Introduction

Parkinson's disease (PD) is a neurodegenerative disease presenting with bradykinesia, rigidity, tremor, and gait disturbance as motor features and is one of the most common movement disorders in people older than 65 years1. Although most of the cases are sporadic, it is estimated that 3–5% of patients have monogenic causes linked to known Parkinson's disease genes, while an additional 16–36% of patients have genetic risk variants that increase the risk of PD1. The pathophysiological mechanisms of those genes mostly involve excessive pathological aggregation of alpha-synuclein, mitochondrial dysfunctions, or deficits in lysosomal degradation to maintain cellular homeostasis2,3. A well-known mutation associated with significant PD risk is the mutations in the glucocerebrosidase gene (GBA1), leading to reduced lysosomal enzyme activity and build-up of glucocerebrosides4. Recently, a lysosome-related gene, PSAP, has also been reported to have an association with PD risk5,6.

The PSAP gene encodes the prosaposin protein and includes 14 exons. After translation, prosaposin is cleaved into 4 active units, saposin A to D, which serve as sphingolipid activators required for certain lysosomal enzymes7. For example, saposin C can activate glucocerebrosidase, which is encoded by GBA1. Mutations in saposin C and B domain of PSAP are implicated in Gaucher’s disease and metachromatic leukodystrophy, respectively8,9. Notably, several pathogenic mutations in the saposin D domains of PSAP were firstly reported in three families with autosomal dominant inheritance in PD6. They further analyzed the saposin D domain, which is encoded in exon 10 to 14, in a combined cohort enrolling 440 Japanese and 705 Taiwanese patients with sporadic PD (SPD). Notably, two intronic variants, rs885828 in intron 12, and rs4747203 in intron 11, were found to be associated with an increased risk of SPD.

Later, more studies have been conducted to investigate the association between PSAP variants and risk of PD. However, the results were conflicting. No significant associations of rs885828 and rs4747203, intronic variants, have been reported using genome-wide association studies and Sanger sequencing of PSAP exon 10–14 in SPD10,11. On the other hand, Lin et al. discovered a protective effect of these two intronic variants12. Chen et al. also reported a significantly reduced risk of PD in patients of SPD who carried PSAP rs4747203 variant13.

Based on aforementioned inconsistent results regarding the effect of PSAP intronic variants on the risk of PD, this study aims to clarify the role of PSAP variants in different subgroups of PD among Taiwanese population by enrolling patients with early-onset PD (EOPD), who have the onset age of PD before 50 years old, and patients with familial PD (FPD), who have a family history of PD, patients with sporadic PD (SPD), who have no family history and have disease onset older than 50 years of age, and health controls (HC). Both exonic variant analysis and intronic variant association studies were utilized. The first part of the study looks at exonic mutations of PSAP in EOPD and FPD, in comparison with HC from a public database. The second part of the study looks at intronic variants in EOPD, FPD and SPD, in comparison with HC from NTUH. Meta-analysis is also conducted to summarize the current findings on the association between PSAP variants and risk of PD.

Results

Demographic characteristics

A total of 887 patients, including 485 SPD, 219 EOPD, and 183 FPD (117 ADPD, 66 ARPD) were enrolled. A total of 711 HC, including 300 from Taiwan Biobank and 412 from NTUH were also enrolled. The mean age, gender proportion, disease onset, and disease status were summarized in Table 1. The mean age at examination was 58.9 ± 11.1 years in HC, 67.0 ± 4.5 years in SPD, 46.7 ± 7.5 years in EOPD and 63.2 ± 11.0 in FPD. The mean age at disease onset was 62.4 ± 3.4 years in SPD, 41.4 ± 8.1 years in EOPD and 58.7 ± 12.3 in FPD. The median Hoehn and Yahr (H&Y) scale was 2 in all PD subgroups.

Table 1 Participant demographics.
Full size table

Mutation analysis of PSAP

Exonic mutations of PSAP in EOPD and FPD patients were compared with HC from community-based public datasets including gnomAD EAS (East Asia population) and Taiwan Biobank. Three pathogenic exonic variants (p.C412Y, p.Q453P, p.C451_L477del) were not found in EOPD and FPD groups6. Two novel exonic variants were identified, including p.A146E on exon 5 (c.437C > A), and p.Y248C on exon 7 (c.743A > G), in 2 cases of EOPD. Those two variants are not reported in public datasets. Based on the multiple sequence alignment analysis with other species, most of them were shown to be evolutionarily conserved amino acid positions in PSAP (Supplementary Fig. S1). According to the criteria of the American College of Medical Genetics, the variant of p. Y248C variant is classified as likely pathogenic variants and the variant of p.A146E is classified as variants of uncertain significance14. We were unable to perform the co-segregation analysis of these two exonic variants in proband’s family due to noncompliance and loss follow up. All variants identified in our cohort are provided in Supplementary Table S1.

