Evaluation of Induced Pluripotent Stem Cell‐Derived Dopaminergic Neurons from Siblings with Gaucher Disease Discordant for Parkinsonism

Abstract

Background

In Gaucher disease (GD), glucocerebrosidase (GCase) deficiency results from biallelic pathogenic GBA1 variants. While GBA1 variants are a major risk factor for Parkinson's disease (PD), most patients with GD never develop parkinsonism.

Objectives

To understand factors impacting PD penetrance in patients with GD by comparing induced pluripotent stem cell (iPSC)‐derived dopaminergic neurons (DANs) from GD siblings discordant for PD.

Methods

iPSCs, reprogrammed from two different sibling pairs where both siblings had GD but only one developed PD, were differentiated into DANs. In one family, DAN enrichment was achieved via geneticin selection and proteomic evaluations were performed.

Results

GCase and lipid substrate levels were similar in GD and GD/PD DANs. After geneticin selection, proteomic analysis of the enriched DANs showed upregulation of molecular chaperones in the GD/PD line.

Conclusion

PD discordance in both sets of GD siblings did not correlate with GCase or lipid substrate levels in DANs, implicating the involvement of other modifiers. Published 2025. This article is a U.S. Government work and is in the public domain in the USA. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

1. Introduction

Dopaminergic neurons (DANs) differentiated from patient‐derived induced pluripotent stem cells (iPSCs) serve as useful in vitro models to study Parkinson's disease (PD), as they partially recapitulate aspects of this multifactorial disorder, including the role of genetic vulnerability. ¹ Variants in GBA1, encoding glucocerebrosidase (GCase), are the most common known genetic risk factor for PD, ² , ³ present in 5%–10% of patients. ⁴ , ⁵ , ⁶ The lysosomal storage disorder Gaucher disease (GD) results from biallelic pathogenic GBA1 variants. However, despite having significantly reduced GCase activity, most patients with GD never develop parkinsonism. ⁷ Given the clinical heterogeneity among individuals sharing identical GBA1 variants, including variable lipid substrate (glucosylceramide [GlcCer] and glucosylsphingosine [GlcSph]) and GCase levels in the peripheral organs, ⁸ we explored whether GCase and lipid substrate levels specifically in DANs might determine PD penetrance among patients with GD.

Studying DANs differentiated from iPSCs derived from individuals discordant for PD provides a means to explore aspects of disease pathogenesis in affected cell types while still preserving the genetic background. ⁹ This is particularly true when evaluating discordant siblings in families. Over a 15‐year period, we closely followed siblings with GD discordant for PD, performing repeated clinical evaluations, ¹⁰ and generated iPSC lines from two of these families. Here we present the molecular characterization of iPSC‐derived DANs from five individuals in the two families, together with relevant longitudinal clinical information.

The inherent heterogeneity in DAN differentiation efficiency across iPSC lines and in culture batches, ¹¹ , ¹² together with the low purity of cultured DANs, ¹³ posed significant technical challenges to our approach. To improve the reliability of downstream comparisons between the siblings, in Family 2 we incorporated a strategy to enable DAN enrichment via geneticin selection by inserting neo into the TH locus, placing its expression under the control of the endogenous TH promotor. Proteomic analyses of enriched DANs revealed lysosomal features associated with GD, as well as changes in molecular chaperones that might potentially contribute to the PD discordance.

2. Methods

2.1. iPSC Lines

The five participants provided informed consent under National Institutes of Health (NIH) clinical protocol 86HG0096. Sanger sequencing of GBA1 was performed as described previously. ¹⁴ Fibroblasts were reprogrammed into iPSCs using non‐integrative Sendai virus.

2.2. iPSC Maintenance and Differentiation into DANs

iPSCs were maintained in Essential 8 Medium (Cat.# A1517001; Thermo Fisher Scientific, Waltham, MA, USA). DANs were differentiated as described previously with minor adaptations. ¹⁵

2.3. Enrichment of DANs

The iPSC lines from the three sisters in Family 2 were edited to enable DAN enrichment with geneticin selection as described previously. ¹⁶

2.4. Proteomics and Lipidomics

Whole cell proteomics were performed at day 75 of DAN maturation as previously presented. ¹⁷ GlcCer and GlcSph were measured with supercritical fluid chromatography‐mass spectrometry (SFC‐MS/MS) separating glucose and galactose species at the Lipidomics Shared Resource (LSR) at the Medical University of South Carolina Hollings Cancer Center. Phosphate (Pi) was also determined in the samples to facilitate data normalization.

