Available via license: CC BY 4.0
Content may be subject to copyright.
PINK1 and MTF2 Modulators Synergize Cell Transplantation
Therapy in Parkinson’s Disease
Yichen Gao
United World College Changshu China, Jiangsu, China
ycgao23@uwcchina.org
Abstract. Parkinson’s Disease (PD), the second most prevalent neurodegenerative disease, has
the key pathological feature of selective degeneration of dopaminergic neurons (DANs). Current
cell therapy based remedy of PD centered on DAN transplantation, whether stem cell derived,
or from fetal tissues, yet the possible dysfunction underlying in the microenvironment and non-
neuronal mechanisms may impede this solution. The unresolved question is if microenvironment
impairments and other non-cell-autonomous signaling may affect the healthy, transplanted
DANs, triggering survival rate reduction. We hypothesized that non-neuronal mechanisms do
exist thus specific drugs should be incorporated along with DANs during transplantation. Here
we performed single cell RNA sequencing (scRNA-seq) analysis using Seurat and AI based
package Monocle3 to confirm the existence of non-cell-autonomous signaling and revealed
PINK1 and MTF2 as the main culprits through the comparison of initial differentiation dataset
and long term transplantation dataset of transplanted DANs. Targeting PINK1, we identified the
drug MTK458 and validated its effectiveness through AI based drug screening and molecular
docking. Together, these findings supported MTK458 to be an effective co-transplant material
designated for alleviating the abnormality in non-neuronal mechanisms or microenvironment.
Keywords: Parkinson’s Disease, Co-Transplantation, Cell Therapy, Dopaminergic Neuron,
Bioinformatics, Single-cell RNA-Sequencing, K-Nearest Neighbors, AlphaFold 3, Machine
Learning.
1. Introduction
Parkinson’s disease (PD) asserts its prominence as the second most pervasive neurodegenerative disease,
affecting more than 8.5 million people worldwide by 2019 [1]. Patients often manifest motor symptoms
— tremor, rigidity, slow movements and balancing difficulties — as well as non-motor symptoms like
sleep disruption (Figure. 1A). PD patient’s neurological analysis found an accumulation of Lewy bodies
formed by ɑ-synuclein deposition and loss of dopaminergic neurons (DANs) in midbrain substantia
nigra pars compacta, and ventral tegmental area [2] (Figure. 1B). The loss of DANs is the direct
contributor to PD’s symptoms, and thus is the key for devising a viable treatment.
For decades, standard treatments focused on dopamine replacement therapy (DRT) to regulate levels
of dopamine, most commonly using the dopamine precursor L-DOPA penetrating the blood-brain
barrier to elevate dopamine levels [3]. Emerging therapies include deep brain stimulation therapy (DBS),
employing electrical impulses to restore abnormal neural activities [4]. However, neither of these PD
treatments tackles the root factor of dopaminergic neuron loss. Despite PD’s proliferating cases, these
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
© 2024 The Authors. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0
(https://creativecommons.org/licenses/by/4.0/).
259
current clinical treatments are said to be palliative: remedial actions that do not inhibit the process of
degeneration. In the search for more fundamental treatments, cell therapies, involving transplantation of
cellular materials into patients, uncovered a new approach of restoring degenerate neurons.
Figure 1. Biological Mechanisms of PD. (A) Loss of dopaminergic neurons. (B) ɑ-synuclein deposition
and formation of Lewy Bodies. Modified based on Goldoni et al., 2022.
While DAN transplantation gained a surge in popularity in recent years, the unresolved difficulty of
diminished cell survival rate, typically lower than 10% in both autologous and allogeneic models, can
potentially render it ineffective [5]. The low survival rate of DAN transplantation is the major hurdle,
commonly due to the selective vulnerability of transplanted DANs and contaminating cell types.
However, a cause typically overlooked is microenvironment or non-cell-autonomous signaling: the
disease-causing agent might not be only within dopaminergic neurons, but also be lurking in nearby
cells and extracellular matrix.
Figure 2. Two Hypotheses of PD emergence and cure. To present two hypotheses of what causes the
low DAN survival, an analogy of the disease environment as an ailing village is displayed. First, the
cause might be “famine”, then new immigrants will not be affected by indigenous villagers. Second, if
the issue is “plague”, the immigrants will have to bring medicines. If PD derived from purely DAN
dysfunction, implanting more DANs can be a remedy; if not, adjustments within the microenvironment
using drugs have to be made.
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
260
Our main outstanding question is: what limits the survival rate of DANs? We hypothesized that there
is a defect in the microenvironment, rendering the strategy of only replenishing DANs insufficient
(Figure 2). To address this hypothesis, there are two questions to be answered: first, whether a non-
neuronal mechanism exists, affecting the transplanted DANs, and second, what molecules can
compensate for such effects and boost the survival of DAN transplant. Thus, a co-transplanted drug
should be added to alleviate the defects in the microenvironment. Here, we utilized multiple scRNA-seq
datasets of PD and with best suited machine learning based tools, we revealed critical genes through
comparison between initial differentiation and long-term transplantation data, thus proposing drugs as a
co-transplanted material to address the challenge in cell transplantation by targeting the
microenvironment dysfunction.
