretzula CA, et al. https://doi.org/10.1074/jbc.M114.550509. TMEM175 deficiency impairs lysosomal and mitochondrial function and 62. Mochida GH, Ganesh VS, de Michelena MI, Dias H, Atabay KD, Kathrein KL, et al. increases α-synuclein aggregation. Proc Natl Acad Sci U S A. 2017;114(9): CHMP1A encodes an essential regulator of BMI1-INK4A in cerebellar 2389–94. https://doi.org/10.1073/pnas.1616332114. development. Nat Genet. 2012;44(11):1260–4. https://doi.org/10.1038/ng.2425. 43. Lassot I, Mora S, Lesage S, Zieba BA, Coque E, Condroyer C, et al. The E3 63. Healy DG, Falchi M, O'Sullivan SS, Bonifati V, Durr A, Bressman S, et al. ubiquitin ligases TRIM17 and TRIM41 modulate α-Synuclein expression by Phenotype, genotype, and worldwide genetic penetrance of LRRK2- regulating ZSCAN21. Cell Rep. 2018;25(9):2484–96. https://doi.org/10.1016/j. associated Parkinson's disease: a case-control study. Lancet Neurol. 2008; celrep.2018.11.002. 7(7):583–90. https://doi.org/10.1016/S1474-4422(08)70117-0. 44. Dermentzaki G, Paschalidis N, Politis PK, Stefanis L. Complex effects of the 64. Benitez BA, Davis AA, Jin SC, Ibanez L, Ortega-Cubero S, Pastor P, et al. ZSCAN21 transcription factor on transcriptional regulation of α-Synuclein in Resequencing analysis of five Mendelian genes and the top genes from primary neuronal cultures and in vivo. J Biol Chem. 2016;291(16):8756–72. genome-wide association studies in Parkinson's disease. Mol Neurodegener. https://doi.org/10.1074/jbc.M115.704973. 2016;11(1):29. https://doi.org/10.1186/s13024-016-0097-0. 45. Salašová A, Yokota C, Potěšil D, Zdráhal Z, Bryja V, Arenas E. A proteomic 65. Spitz M, Pereira JS, Nicareta DH, Abreu Gde M, Bastos EF, Seixas TL, et al. analysis of LRRK2 binding partners reveals interactions with multiple Association of LRRK2 and GBA mutations in a Brazilian family with signaling components of the WNT/PCP pathway. Mol Neurodegener. 2017; Parkinson's disease. Parkinsonism Relat Disord. 2015;21(7):825–6. https://doi. 12(1):54. https://doi.org/10.1186/s13024-017-0193-9. org/10.1016/j.parkreldis.2015.03.029. 46. Tsai PI, Lin CH, Hsieh CH, Papakyrikos AM, Kim MJ, Napolioni V, et al. PINK1 66. Escott-Price V, International Parkinson’s disease genomics consortium, Nalls phosphorylates MIC60/Mitofilin to control structural plasticity of MA, Morris HR, Lubbe S, Brice A, et al. Polygenic risk of Parkinson disease is mitochondrial crista junctions. Mol Cell. 2018;69(5):744–56. https://doi.org/1 correlated with disease age at onset. Ann Neurol. 2015;77(4):582–91. https:// 0.1016/j.molcel.2018.01.026. doi.org/10.1002/ana.24335. 47. Nalls MA, Pankratz N, Lill CM, Do CB, Hernandez DG, Saad M, et al. Large- 67. Blauwendraat C, Reed X, Krohn L, Heilbron K, Bandres-Ciga S, Tan M, et al. scale meta-analysis of genome-wide association data identifies six new risk Genetic modifiers of risk and age at onset in GBA associated Parkinson's loci for Parkinson’s disease. Nat Genet. 2014;56(9):1–7. disease and Lewy body dementia. Brain. 2020;143(1):234–48. https://doi. 48. Abou-Sleiman PM, Muqit MM, McDonald NQ, Yang YX, Gandhi S, Healy DG, org/10.1093/brain/awz350. et al. A heterozygous effect for PINK1 mutations in Parkinson's disease? Ann 68. Gan-Or Z, Liong C, Alcalay RN. GBA-associated Parkinson's disease and other Neurol. 