Detailed methods can be found in the Supplementary Materials and the antibodies used are specified in Table S1.

3. Results

3.1. Clinical Presentation of the Siblings with GD Discordant for Parkinsonism

Family 1 (Fig. 1A) included two brothers (GBA1: N370S/N370S; p.N409S/p.N409S). Their mother was affected with PD. The older brother (HT932) was diagnosed with GD at age 49 years when he presented with bone pain, hepatosplenomegaly, and thrombocytopenia. He was started on enzyme replacement therapy (ERT), and his clinical course stabilized. He is now 73 years old and has been followed multiple times in our clinic over the past two decades without evidence of parkinsonism to date. Two separate F18‐dopa positron emission spectrometry (PET) scans showed normal uptake. His GD diagnosis prompted evaluation of his younger brother (HT810) who was also diagnosed with GD at age 47 years. Although he was found to have mild splenomegaly and osteoporosis, ERT was never initiated. At age 52 years, HT810 developed left hand tremor and bradykinesia, subsequently exhibiting shuffling gait with freezing episodes and stooped posture. Initially he responded to levodopa, but required periodic dose increases over time, causing motor complications. He also suffered from anxiety, depression, had two episodes of marked psychosis, and, later in the disease course, developed dementia. He died at age 71 years. An autopsy revealed severe and diffuse Lewy body pathology and neuronal loss. Skin fibroblasts were obtained from HT932 at age 56 years and HT810 at 54 years (after PD diagnosis) and reprogrammed to iPSC lines.

FIG 1.

Characterization of patient‐induced pluripotent stem cell (iPSC)‐derived dopaminergic neurons (DANs). (A) Summary of the clinical features of each iPSC donor. (B,C) Western blot showing levels of glucocerebrosidase (GCase), tyrosine hydroxylase (TH), and glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) in the five lines of DANs with quantification. (D) GCase activity levels in DANs from the five lines. (E) Glucosylceramide (GluCer) and glucosylsphingosine (GluSph) levels in HT707, HT708, and HT711 DANs. The indicated values are the average from three independent measurements from biological replicates (unit: ×10⁻¹ pmole/per nmole phosphate). DX, diagnosis; N/A, not applicable; M, male; F, female; ERT, enzyme replacement therapy. [Color figure can be viewed at wileyonlinelibrary.com]

Family 2 (Fig. 1A) included two sisters (HT707 and HT708) (GBA1: N370S/c.203del + R257X; p.N409S/c.203del + p.R296X) who were evaluated longitudinally at the National Institutes of Health (NIH). ¹⁰ The older sister, HT707, was diagnosed with GD at age 12 years, exhibiting skeletal involvement, anemia, thrombocytopenia, and massive organomegaly, leading to a partial splenectomy. Although later, ERT resulted in improved hematologic parameters and organ size, skeletal complications and pulmonary hypertension developed. Assessment by a movement disorder specialist at ages 60 and 62 years revealed no evidence of parkinsonism. She died at age 63 years due to metastatic osteosarcoma, and no Lewy body pathology nor abnormal basal ganglia were observed at autopsy. Her younger sister, HT708, diagnosed with GD at age 8 years, presented with hepatosplenomegaly, thrombocytopenia, and fatigue that responded to ERT. At age 55 years, she had onset of left‐hand tremor. Examination at age 56 years revealed bradykinesia, rigidity, and a unilateral left pill‐rolling tremor, all responsive to L‐dopa. After 6 years of treatment, she was enrolled in a clinical trial for continuous infusion of L‐dopa subcutaneously, decreasing her experienced motor fluctuations. Eight years after PD onset, her Montreal Cognitive Assessment (MoCa) score was 26 points. HT707 and HT708 also had a middle sister, HT711, who carries no GBA1 mutations and had no signs of PD during examinations at ages 58 and 63 years (Fig. 1A). Skin fibroblasts were obtained from HT707 at age 60 years, HT708 at 56 years (after PD diagnosis), and HT711 at 58 years, and reprogrammed to iPSC lines.