2. Methods
2.1. Dataset Overview
In order to address our hypothesis, we employed the comparison between disease versus healthy datasets
to confirm the existence of non-neuronal signaling in the PD microenvironment. We then utilized data
comparison between differentiation versus transplantation datasets to observe the differences before and
after transplantation, and further compared it with the natural development dataset to reveal specific
genes that could be causing this dysfunction. We also integrated an additional midbrain organoid
scRNA-seq datasets for further comparative analysis.
2.1.1. Disease versus Healthy Datasets. We employed the open source single cell RNA sequencing
(scRNA-seq) dataset from Khan et al., 2021 (GSE187012), which utilized a combined method of
neurotoxin maneb and paraquat to model PD mice. A healthy control group (GSM5667021), without
the injection of neurotoxin, and an PD disease modeled experiment group (GSM5667021) were both
investigated in this study. Both groups record the gene expression levels in substantia nigra pars
compacta at single cell resolution. We employed these two datasets to analyze the gene expression
differences in non-neuronal cells and for pathway examination.
2.1.2. Differentiation versus Transplantation Datasets. Datasets from Tiklová et al. (2020)[6] were
used to compare gene expression between the course of DAN differentiation and transplantation. PD
rats, modeled with neurotoxin 6-hydroxydopamine (6-OHDA), were transplanted with fetal tissues and
human embryonic stem cells (hESCs) derived neuronal grafts in ventral midbrain. The differentiation
dataset (GSE132758), recording the differentiation stage of early transplantation, and the transplanted
dataset (GSE118412), measuring the expression in transplanted graft after long term survival, gave
insights into how transplanted DANs were affected by non-cell-autonomous mechanisms.
2.1.3. Analysis of Developing Brain Gene Expression By Braindex. We implemented a developmental
database to compare and consolidate our results. Braindex, a brain development expression portal
created by Kleinman Lab, profiles gene expression of mouse brain in conditions of cancer and neuro-
development at single cell resolution [7]. We employed this database to obtain the gene expression
across developmental stages to examine which of our target genes revealed by scRNA were more
significant, by comparing its trend during healthy development, and during the transplantation process
of growth.
2.2. Single-cell RNA Sequencing Analysis
2.2.1. Pre-processing and Cell Type Identification. Our analysis is initiated via the R package Seurat
V5, tailored specifically for single-cell RNA-sequencing (scRNA-seq) data investigation. To filter out
low-quality cells and increase the accuracy of downstream analysis, we began by performing quality
control, removing cells with unique feature counts fewer than 500, and mitochondrial counts more than
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
261
5%. The data from different rats in the same database are then integrated, and normalized using
“LogNormalize” to stabilize variance. After normalization, 2000 highly variable features were identified
using “FindVariableFeatures”. These variable features were scaled using “ScaleData” to prevent the
highly expressed genes from overshadowing others. Next, dimension reduction was applied using
Principal Component Analysis (PCA), capturing the source of variability in the dataset. Finally, we
clustered the cells with “FindNeighbors” and “FindClusters”, categorizing cells by gene expression and
displaying visually using Uniform Manifold Approximation and Projection (UMAP). These cell clusters
were further analyzed using “FindMarkers” to reveal marker genes for the purpose of identifying their
specific cell types. These cell type identifications were reinforced with violin plots of corresponding
marker genes. This same procedure was implemented on all five datasets before their individual analysis.
2.2.2. Trajectory and Pseudotime Analysis. The machine learning package Monocle3 was used on top
of Seurat to perform trajectory and pseudotime analysis for a more comprehensive comparison between
differentiation and transplantation data. We transferred the data accessed by Seurat to a cds object in
Monocle3, so that pre-processing and cell type identification don’t need to be reiterated. The same
clusters were visualized in Monocle3 for double check, then the trajectory was learnt by using the
machine learning function “learn_graph”, displayed on the same UMAP. Furthermore, the cell types
were ordered in pseudotime by “order_cells”, visualized in color from yellow to purple, indicating the
relative time of differentiation progression. Ultimately, box plots of pseudotime of cell types were
plotted with ggplot2, an R visualization package.
2.2.3. Gene Set and Pathway Analysis. Via the R package clusterProfiler, we performed gene set
enrichment analysis to interpret the importance and functions of specific pathways. We created a gene
list using markers from the seurat object along with the marker gene expression and Log2 Fold Change.
Then we employed functions “enrichGO” and “enrichKEGG” to carry out functional enrichment
analysis. Furthermore, we executed gene set enrichment analysis using “gseGO”, obtaining ridgeplots
and dotplots to analyze the importance of each pathway.
2.3. Drug Mining Tools
We target significant genes revealed by scRNA-seq and the developmental database with the most
advanced tool currently available, the machine learning based tool AlphaFold 3 for protein structure
visualization and SwissDock for molecular docking. We also compared the protein structure in the
Protein Data Bank (PDB) and predicted by AlphaFold 3 during drug docking.