2006;60(4):414–9. https://doi.org/10.1002/ana.20960. Synucleinopathies. Curr Neurol Neurosci Rep. 2018;18(8):44. https://doi.org/1 49. Djarmati A, Hedrich K, Svetel M, Lohnau T, Schwinger E, Romac S, et al. 0.1007/s11910-018-0860-4. Heterozygous PINK1 mutations: a susceptibility factor for Parkinson disease? 69. Gialluisi A, Andlauer TFM, Mirza-Schreiber N, Moll K, Becker J, Hoffmann P, Mov Disord. 2006;21(9):1526–30. https://doi.org/10.1002/mds.20977. et al. Genome-wide association scan identifies new variants associated with 50. Nuytemans K, Theuns J, Cruts M, Van Broeckhoven C. Genetic etiology of a cognitive predictor of dyslexia. Transl Psychiatry. 2019;9(1):77. https://doi. Parkinson disease associated with mutations in the SNCA, PARK2, PINK1, org/10.1038/s41398-019-0402-0. PARK7, and LRRK2 genes: a mutation update. Hum Mutat. 2010;31(7):763– 80. https://doi.org/10.1002/humu.21277. Publisher’s Note 51. Sun M, Latourelle JC, Wooten GF, Lew MF, Klein C, Shill HA, et al. Influence Springer Nature remains neutral with regard to jurisdictional claims in of heterozygosity for parkin mutation on onset age in familial Parkinson published maps and institutional affiliations. disease: the GenePD study. Arch Neurol. 2006;63(6):826–32. https://doi.org/1 0.1001/archneur.63.6.826. 52. Klein C, Djarmati A, Hedrich K, Schäfer N, Scaglione C, Marchese R, et al. PINK1, Parkin, and DJ-1 mutations in Italian patients with early-onset parkinsonism. Eur J Hum Genet. 2005;13(9):1086–93. https://doi.org/10.1038/ sj.ejhg.5201455. 53. Rimsa V, Eadsforth TC, Hunter WN. Structure of the SCAN domain of human paternally expressed gene 3 protein. PLoS One. 2013;8(7):e69538. https://doi. org/10.1371/journal.pone.0069538. 54. Lee C, Guo J, Zeng W, Kim S, She J, Cang C, et al. The lysosomal potassium channel TMEM175 adopts a novel tetrameric architecture. Nature. 2017; 547(7664):472–5. https://doi.org/10.1038/nature23269. 55. Ren Z, Yang N, Ji C, Zheng J, Wang T, Liu Y, et al. Neuroprotective effects of 5-(4-hydroxy-3-dimethoxybenzylidene)-thiazolidinone in MPTP induced parkinsonism model in mice. Neuropharmacology. 2015;93:e218. 56. Kim J, Lee S, Ko S, Kim-Ha J. dGIPC is required for the locomotive activity and longevity in Drosophila. Biochem Biophys Res Commun. 2010;402(3): 565–70. https://doi.org/10.1016/j.bbrc.2010.10.095. 57. Lorenzo DN, Badea A, Zhou R, Mohler PJ, Zhuang X, Bennett V. βII-spectrin promotes mouse brain connectivity through stabilizing axonal plasma membranes and enabling axonal organelle transport. Proc Natl Acad Sci U S A. 2019;116(31):15686–95. https://doi.org/10.1073/pnas.1820649116. 58. Gao X, Carroni M, Nussbaum-Krammer C, Mogk A, Nillegoda NB, Szlachcic A, et al. Human Hsp70 Disaggregase reverses Parkinson’s-linked α-Synuclein amyloid fibrils. Mol Cell. 2015;59(5):781–93. https://doi.org/10.1016/j.molcel.2 015.07.012. 59. Venturelli E, Villa C, Fenoglio C, Clerici F, Marcone A, Benussi L, et al. Is KIF24 a genetic risk factor for Frontotemporal lobar degeneration? Neurosci Lett. 2010;482(3):240–4. https://doi.org/10.1016/j.neulet.2010.07.047.