Genotypic and allelic analysis of 4 intronic variants around Saposin D domain

The Chi-Square test for Hardy–Weinberg equilibrium were all non-significant in all four intronic variants (p-value 0.42 in rs4747203, 0.11 in rs885828, 0.67 in rs142614739 and 0.85 in rs749823). The allelic and genotypic frequency of 4 intronic variants (adjusted by age and sex), rs4747203, rs142614739/rs74733861 (this variant is referred to rs142614739 in the following text), rs749823 and rs885828, between each group of PD patients and HC were compared respectively (Table 3). The results of genotypic and allelic frequencies of rs4747203, rs142614739, rs749823 and rs885828 in each group were shown in Table 2. In the logistic regression adjusting age and sex, the rs142614739 was significantly associated with increased disease risk in EOPD by additive model (odds ratio [OR] = 1.85, 95% confidence interval [95% CI]) = 1.33–2.58, p < 0.001). There was no statistically significant association of rs4747203 and rs885828 between total and other subgroups of PD patients and HC, although there was a protective trend of rs885828 variants in EOPD (OR = 0.77, 95% CI = 0.60–0.99, p = 0.04). Also, the additive models of rs4747203 showed a statistically non-significant protective effect in EOPD (OR = 0.79, 95% CI = 0.62–1.02, p = 0.07).

Table 2 The genotypic, dominant, and recessive genotypes and allelic frequency of four intronic variants between patients with different subgroups of Parkinson’s disease and healthy controls.
Full size table

Meta-analysis of 4 intronic variants

Pooled analysis of the effects of rs142614739, rs885828, rs4747203 and rs749823 on the risk of PD was performed after incorporating multiple studies (Fig. 1). The sample size of each study was summarized in Supplementary Table S3. As shown in Fig. 1A, rs142614739 was significantly associated with the risk of PD (common effect OR = 1.24, 95% CI = 1.07–1.44; random effect OR = 1.24, 95% CI 1.03–1.50). In contrast, rs885828, rs4747203 and rs749823 were not significantly associated with the risk of PD (Fig. 1B–D).

Figure 1
figure 1

Meta-analysis of the effects of rs142614739, rs4747203, rs885828 and rs749823 on the risk of Parkinson’s disease: rs142614739 was significantly associated with the risk of PD in both common effect (OR = 1.24, 95% CI = 1.07–1.44) and random effect (OR = 1.24, 95% CI = 1.03–1.50) in patients with total PD including EOPD, SPD, and FPD (A). In contrast, rs4747203 (common effect: OR = 0.99, 95% CI = 0.97–1.01; random effect: OR = 0.96, 95% CI = 0.91–1.01), rs885828 (common effect: OR = 1.00, 95% CI = 0.98–1.01; random effect: OR = 1.00, 95% CI = 0.94–1.05) and rs749823 (common effect: OR = 0.99, 95% CI = 0.95–1.03; random effect: OR = 1.23, 95% CI = 0.92–1.43) were not associated with the risk of PD (B–D).

Full size image

Discussions

In this study, we performed a comprehensive sequencing of 14 exons in PSAP and discovered 2 novel PD-related exonic variants in patients with FPD. Furthermore, we investigated 4 intronic variants around saposin D domain and we found that rs142614739 significantly increased the risk of younger-onset (< 50 years old) PD patients (Table. 3, Fig. 1) rather than in patients with sporadic PD (onset age > 50 years-old) or familial PD (positive family history). The meta-analysis further confirmed the risk effect of rs142614739 and concluded the neutral effect of the previously reported intronic variants (rs885828, rs4747203, and rs749823).