2.3.1. AlphaFold 3. As a deep learning based model for predicting protein structure, AlphaFold3
performs in high accuracy with nearly every protein in Protein Data Bank (PDB). AlphaFold 3 utilizes
pair-wise representation of chemical complexes and generates their atomic arrangements, allowing it to
perform highly in predicting protein structures and interactions.
2.3.2. SwissDock. In addition, molecular docking was done to explore how effectively the drug targets
the gene. To achieve the goal, I utilized SwissDock, a docking program using EADock ESS engine to
analyze how a ligand and a protein forms a stable compound [8]. Generating and ranking the binding
models with solvent effects and CHARMM energies, the algorithm enables SwissDock to perform
accurate docking within minutes.
3. Results
3.1. Verifying Non-neuronal Mechanisms in PD
According to our hypothesis, the first question is whether only refilling DANs is sufficient to relieve the
defects in brain regions affected by PD. Are the defects only purely DAN-derived or are they also caused
by other non-neuronal glial cells and extracellular matrix? To approach this question, we explored the
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
262
non-cell autonomous pathogenic mechanisms of PD. In order to verify, we found datasets of PD disease-
modeled condition and healthy conditions. By comparing the gene expression in microglia,
oligodendrocytes, and astrocytes instead of dopaminergic neurons via scRNA-seq, we revealed a large
disparity in the microenvironments in disease versus healthy state, thus suggesting that non-cell-
autonomous signaling certainly contributed to the disease. The PD disease dataset (GSE187012) [9],
which contained a healthy control group, and a disease modeled experimental group treated with maneb
and paraquat, was analyzed to uncover nine different cell types each, visualized by UMAP reduction,
each cluster representing a unique cell type or subtype (Figure 3). These cell types were identified using
their associated genes. Notably, not only two cells related to dopamine production, dopaminergic
neurons (DANs) and dopaminergic progenitor cells (DPs), were unveiled, but also, two subtypes of
microglias were discovered in both groups: HM (homeostatic microglia) and DAM (disease-associated
microglia). Specifically, DAM, a type of microglia related to neurodegenerative diseases may suggest
abnormality within the microenvironment. Moreover, neuroblastoma cells, a type of malignant cancer
cell, were also detected.
Figure 3. Cluster Diagram with UMAP reduction in Healthy and Disease Data.
Remarkably, we found disease related differences between the microenvironments of disease and
healthy conditions by analyzing the expression of PD associated genes in non-neuronal cells (Figure 4).
FBXO7, which is typically in deficiency in PD models causing impaired mitochondrial function, is also
expressed in low levels in oligodendrocytes of disease-modeled dataset, but high in healthy dataset.
SYNJ1, found to be reduced in MPTP-induced Parkinson mice, is also expressed in low levels in DAMs
of disease-modeled dataset compared to the healthy dataset. SYT11, on the other hand, reported to be
accumulated in PD mice, is found in high levels in astrocytes of disease-modeled dataset while low in
healthy dataset.
Figure 4. Comparing Disease Associated Genes in Microenvironment
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
263
In the end, the differences in expression levels of disease associated genes in non-neuronal cells
suggests differences occurring within the microenvironment. This could contribute to microenvironment
dysfunction that leads to degenerative death of healthy dopaminergic neurons. Therefore, we arrive at
the finding that it may not be sufficient to just restock the DAN for PD treatment. This microenvironment
or non-cell autonomous role in PD has been reported previously on glial cells, supporting our results.
Microglia, especially DAMs, for example, have been found to play damaging roles in neurodegenerative
diseases, and are especially pivotal in exacerbating neuroinflammation and autophagy impairment [10].
3.2. Comparing Differentiation and Transplantation Data of PD
As demonstrated in 3.1, PD is indeed associated with non-neuronal cells and microenvironment
abnormality. This raised the second question: would this malfunctioning microenvironment cause the
healthy DANs transplanted to be affected? To further investigate, we compared the differentiation and
transplantation conditions in a 6-hydroxydopamine PD model transplanted with ventral midbrain grafts
to examine the change in gene expression in DANs during and after transplantation. We utilized two
datasets, one differentiation dataset (GSE118412) and one transplantation dataset (GSE132758) [6],
which were generated from the same lab to reduce variance. Using scRNA-seq, the specific cell types
of each dataset are revealed before identification of specific disease-causing pathways.
3.2.1. Identifying Cell Types in Differentiation and Transplantation Data. In the differentiation dataset,
seven distinct cell types were disclosed using their corresponding cell marker genes, plotted in a UMAP
cluster plot (Figure 5). Similarly, eight cell types in the transplantation data were revealed. Noticeably,
two types of cell related to dopamine production, dopaminergic neurons and their progenitor cells, were
uncovered using the two marker genes: TH, crucial to synthesis of dopamine, and NR4A2 (also known
as Nurr1), essential for development and maintenance of dopaminergic neurons. After further analysis
with mature neuronal markers and markers of progenitor cells, DANs and DPs could be distinguished.