3.2. Comparable Levels of GCase, GCase Activity, and Lipid Substrates in GD and GD/PD DANs

All five patient‐derived iPSC lines were differentiated into DANs following a well‐established protocol and matured to day 65–75 for evaluation. ¹⁵ Across the five DANs, GCase protein levels were downregulated in HT707, HT708, HT810, and HT932 compared with HT711. No substantial differences were observed between the two clinically discordant siblings in each family (Fig. 1B,C), suggesting that the GBA1 genotype has a more substantial impact on GCase levels than PD status. Furthermore, the GCase activity (Fig. 1D), while low in each individual with GD, did not correlate with PD status. Particularly, HT810 DANs had higher GCase activity than HT932. GlcCer and GlcSph were measured by SFC‐MS/MS in DANs derived from Family 2. GlcSph was elevated in both HT707 and HT708, consistent with GlcSph being a biomarker for GD (Fig. 1E).

Substantial variation in DAN differentiation efficiency was noted among the iPSC lines, as indicated by variable levels of tyrosine hydroxylase (TH) (Fig. 1B). Because comparable DAN differentiation efficiency among the lines could be important for further characterization, we introduced a DAN enrichment strategy ¹⁶ into the iPSC lines from Family 2.

3.3. DAN Enrichment Via Geneticin Selection in Family 2 Enhanced Downstream Comparisons Between Siblings

We addressed our concern regarding DAN differentiation variability inherent to iPSC lines by introducing geneticin selection for TH‐expressing (TH+) DANs into iPSC lines HT707, HT708, and HT711. A neo gene was inserted into the TH locus after exon 14 following a P2A self‐cleaving peptide as described elsewhere, ¹⁶ generating the edited iPSC lines HT707 TH‐neo, HT708 TH‐neo, and HT711 TH‐neo. After DAN differentiation and geneticin selection, TH expression in the three edited lines was clearly more similar (Fig. 2A). To further characterize geneticin‐enriched DANs, we generated cellular proteomic datasets and evaluated the results using dimensionality reduction by Uniform Manifold Approximation and Projection (UMAP) (Fig. 2B). HT707 TH‐neo, HT708 TH‐neo, and HT711 TH‐neo DANs clustered together after geneticin selection, demonstrating reduced heterogeneity among DANs. Notably, HT711 and HT711 TH‐neo clustered well, as did HT707 TH‐neo and HT707 TH‐neo + geneticin, suggesting that DAN differentiation was not altered by the insertion of neo or geneticin selection. GCase levels in the enriched DANs from the GD sisters remained similar and were considerably lower than in the wild‐type sister (Fig. 2A). No difference was observed in α‐synuclein levels measured in the 1% Triton‐X100 soluble fraction (Fig. 2A).

FIG. 2.

Phenotyping enriched dopaminergic neurons (DANs). (A) Western blotting analysis of enriched DANs from HT707 TH‐neo, HT708 TH‐neo, and HT711 TH‐neo showing tyrosine hydroxylase (TH), α‐synuclein (α‐syn), glucocerebrosidase (GCase), and glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) as the loading control. (B) Uniform Manifold Approximation and Projection (UMAP) clustering DAN cellular proteomics shows improved comparability after geneticin selection. (C) A network of lysosomal hydrolases that is downregulated in HT707 TH‐neo and HT708 TH‐neo DANs. (D) A network of molecular chaperones that is upregulated in HT708 TH‐neo DANs compared with HT707 TH‐neo. (E) Top 20 enriched gene ontology (GO) term pathways comparing HT708 TH‐neo and HT707 TH‐neo DANs. [Color figure can be viewed at wileyonlinelibrary.com]

To characterize neuronal proteomic changes related to GD, we identified shared changes in geneticin‐enriched DANs from HT707 TH‐neo and HT708 TH‐neo compared with HT711 TH‐neo. These included 80 upregulated and 118 downregulated proteins (cut‐off: fold change 1.5, adj. P‐value < 0.01). Protein–protein interaction network functional enrichment analysis using the STRING database revealed a network of lysosomal proteins among the shared downregulated proteins, including GCase, Cathepsin A (CTSA), Cathepsin L (CTSL), and β‐galactosidase (GLB1) (Fig. 2C), indicating broad lysosomal changes related to GD. To delineate neuronal changes specifically linked to PD, we investigated the differences between HT707 TH‐neo and HT708 TH‐neo. A network of molecular chaperones, including HSPA6 of the 70 kDa heat shock protein family (HSP6) and its co‐chaperone BAG3, were upregulated in HT708 TH‐neo, reflecting disruptions in protein homeostasis potentially contributing to the PD phenotype (Fig. 2D). Gene ontology (GO) term enrichment analysis of the proteomics data also indicated major differences in synapse assembly between HT707 TH‐neo and HT708 TH‐neo, identifying “synaptic signaling” and “synapse assembly” among the top downregulated biological categories (Fig. 2E).