Table 3 The association tests of each variant in the risk of Parkinson’s disease by genotypic and allelic frequency analysis.
Full size table

The biogenesis of saposin A-D cleaved from full-length pro-peptide Prosaposin presented in endosomes intracellularly7. In particular, saposin C can activate glucocerebroside in lysosome, which is linked to the pathogenesis of PD3. Moreover, the PSAP gene was found to be upregulated in the substantia nigra of PD versus controls15. Previous evidence implies that any deficit in the functional chain of prosaposin could potentially impair the endo-lysosomal pathways, and eventually result in PD-associated biochemical aberrances and phenotypes. There were 2 exonic variants disclosed in this study. The missense mutation p. Y248C is located at the saposin B domain. Previous study has shown that functional deficit of saposin B leads to diminished degradation of glycosphingolipids in the brain16. The other missense mutation p.A14E is located between saposin A and B domain. We speculated that this mutation may interfere with the peptide splicing of prosaposin or influence other post-translation modification processes. Unfortunately, co-segregation analysis cannot be performed due to loss of follow-up so that the pathogenicity of those variants cannot be further validated. Furthermore, only exons in patients with EOPD and FPD were sequenced while patients with SPD and HC were not. Rare variant burden analysis could not be performed to compare the frequency of rare variant between cases and controls. We could only use the public information from Taiwan Biobank and GnomAD as an alternative to reflect the possible genetic background of general populations. Functional study, such as changed lysosomal burden in the cell model and motor impairment in the animal model, can also be considered in the future6.

The role of two intronic variants (rs885828, and rs4747203) remain controversial6. However, numerous studies reported neutral or even protective effect in these 2 intronic variants10,11,12,13,17. A protective effect of rs4747203 was reported and a risk effect of rs14271473912,13 was observed. While the largest cohort in the meta-analysis from 23andMe including 1,400,000 HC and 37,688 PD did not show no association of rs4848203, rs885828, and rs749823 variants with the risk of PD (Supplementary Table S3)18. It is possible that intronic variants may influence mRNA splicing or post-transcriptional regulation19. However, the functional impact of intronic variants in mRNA splicing is still a challenge in the genomic research field19. More importantly, although some software is developed to predict the cleavage site of peptides by various proteases in databases20, it is still hard to answer whether these intronic variants located between exon 10–11 are able to impair the topology of pre-peptide and/or derived peptides intracellularly. In order to support the genomic impact at the cellular level, the expression level of RNA or protein products may reflect the difference21. Facchi et al. has found that PSAP RNA expression level in the whole blood was significantly correlated with rs885828 genotypes11. Although the exact mechanisms that these intronic variants are involved remain to be discovered, these findings provide an alternative method to examine the biological effect of those intronic variants.

In fact, the presence of variants in rs885828 and rs4747203 are in linkage disequilibrium as mentioned by Sosero et al17. This phenomenon is more prominent in Caucasian (r2 = 1, D′ = 1) and less commonly seen in Japanese healthy controls (D′ = 0.7) in rs88582817. Therefore, ethnic variation in the allele frequency of these two intronic variants between Asian and Caucasian may account for the different analytic results. Interestingly, few except one study mentioned the existence of PSAP rs142614739, which is located between rs885828 and rs4747203 at the intron 10–1112 and has been merged into rs74733861 in dbSNP and ClinVar Miner22,23. This variant is a 6-nucleotide deletion between exons 10 and 11, which may impair the following mRNA splicing or peptide cleavage. Lin et al. also found that two out of four common haplotypes, including H2 (rs885828 T allele, rs142614738 wt, rs4747203 G allele) and H4 (rs885828 C allele, rs142614739 del, rs4747203 A allele), were significantly associated with PD12. As for the European population, there is a moderate linkage disequilibrium between rs142614739 and rs885828 (R2 = 0.3579, D’ = 1) and between rs142614739 and rs4747203 (R2 = 0.3579, D′  = 1) in the database of LDpop Tool (https://ldlink.nih.gov/?tab=ldpop). Thus, the significance of rs142614739 in the European population is still uncertain. What we found is that rs142614739 del is associated with EOPD in our cohort and meta-analysis. One interpretation is that this intronic variant could affect the splicing or the stability of mRNA. The other explanation is that those intronic variants are makers of nearby genetic mutation due to linkage disequilibrium.