Figure 5. Differentiation VS Transplantation Cell Clusters with UMAP visualization. Left:
Differentiation; Right: Transplantation. The identified cell types are exhibited using the dimension
reduction technique of Uniform Manifold Approximation and Projection (UMAP) in a cluster diagram,
with each cluster and color representing a unique cell type
3.2.2. Trajectory Analysis of Differentiation and Transplanted DANs. To further compare differences
within differentiation and transplantation dataset, we conducted trajectory analysis, visualized by
UMAP (Figure 6), and plotted their relative pseudotime with a box plot (Figure 7). We found a normal
progression in the differentiation dataset, with DAN emerging after DP at the latest stage. Surprisingly,
we discovered abnormal progression in transplantation data: there are other cell types between DAN
and DP. This might indicate defects within DANs after long-term transplantation, potentially caused by
microenvironment dysfunction or non-cell-autonomous mechanisms.
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
264
Figure 6. Trajectory Analysis of Cell Types. Left: Differentiation; Right: Transplantation. The color
indicates the progression of cell types, with purple the earliest, yellow the latest.
Figure 7. Box Plot of Pseudotime Data. Left: Differentiation; Right: Transplantation.
3.2.3. Gene Set Enrichment Analysis. In order to further confirm our hypothesis that the abnormality
lurks in the transplantation microenvironment, we compared pathways by using gene set enrichment
analysis of the differentiation and transplantation data. Using a ridgeplot showing the distribution of
enrichment across gene sets, we surmise that dysfunctions or abnormalities may underlie within the
transplantation microenvironment (Figure 8). While the differentiation dataset contains gene sets with
variable enrichment distribution, the gene sets in the transplantation dataset show similar enrichment
distribution. This difference may be attributed to microenvironment dysfunction or other defects in non-
neuronal mechanisms. Moreover, surprisingly, by using a dotplot of activated and suppressed gene sets,
we discovered that although both datasets show traits of activated differentiation pathways, the
transplantation dataset shows a higher presence in mitotic transition, metaphase and anaphase, and
nuclear division that parallels cell proliferation (Figure 9). This might indicate gene dysregulation and
other issues, such as cancers.
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
265
Figure 8. Ridgeplot of Gene Set Enrichment.
Figure 9. Dotplot of Activated and Suppressed Gene Sets.
3.2.4. Gene Expression Differences in Differentiation and Transplantation. We compared the
expression of PD related genes in differentiation and transplantation dataset to reveal differences that
could be caused by non-neuronal mechanisms. Fifteen genes (PINK1, PRKN, GBA, PARK7, MTF2,
PPP6R2, ATP13A2, VPS35, MAPT, SYT11, SNCA, ADD1, IRS2, USP8 and USP25) were chosen
from GWAS [11]. By comparing these disease related genes in the two datasets, inconsistency was
found with genes PINK1, PARK7, PRKN, GBA, MTF2 and PPP6R2 (Figure. 10), while the other nine
genes exhibit little differences. While PINK1 and PARK7 are highly expressed in differentiation and
lowly in transplantation, the other four genes display the inverse.
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
266
Figure 10. Expression levels of Six Significant Genes in Differentiation and Transplantation
Within the six genes found to show differences, deficiency of PRKN and GBA were typically found
in PD, as PRKN deficiency harms mitophagy pathways and GBA insufficiency could damage the
autophagy-lysosome system [12,13]. PINK1 and PARK7 on the other hand are more complicated: their
mutations were not directly associated with higher or lower expression levels. Mutations in PINK1 are
associated with disrupted mitophagy, while mutations in PARK7 could trigger the gene to be oxidized,
which is a biomarker of PD, and cause oxidative stress [14]. Dysregulation in PINK1, Parkin and
PARK7 are all associated with Early Onset Parkinson’s Disease (EOPD) [15]. MTF2 and PPP6R2 are
novel genes, and it is not known whether it was upregulated or downregulated in PD, nor is it known
about their specific roles in the disease. The drastic differences shown in these six genes between
differentiation and transplantation data indeed confirms that transplanted cells were affected, or even
potentially damaged by the impairments in non-neuronal signaling.
3.3. Characterizing Critical Genes and Pathway
3.3.1. Filtering the Key Genes underlying abnormality in Transplantation. As shown in 3.2, there are
critical differences between transplantation and differentiation, and there are abnormalities in the
transplantation data. This result suggests that the healthy DANs transplanted may be influenced by the
dysfunctional signals in the microenvironment. But what are the most crucial genes causing this
influence? Among the six genes found, we compared the trend of expression in the differentiation and
transplantation data to the trend of development in the natural growth phase. In the end, the trend in
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
267
differentiation and transplanted versus general healthy developmental trend in the four genes, PRKN,
GBA, PARK7 and PPP6R2, did not display significant differences, suggesting that the gene expressions
were not severely affected by microenvironment and non-neural dysfunction. However, differences in
general trend was observed in PINK1 and MTF2. The expression levels of PINK1 decreases from
differentiation to transplantation, which is in contrast with the developmental trend in the healthy
environment that shows increase in expression along with growth. Similarly, the expression of MTF2
increases in differentiation to transplantation but decreases in general healthy growth trend (Figure 11).