4. Discussion

Despite two decades of intense research, genetic factors that determine PD penetrance among individuals carrying GBA1 variants are still largely unknown. ¹⁸ , ¹⁹ iPSCs derived from genetically‐related individuals with disparate clinical outcomes may provide a feasible method to recapitulate cell‐type specific pathological changes. Here, we present five iPSC lines derived from two families sharing a genetic predisposition to PD but with disparate clinical outcomes, accompanied by associated longitudinal clinical data. In Family 1, HT932 has been followed at regular intervals for the past 18 years, and at age 73 years shows no prodromal features of PD, unlike his brother, who died at age 71 years after a 19‐year‐long PD course. In Family 2, comprehensive clinical information, including a post mortem neuropathologic examination, confirmed that HT707, unlike her sister with PD, had no indication of parkinsonism, despite harboring identical biallelic GBA1 pathogenic variants that resulted in low GCase expression in her iPSC‐derived DANs.

In Family 2, we addressed the substantial DAN differentiation heterogeneity inherent in iPSC lines, even among lines derived from genetically‐related individuals, by incorporating our newly developed enrichment strategy. ¹⁶ We verified comparable TH expression across the enriched DAN lines and generated cellular proteomics datasets. UMAP confirmed improved clustering in the DANs among the three siblings after enrichment. Proteomic analysis revealed that the GD siblings shared downregulated lysosomal hydrolases, while the PD sibling had a network of molecular chaperones upregulated. These findings emphasize the critical importance of securing comparable differentiation efficiency for uncovering disease‐relevant phenotypes based on comparisons between individuals. The similar α‐synuclein levels observed among the enriched DANs (Fig. 2A) indicate that under our culture conditions and at the maturation stage examined, α‐synuclein levels do not reflect PD status. Our observation is consistent with previous studies also reporting similar α‐synuclein levels in iPSC‐DANs from controls and patients with PD carrying GBA1 variant N370S, ²⁰ and from a set of monozygotic twins sharing GBA1 genotype N370S/WT but clinically discordant for PD. ²¹ A recent study reported accumulation of phosphorylated α‐synuclein, the classical hallmark of PD pathology, in neurons converted directly from patient fibroblasts, which maintained age‐related properties of the donors. ²² However, this pathological phenotype was absent in neurons generated from iPSCs from the same patients, potentially due to the reprogramming process resetting the age and epigenetic signature of the patients. ²² The iPSC approach employed here likely shares the same limitation; and in the future, converting fibroblasts from our patients directly to neurons may be warranted. Because the upregulation of HSPA6 and BAG3 in HT708 occurred in the absence of α‐synuclein accumulation, we speculate that these elevations may reflect impaired protein homeostasis, potentially contributing to PD pathogenesis (Fig. 2D), but the lack of an α‐synuclein phenotype limited us from evaluating the effects of HSPA6 and BAG3 modulation on its levels.

In both families, DA neurons from the siblings with GD exhibited comparable GCase levels, which were lower than in the non‐GD control HT711. HT707 and HT708 DANs showed an equivalent, yet small accumulation of GluSph (Fig. 1E). This aligns with their diagnosis of non‐neuronopathic GD, where no GD‐related neuronal pathology and little lipid substrate accumulation is observed in post mortem brain tissues. ²³ , ²⁴ Since DANs from HT707 and HT708, and HT810 and HT932 had similar GCase and lipid substrate levels, neither explains the PD discordance in the siblings, implicating genetic modifiers acting through other pathways. An improved understanding of these genetic modifiers is critical for identifying individuals with GBA1 variants who may benefit from enhanced PD monitoring and early clinical intervention.

5. Author Roles

(1) Research Project: A. Study Concept and Design, B. Organization, C. Data Acquisition; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript Preparation: A. Writing of the First Draft, B. Editing the Final Version.

E.H.: 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B

K.R.: 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B

G.P.: 1A, 1B, 1C, 2A, 2B, 2C

Y.H.: 1B, 1C, 2A, 2B, 2C

Z.L.: 1B, 1C, 2A, 2B, 2C

C.M.: 1B, 1C, 2A, 2B, 2C

K.A.: 1B, 1C, 2A, 2B, 2C

Y.A.Q.: 1B, 1C, 2A, 2B, 2C

E.R.: 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B

G.L.: 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B

N.T.: 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B

E.S.: 1A, 2C, 3A, 3B

Y.C.: 1A, 2C, 3A, 3B

Supporting information

Data S1. Supporting Information.

Table S1. Antibodies used.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.