In the study published by Sosero et al., the variants rs4747203 and rs885828 are not associated with the risk of PD and the age at onset17. Nevertheless, we found that rs142614739 was associated with increased risk of EOPD, while the effect is less prominent in sporadic PD older than 50 years old at the onset of symptoms. In the study published by Oji et al., they also mentioned that the included Japanese cohort has significantly younger average age at onset (50.9 ± 13.7 years)6. This age difference might explain the different effect of the intronic variant compared with other studies. A larger association study in multi-ethnic populations are needed to determine whether PSAP variants are associated with the younger onset of PD.

There are limitations in this study. First, although we enroll patients with different age of onset and presence of family history or not, the overall sample size is relatively small in PD patients which may undermine the statistical power to evaluate the effect of each variant. Second, the female is predominant in our control group, which may indicate more healthy volunteers in our hospital. Previous study has shown that gender may influence the health care-seeking behavior, especially in women participants24. However, no obvious gender effect has been reported in the polymorphism distribution of PSAP gene before. We also adjust sex and age in statistical analysis. Therefore, we presumed that the gender imbalance may not have a major impact on the results. Third, lack of burden analysis and co-segregation study makes the clinical effect of those rare exonic variants uncertain. Fourthly, although this study aimed to investigate the associations between pathological traits and genetic variations, environmental factors may also influence the risks of developing PD. Thus the results should be carefully interpreted because respective contributions of genetic and environmental factors were not conducted in this study. Lastly, there were also several limitations in the meta-analysis. The age of disease onset was not provided in all studies included in the meta-analysis. Therefore, the association between age of disease onset or other phenotypic traits with PSAP variants cannot be analyzed in our meta-analysis. Moreover, rs142614739 has been only published in 2 studies so the overall population size is still low in the meta-analysis. The exact mechanism of these intronic variants remains uncertain and warrants functional studies of PSAP in the lysosomal pathway and even alpha-synuclein propagation pathways.

In conclusion, two exonic variants in the PSAP gene were highlighted, one of which is possibly pathogenic. The analysis of intronic variants reveals that rs142614739 was associated with significantly increased risk of EOPD, suggesting a potential modification of disease age at onset. On the other hand, rs885828 and rs4747203 do not have a significant effect on the risk of PD in this study and meta-analysis. Further functional study on the mechanism of rs142614739 leading to increased risk of PD is highly warranted.

Method

Participants and study design

The study design of this case-control association study is summarized in Fig. 2. Patients with SPD, EOPD, FPD, and HC were enrolled at the Excellent Centre for Parkinson and Movement Disorder, Department of Neurology, National Taiwan University Hospital (NTUH). The diagnosis of PD was based on the UK Parkinson’s Disease Society Brain Bank clinical diagnostic criteria by two movement disorder specialists (M-C Kuo, R-M Wu)25. The EOPD is defined as the age at symptom onset before 50 years old. The SPD is classified as the patients who had presentation of parkinsonian motor features at age older than fifty. The FPD is defined as the presence of a positive family history of PD in the proband’s family in either autosomal dominant (AD) or autosomal recessive (AR) patterns. Volunteers or spouses of the diseases were enrolled and examined by doctors or nurses at the Excellent Centre for Parkinson and Movement Disorder, Department of Neurology, NTUH to exclude major systemic or neurodegenerative diseases. All participants provided written informed consent. Ethics approval was obtained from the Ethics Committee of the National Taiwan University Hospital, Taipei, Taiwan.

Figure 2
figure 2

Participants and study design: Three subgroups of Parkinson’s disease (PD) patients, including familial PD (FPD), early onset PD (EOPD) and old-onset/sporadic PD (SPD) were enrolled. Exons 1 to 14 were Sanger sequenced in FPD and EOPD. Exonic and intronic variants we identified were compared with public data on GnomAD and Taiwan Biobank first. Among four intronic variants on intron 11 and 12 on the saposin D domain of PSAP gene, including rs4747203, rs885828, rs142614739, and rs749823, were sequenced in additional cohorts including SPD and both community- and hospital-based controls to compare the genotypic and allelic frequencies between groups.