Figure 11. Gene Expression Trend of PINK1, MTF2, PARK7, PRKN, GBA and PPP6R2 in Brain
Development visualized by Braindex [7]. The x-axis shows age, the y-axis shows the proportion of gene
positive cells, and colors indicate different cell types. The inconsistency of PINK1 and MTF2 between
differentiation to transplantation trend and general development trend reveals that defects might be
lurking within their pathways, and that they might be the main culprits in reducing the survival of
transplanted DANs. Thus, they are chosen for further analysis.
3.3.2. Gene Regulatory Pathway Analysis. As detailed in 3.3.1, PINK1 shows a decreasing trend during
transplantation and increasing trend in healthy development, whereas MTF2 has an inverse pattern.
Therefore, we hypothesized that there is a negative regulatory relationship between them which might
be caused by a linked pathway. The first question when tackling this hypothesis is whether a negative
correlationship exists. Through the use of a pseudotime plot, we discovered a strong opposite trend of
PINK1 and MTF2 expression in both healthy and disease microenvironments (Figure 12). However,
fluctuations in expression in disease microenvironment are drastically larger than they are in healthy
microenvironment, which may indicate that their pathways play important roles in PD.
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
268
Figure 12. Pseudotime Expression of MTF2 and PINK1.
To uncover their linked regulatory pathways, we utilized the String network to reveal associated
genes (Figure 13). We discovered the gene EED, which has a direct link to PINK1 while also interacting
with a myriad of genes linked to MTF2. EED is a key component of PRC2, and plays a prominent role
in growth, self-renewal and differentiation of stem cells; in the mammalian neural system, it could
regulate neuronal differentiation and neurogenesis [16]. Though the effects of EED on PD related genes
is unclear, it can give insights into a novel pathway of neurodegeneration.
Figure 13. Relationship between PINK1 and MTF2 with String network.
To further unearth what specific issues causes the dysregulation in PINK1, genes reported upstream
to PINK1 are analyzed. In healthy mitochondria, PINK1 are cleaved in the inner mitochondrial
membrane (IMM), a process facilitated by PARL, leading to PINK1 degradation. UBR1, a gene binding
to N terminals of PINK1, assists degradation. The comparison between differentiation and
transplantation dataset suggests that PARL and UBR1 both have notable changes during the course of
transplantation in DANs and other non-neuronal cells in the microenvironment like oligodendrocytes,
which may suggest potential disturbances (Figure 14).
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
269
Figure 14. Expression Level of PARL and UBR1 in Differentiation and Transplantation.
3.4. Drug Mining
As suggested in 3.3.2, PINK1 and MTF2 are negatively correlated. Between them, ultimately, PINK1
was chosen for drug screening (Figure 15), because MTF2 is a novel gene coding a transcription factor:
there is a lack of information to its novelty and moreover, transcription was often considered
“undruggable” due to its disordered essence.
Figure 15. Protein Structure of PINK1 Revealed by AlphaFold3. The colors dark blue, light blue, yellow
and orange indicate a decrease in confidence level respectively.
Previous results point to a drug activating PINK1, as PINK1 levels decreased during transplantation,
inconsistent with the increasing expression in natural healthy development. We found three confirmed
molecules that successfully activate PINK1: N6-Furfuryladenine, a neo-substrate kinetin [17]; MTK458,
a small molecule mitigating pUb and α-synuclein build-up [13]; and Niclosamide, an anthelmintic drug
that temporarily disrupts the mitochondrial membrane potential [18].
Using Swissdock, how firmly the connection formed by PINK1 and the drugs was revealed. While
N6-Furfuryladenine has the maximum affinity of -5.845 kcal/mol (Figure 16A) and Niclosamide with -
5.575 kcal/mol (Figure 16A), MTK458 with -8.256 kcal/mol (Figure 16A) is the drug docking the
closest. Dimethyl Sulfoxide, a versatile solvent used as a negative control, has an affinity of -1.900
kcal/mol (Figure 16D).
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
270
Figure 16. Molecular Docking of PINK1 from PDB with MTK458, N6-Furfuryladenine and
Niclosamide. Each panel shows the twenty locations of docking with the greatest affinity for each drug,
with a smaller calculated affinity number in kcal/mol signifying greater affinity. The dash line indicates
interactions, hydrogen bonds in blue, ionic interactions in yellow, cation pi interactions in orange, and
hydrophobic contacts in gray.
4. Discussion
4.1. Limitations
This study is novel and important by shedding light on a new co-transplantation strategy to enhance the
survival of dopaminergic neurons, but there are a few minor reservations. Firstly, for the purpose of
identifying associated genes, we compared differentiation versus transplantation dataset. However, to
reduce variance, we chose two datasets from the same lab, which could result in a caveat. If biases occur
within this lab, it could result in systematic errors. Despite these potential flaws, we were unable to find
a more faultless and comprehensive system. Thus, we utilized a natural developmental database to
compare as well to reduce the potential sequencing biases. In the future, as more differentiation and
transplantation data are gathered, these datasets can be integrated for multimodal analysis, increasing
reliability and revealing more drug targets for co-transplantation. Secondly, transplantation datasets used
in this study were obtained by mouse or rat models, but not on PD patients, since postmortem patient
samples for scRNA-seq are not available. If pathways are different in rats, mice and humans, then the
identified drugs might not effectively target the identified genes. Dealing with limited human datasets,
spatial transcriptomics using slide-seq can bridge the gap when translating from animals to humans.
4.2. Laboratory Validation
To test our conclusions, we propose a co-transplantation experiment, combining MTK458 with human
iPSC derived DANs to be transplanted in PD modeled mice. Long term survival rates of DANs with
drug co-transplantation could be compared with conditions with sole transplantation, while further
scRNA-seq analysis could be used to test the expression of PINK1 and associated genes in DANs and
cells within the microenvironment. However, using immortalized cell lines in this approach might not
be suitable, because they can not fully capture the features of DANs.
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
271
Another method to be employed is using the gene editing tool CRISPR (Clustered Regularly
Interspaced Short Palindrome Repeats) to filter the six genes identified in 3.2.3, and reveal which are
responsible for low survival rates of DANs during transplantation. Using CRISPR, the six genes could
be deactivated sequentially in cultured DANs, generating different cell lines, each with a distinct gene
knockout. Subsequently, by transplanting the edited DANs into mouse or rat models, if DANs with
specific gene disabled exhibit higher or lower survival rates than others, these genes can be indicated as
important. These significant genes could be further analyzed by targeting their pathways and associated
proteins using drugs, enhancing their activation or increasing their degradation, to observe if survival
rates are improved. Side effects of drugs such as inflammation could also be experimented, by necropsy
and tissue analysis [5].
4.3. In Vivo Reprogramming as Alternative Solution
Targeting non-neuronal mechanisms with PINK1 related pathways could potentially enhance the
survival rate of dopaminergic neurons not only in transplantation therapy, but also during in vivo
reprogramming. In vivo reprogramming is critical for rendering differentiated cells as dopaminergic
neurons in the patients’ midbrain, alleviating symptoms by remedying the degenerative death of DANs.
Astrocytes had been suggested to have the ability to be reprogrammed to DANs via a one-step
conversion of diminishing RNA-binding protein PTB [19,20], though disputes regarding the identity of
reprogrammed cell emerged [21]. This solution bypasses the need for transplantation, however, once
new DANs were produced by reprogramming, their survival may also depend on the health of its
microenvironment. By using drugs revealed in here, non-neuronal defects may be improved, thus
enhancing the efficacy and rate of success of in vivo reprogramming.
4.4. Future implications
This study demonstrates a range of implications for DAN transplantation therapy. First, PINK1
reduction can be caused by different genetic mutations or environmental factors. PINK1 has been
reported to be regulated by multiple pathways and molecules, including insulin and small interfering
RNAs (siRNA) [22,23]. Further identification of biomarkers through genetic profiling of individual
patients in personalized treatments could pinpoint the root agents causing dysregulation of PINK1 to
address their personal needs. This can be coupled with autologous iPSC derived DAN transplantation
to maximize personalization. PD is highly heterogeneous, driven by issues within multiple systems,
neurotransmitters and genes, thus genetic variability between individuals could affect treatments
significantly.
Secondly, PINK1 dysregulation found in this study could suggest how mitochondrial dysfunction
may regulate the survival of transplanted dopaminergic neurons. Defects in PINK1-parkin pathway is
the most common cause of Early Onset Parkinson’s Disease (EOPD), due to the disrupted mitophagy
leading to neuronal damages [15]. Prolonged mitochondria loss could cause insufficient ATP production,
cellular toxicity, accumulation of reactive oxygen species (ROS) and calcium current abnormality,
triggering neuronal loss [24].
Thirdly, organoids could also function as an alternative approach for transplantation therapy for PD
[25]. Organoids, three-dimensional cell culture systems often derived from stem cells, may be a potential
source other than grafts to restore DANs and alleviate PD symptoms. By adding supplements required
by DANs to support their growth and function, organoids could be a useful source of transplantation.
Treatments could also be further optimized by developing patient-derived organoids to personalized
medicine.
Moreover, immune modulators such as regulatory T cells could be combined with PINK1 during
transplantation to elevate the survival of transplanted dopaminergic neurons. The dysfunction of
mitophagy plays a pivotal role in triggering neuroinflammation, which could be especially harmful to
dopaminergic neurons in the transplanted graft. Regulatory T cells could thus be used as a suppressor to
neuroinflammation by mediating the response to needle trauma, alleviating neuronal and synaptic
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
272
damages [26]. Combining the MTK58 and regulatory T cells in co-transplantation could act as a dual
insurance to achieve better clinical outcomes.
4.5. Conclusion
This study has demonstrated a new strategy of enhancing DAN survival in transplantation to the
midbrain by modulating PINK1 pathway (Figure 17). We have first verified that non-neuronal defects
may be lurking in the microenvironment. We then revealed PINK1 and MTF2 inconsistency during
development versus during transplantation. Here we proposed MTK458 as a co-transplanted chemical
to activate PINK1 aiming to reduce the effect from non-neuronal signaling, bringing increased efficiency
and practicality to PD cell replacement therapy.
Figure 17. Proposed Working Model of MTK58 Enhancing Transplanted DAN Survival Rate by
Activating PINK1.