Full size image

DNA preparation and sequencing

Sanger sequencing was performed for PSAP exon 1–14 in patients of FPD and EOPD to determine if there is any pathogenic mutation hotspot and PSAP intron 11–12 in all PD patients and HC for the analysis of 4 previously-reported intronic variants. In the groups of FPD and EOPD, pathogenic/likely pathogenic variants in other previously known PD causative genes were excluded by multiplex ligation-dependent probe amplification and targeted next-generation sequencing panel26. For DNA preparation, 10 ml of venous blood was collected and genomic DNA was extracted by Blood/Cell DNA Maxi Kit (Geneaid GDM025). DNA was then amplified by polymerase chain reaction (PCR). Different primers were used to amplify genomic segments of each PSAP exon or intron. Primer sequences are listed in Supplementary Table S2. PCR reactions were performed in a total volume of 25 μl, containing 12.5 μl of PCR Master Mix (Cat.No: M7502, Promega), 9.5 μl nuclease-free water, 10 μM of forward-primer, 10 μM of reverse primer, and 200 ng of genomic DNA. PCR conditions were 2 min 30 s of initial denaturation at 95 °C, and 35 cycles of 30 s of denaturation at 95 °C, 30 s at annealing temperature, and 35 s of extension at 72 °C. PCR products were purified using Presto™ 96 Well PCR Cleanup Kit (96DFH10, Geneaid). Direct sequencing was performed using BigDye™ Terminator chemistry (version 3.1; Life Technologies) as recommended by the manufacturer. Then DNA was precipitated by ethanol and loaded on an Applied Biosystem 3730 DNA Analyzer (Life Technologies). For sequencing analysis, Chromas and Unipro UGENE were used.

Exonic mutation analysis

Mutations were compared with the public reference database, including gnomAD (https://gnomad.broadinstitute.org/) and Taiwan Biobank (https://taiwanview.twbiobank.org.tw/variant.php)27. Variants identified were annotated according to GenBank accession number NM_002778, corresponding to Homo sapiens PSAP, transcript variant 1. For the analysis of exons in PSAP, missense variants were only considered deleterious when they were predicted to be damaging by SIFT and PolyPhen-2. Rare variants were further classified as pathogenic, likely pathogenic, variant of uncertain significance (VUS), or likely benign or benign, according to the recommendations of the American College of Medical Genetics14. Amino acid sequence comparison across different species is done to examine the evolutionary conservation via multiple sequence alignment by constraint-based alignment tool for multiple protein sequences (COBALT)28.

Statistical analysis

Demographics of participants in each group are summarized in mean ± standard deviation for continuous variables and median ± interquartile range for the H&Y scale. Chi-square test was used to test for the Hardy–Weinberg equilibrium of four intronic variants in the controls. To adjust age and sex, multivariate logistic regression incorporating age, sex and each variant status was used. Additive inheritance model (genotype AA as 2, AB as 1 and BB as 0) was used in the regression. The Wald test was applied to test for the significance of the genotype variable. We used Bonferroni correction to adjust for multiple testing for four variants, after which a p-value less than 0.0125 (0.05 divided by 4) was regarded as statistically significant, as detailed in a previous study27. The strength of the association between the variants and Parkinson’s disease susceptibility was evaluated by OR.

Meta-analysis

The publications were searched from the databases of Medline, PubMed, and ISI Web of Knowledge. The last search was updated on Oct, 31, 2023. The search terms “prosaposin gene”, “saposin”, “Parkinson’s disease”, “SNPs” and “polymorphism” were used. The references of associated publications were carefully evaluated to obtain as many as possible. Only those published studies with full-text articles in English were included in this meta-analysis. For overlapping and republished studies, only the first one with the largest samples was included. The details of our inclusion criteria were as follows: (1) evaluation of the polymorphism of PSAP rs4747203, rs885828, rs749823 and rs142614739 on PD risk; (2) well-designed case–control studies using similar research methods (with strict case selection, international diagnosis standard of PD, and appropriate statistical methods); (3) with sufficient published data to estimate an OR with 95% CI; (4) written in English and (5) containing detailed and useful genotype frequencies or alleles frequencies. Studies without clear and full information of samples were excluded. Two reviewers extracted information from all eligible publications independently, according to the above inclusion criteria. The sample size of each included sizes are listed in the Supplementary Table S3.

The risk of PD associated with PSAP rs142614739, rs4747203, and rs885828 was estimated for each study by ORs with 95% CI. For all studies, we estimated the association under the allele contrast model due to the availability of the published data. The Q-statistic was used to investigate the degree of heterogeneity between the trials, and a p value of more than 0.10 for the Q-test indicated a lack of heterogeneity among studies. We used the mixed-effect model, combining common effects and random effects from each of the studies based on the DerSimonian and Laird (DL) method. All of the meta-analyses were performed with R “meta” package29.