References
[1] Parkinson disease. (2023, August 9). Www.who.int. https://www.who.int/news-room/fact-
sheets/detail/parkinson-disease#:~:text=The%20prevalence%20of%20PD%20has%
20doubled%20in%20the
[2] Goldoni, R., Dolci, C., Boccalari, E., Inchingolo, F., Paghi, A., Strambini, L., Galimberti, D., &
Tartaglia, G. M. (2022). Salivary biomarkers of neurodegenerative and demyelinating diseases
and biosensors for their detection. Ageing Research Reviews, 76, 101587. https://doi.org/10.
1016/j.arr.2022.101587
[3] Sontheimer, H. (2021). Diseases of the Nervous System. Elsevier.
[4] Bloem, B. R., Okun, M. S., & Klein, C. (2021). Parkinson’s disease. The Lancet, 397(10291),
2284–2303. https://doi.org/10.1016/S0140-6736(21)00218-X
[5] Kim, J. J., Vitale, D., Otani, D. V., Lian, M. M., Heilbron, K., Iwaki, H., Lake, J., Solsberg, C.
W., Leonard, H., Makarious, M. B., Tan, E.-K., Singleton, A. B., Bandres-Ciga, S., Noyce, A.
J., Blauwendraat, C., Nalls, M. A., Foo, J. N., & Mata, I. (2024). Multi-ancestry genome-wide
association meta-analysis of Parkinson’s disease. Nature Genetics, 56(1), 27–36. https://doi.
org/10.1038/s41588-023-01584-8
[6] Tiklová, K., Nolbrant, S., Fiorenzano, A., Björklund, Å. K., Sharma, Y., Heuer, A., Gillberg, L.,
Hoban, D. B., Cardoso, T., Adler, A. F., Birtele, M., Lundén-Miguel, H., Volakakis, N.,
Kirkeby, A., Perlmann, T., & Parmar, M. (2020). Single cell transcriptomics identifies stem
cell-derived graft composition in a model of Parkinson’s disease. Nature Communications,
11(1), 2434. https://doi.org/10.1038/s41467-020-16225-5
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
273
[7] Jessa, S., Blanchet-Cohen, A., Krug, B., Vladoiu, M., Coutelier, M., Faury, D., Poreau, B., De
Jay, N., Hébert, S., Monlong, J., Farmer, W. T., Donovan, L. K., Hu, Y., McConechy, M. K.,
Cavalli, F. M. G., Mikael, L. G., Ellezam, B., Richer, M., Allaire, A., & Weil, A. G. (2019).
Stalled developmental programs at the root of pediatric brain tumors. Nature Genetics, 51(12),
1702–1713. https://doi.org/10.1038/s41588-019-0531-7
[8] Grosdidier, A., Zoete, V., & Michielin, O. (2011). SwissDock, a protein-small molecule docking
web service based on EADock DSS. Nucleic Acids Research, 39(suppl), W270–W277. https:
//doi.org/10.1093/nar/gkr366
[9] Khan, A. H., Lee, L. K., & Smith, D. J. (2022). Single-cell analysis of gene expression in the
substantia nigra pars compacta of a pesticide-induced mouse model of Parkinson’s disease.
Translational Neuroscience, 13(1), 255–269. https://doi.org/10.1515/tnsci-2022-0237
[10] Gao, C., Jiang, J., Tan, Y., & Chen, S. (2023). Microglia in neurodegenerative diseases:
mechanism and potential therapeutic targets. Signal Transduction and Targeted Therapy, 8(1),
1–37. https://doi.org/10.1038/s41392-023-01588-0
[11] Kim, T. W., Koo, S. Y., Riessland, M., Chaudhry, F., Kolisnyk, B., Cho, H. S., Russo, M. V.,
Saurat, N., Mehta, S., Garippa, R., Betel, D., & Studer, L. (2024). TNF-NF-κB-p53 axis
restricts in vivo survival of hPSC-derived dopamine neurons. Cell, 187(14), 3671-3689.e23.
https://doi.org/10.1016/j.cell.2024.05.030
[12] Chin, R. M., Rakhit, R., Ditsworth, D., Wang, C., Bartholomeus, J., Liu, S., Mody, A., Laishu, A.
, Eastes, A., Tai, C., Kim, R. Y., Li, J., Hansberry, S., Khasnavis, S., Rafalski, V., Herendeen,
D., Garda, V., Phung, J., de Roulet, D., & Ordureau, A. (2023). Pharmacological PINK1
activation ameliorates Pathology in Parkinson’s Disease models. https://doi.org/10.1101/
2023.02.14.528378
[13] Pradas, E., & Martinez-Vicente, M. (2023). The Consequences of GBA Deficiency in the
Autophagy–Lysosome System in Parkinson’s Disease Associated with GBA. Cells, 12(1),
191–191. https://doi.org/10.3390/cells12010191
[14] Gonçalves, F. B., & Morais, V. A. (2021). PINK1: A Bridge between Mitochondria and
Parkinson’s Disease. Life, 11(5), 371. https://doi.org/10.3390/life11050371
[15] Ma, K. Y., Fokkens, M. R., van Laar, T., & Verbeek, D. S. (2021). Systematic analysis of PINK1
variants of unknown significance shows intact mitophagy function for most variants. Npj
Parkinson’s Disease, 7(1). https://doi.org/10.1038/s41531-021-00258-8
[16] Schumacher, A., Faust, C., & Magnuson, T. (1996). Erratum: Positional cloning of a global
regulator of anterior–posterior patterning in mice. Nature, 384(6610), 648–648. https://doi.
org/10.1038/384648a0
[17] Osgerby, L., Lai, Y.-C., Thornton, P. J., Amalfitano, J., Le Duff, C. S., Jabeen, I., Kadri, H.,
Miccoli, A., Tucker, J. H. R., Muqit, M. M. K., & Mehellou, Y. (2017). Kinetin Riboside and
Its ProTides Activate the Parkinson’s Disease Associated PTEN-Induced Putative Kinase 1
(PINK1) Independent of Mitochondrial Depolarization. Journal of Medicinal Chemistry, 60(8),
3518–3524. https://doi.org/10.1021/acs.jmedchem.6b01897
[18] Barini, E., Miccoli, A., Tinarelli, F., Mulholland, K., Kadri, H., Khanim, F., Stojanovski, L., Read,
K. D., Burness, K., Blow, J. J., Mehellou, Y., & Muqit, M. M. K. (2018). The Anthelmintic
Drug Niclosamide and Its Analogues Activate the Parkinson’s Disease Associated Protein
Kinase PINK1. Chembiochem, 19(5), 425–429. https://doi.org/10.1002/cbic.201700500
[19] Qian, H., Kang, X., Hu, J., Zhang, D., Liang, Z., Meng, F., Zhang, X., Xue, Y., Maimon, R.,
Dowdy, S. F., Devaraj, N. K., Zhou, Z., Mobley, W. C., Cleveland, D. W., & Fu, X.-D. (2020).
Reversing a model of Parkinson’s disease with in situ converted nigral neurons. Nature,
582(7813), 550–556. https://doi.org/10.1038/s41586-020-2388-4
[20] Zhou, H., Su, J., Hu, X., Zhou, C., Li, H., Chen, Z., Xiao, Q., Wang, B., Wu, W., Sun, Y., Zhou,
Y., Tang, C., Liu, F., Wang, L., Feng, C., Liu, M., Li, S., Zhang, Y., Xu, H., & Yao, H. (2020).
Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease
in Mice. Cell, 181(3), 590-603.e16. https://doi.org/10.1016/j.cell.2020.03.024
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
274
[21] Wang, L.-L., Serrano, C., Zhong, X., Ma, S., Zou, Y., & Zhang, C.-L. (2021). Revisiting astrocyte
to neuron conversion with lineage tracing in vivo. Cell, 184(21), 5465-5481.e16. https://doi.
org/10.1016/j.cell.2021.09.005
[22] Hees, J. T., Wanderoy, S., Lindner, J., Helms, M., Murali Mahadevan, H., & Harbauer, A. B.
(2024). Insulin signalling regulates Pink1 mRNA localization via modulation of AMPK
activity to support PINK1 function in neurons. Nature Metabolism, 1–17. https://doi.org/10.
1038/s42255-024-01007-w
[23] Deng, H., Jankovic, J., Guo, Y., Xie, W., & Le, W. (2005). Small interfering RNA targeting the
PINK1 induces apoptosis in dopaminergic cells SH-SY5Y. Biochemical and Biophysical
Research Communications, 337(4), 1133–1138. https://doi.org/10.1016/j.bbrc.2005.09.178
[24] Norat, P., Soldozy, S., Sokolowski, J. D., Gorick, C. M., Kumar, J. S., Chae, Y., Yağmurlu, K.,
Prada, F., Walker, M., Levitt, M. R., Price, R. J., Tvrdik, P., & Kalani, M. Y. S. (2020).
Mitochondrial dysfunction in neurological disorders: Exploring mitochondrial transplantation.
Npj Regenerative Medicine, 5(1). https://doi.org/10.1038/s41536-020-00107-x
[25] Fu, C.-L., Jiang, X., Dong, B.-C., Li, D., She, X.-Y., & Yao, J. (2024). Protocol for transplantation
of cells derived from human midbrain organoids into a Parkinson’s disease mouse model to
restore motor function. STAR Protocols, 5(3), 103251–103251. https://doi.org/10.1016/j.xpro.
2024.103251
[26] Park, T.-Y., Jeon, J., Lee, N., Kim, J., Song, B., Kim, J.-H., Lee, S.-K., Liu, D., Cha, Y., Kim, M.
, Leblanc, P., Herrington, T. M., Carter, B. S., Schweitzer, J. S., & Kim, K.-S. (2023). Co-
transplantation of autologous Treg cells in a cell therapy for Parkinson’s disease. Nature,
619(7970), 606–615. https://doi.org/10.1038/s41586-023-06300-4
Proceedings of the 4th International Conference on Biological Engineering and Medical Science
DOI: 10.54254/2753-8818/64/2024.18003
275
