Parkinson’s disease (PD) is a neurodegenerative disease resulting from the degeneration of dopaminergic neurons in the substantia nigra pars compacta (SNpc). Motor symptoms include tremors, rigidity, and bradykinesia (1). Currently, levodopa therapy (2), which employs dopamine precursors that can cross the blood-brain barrier, is extensively used. Levodopa temporarily alleviates motor symptoms in patients with PD. However, persistent administration of high doses of levodopa may result in levodopa-induced dyskinesia (3) and hasten disease progression. Therefore, cell transplantation is currently considered to be the most promising treatment for PD (4-6). In brain regions lacking dopaminergic neurons, neural stem cells (NSCs) or neural progenitor cells (NPCs) that can differentiate into dopaminergic neurons are transplanted. Several studies (4-6) suggest that cell transplantation alleviates the motor symptoms of PD. However, cell transplantation for the treatment of PD is limited by adverse effects and high costs, making its therapeutic application difficult (7). To overcome this obstacle, we reprogrammed astrocytes directly into dopaminergic neurons to treat PD. Numerous studies have investigated the direct reprogramming of non-neuronal cells into induced neurons (iNs). A previous study demonstrated, for the first time, that the expression of three transcription factors, BRN2 (Pou3f2), achaete-scute family bHLH transcription factor 1 (ASCL1), and MYT1L, can transform mouse fibroblasts into functional iNs. Furthermore, NeuroD overexpression has been used to convert human fibroblasts into iNs, according to research (8).

Astrocytes (9) are non-neuronal cells prevalent in the brain that play a role in neurotransmitter transport. To exploit these properties, we induced the direct reprogramming of dopaminergic neurons (iDAN) by overexpressing transcription factors in astrocytes. ASCL1 is a basic helix-loop-helix transcription factor that activates gene expression by binding to the E-box motif (5’-CANNTG-3’). It is a primary factor that initiates and regulates the development of the central nervous system and promotes neurogenesis. According to another study, ASCL1 overexpression can directly reprogram fibroblasts (10) and astrocytes (11) into neurons. ASCL1 facilitates the direct reprogramming of fibroblasts chromatin remodeling into mature neurons by opening closed chromatin at the target site, causing nucleosome phasing (12). In a previous study conducted in our laboratory, we discovered that an ASCL1^5SA mutant, which involved artificial blocking of phosphorylation through five serine to alanine mutations, facilitated the direct reprogramming of mouse astrocytes into neurons (13). To determine whether the ASCL1^5SA mutant had a greater positive effect than ASCL1, a comparative study was conducted. When ASCL1, NURR1, LMX1A, SHH, and BclXL were overexpressed in the ventral midbrain of rodents, the astrocytes in this region were found to be directly reprogrammed into dopaminergic neurons (14). NURR1 and LMX1A are essential for the survival and maintenance of dopaminergic neurons in the midbrain (14). Overexpression of ASCL1, NURR1, LMX1A, SHH, and BclXL in rat cortical (CTX) and ventral midbrain (VM) astrocytes resulted in their direct reprogramming to dopaminergic neurons. In addition, the ASCL1 factor alone enabled rat VM astrocytes to transform directly into dopaminergic neurons. Following ASCL1 overexpression in VM astrocytes, numerous dopaminergic neurons (TH+/TUJ1+ cells) have been identified. In addition, the number of dopaminergic neurons increased during the 3-week differentiation period. However, these outcomes have not been observed in CTX astrocytes. In this study, the use of ASCL1 for direct reprogramming of dopaminergic neurons produced diverse results in different brain regions.

Overexpression of ASCL1, NURR1, LMX1A, SHH, and BclXL combinations in CTX and VM astrocytes enables iDAN direct reprogramming

Upon overexpression of ASCL1 + NURR1 + LMX1A (ANL), ASCL1^5SA + NURR1 + LMX1A (5SANL), NURR1 + SHH + BclXL + ASCL1 (NHB + A), or NURR1 + SHH + BclXL + ASCL1^5SA (NHB + 5SA) in CTX astrocytes, direct reprogramming into dopaminergic neurons with VM characteristics was observed (14). Supplementary Fig. 1 depicts the characteristics of CTX and VM astrocytes. To confirm the reproducibility of these results in rat CTX and VM astrocytes, retroviruses were used for overexpressing ANL, 5SANL, NHB + A, or NHB + 5SA (Fig. 1A). On the third day of direct reprogramming, TUJ1+ cells were found in all CTX and VM astrocyte groups. On day 5, TUJ1+ cells more than doubled in all groups except in the control group, and TH+ cells began to appear (Fig. 1B, D). ANL overexpression in CTX and VM astrocytes showed the greatest increase in TH+/TUJ1+ cells on day 5 and day 7 of direct reprogramming, respectively. Subsequently, among both CTX and VM astrocytes, the number of TH+/TUJ1+ cells decreased as the direct reprogramming period elapsed (Fig. 1C, E). These results are consistent with those of previous studies (14). The combination of ANL, 5SANL, NHB + A, or NHB + 5SA factors effectively enabled direct reprogramming of dopaminergic neurons in mouse (data not shown) and rat astrocytes.

Direct reprogramming of rat VM astrocytes into iDANs using only the ASCL1 factor

Our objective was to identify key transcription factors that facilitate the direct reprogramming of astrocytes into dopaminergic neurons. We examined whether overexpression of ASCL1 in CTX and VM astrocytes can directly reprogram iDANs by inducing TH+ cells. However, TH+ cells were not observed among CTX astrocytes that did not express NURR1 or LMX1A (Fig. 2A). No significant differences were observed during reprogramming. Therefore, NURR1 and LMX1A were determined to be essential factors for the direct reprogramming of CTX astrocytes to iDANs. In contrast, TH+/TUJ1+ cells appeared among VM astrocytes expressing NURR1 and LMX1A, as well as in the ASCL1 and ASCL^5SA mutant single-factor groups (Fig. 2A). The number of TH+/TUJ1+ cells increased among VM astrocytes during long-term differentiation. Interestingly, ANL-overexpressing VM astrocytes displayed the highest number of TH+/TUJ1+ cells on day 7 of differentiation, with a slight decrease observed on day 14. In contrast, the VM astrocytes overexpressing the ASCL1 single factor showed no significant changes in TH+/TUJ1+ cells on day 7 of differentiation compared with the ANL astrocytes. However, on day 14 of differentiation, the number of TH+/TUJ1+ cells among the VM astrocytes overexpressing ASCL1 was higher than that among the ANL astrocytes. This result was consistent with that of the quantitative analysis (Fig. 2C). These results suggest that, in VM astrocytes, NURR1 and LMX1A play a role in rapid reprogramming to iDANs at the beginning of differentiation, whereas ASCL1 induces stepwise iDAN reprogramming according to the differentiation period.

In summary, we found that direct reprogramming from CTX astrocytes to iDANs is not possible with ASCL1 alone without the midbrain transcription factors NURR1 and LMX1A. VM astrocytes can be reprogrammed into iDANs by overexpression of ASCL1 or ASCL1^5SA as single factors. Direct reprogramming of CTX and VM astrocytes using NURR1 and LMX1A without ASCL1 did not have a significant effect (Fig. 2B). Therefore, ASCL1 is the most important transcription factor for direct iDAN reprogramming.

ASCL1 shows prolonged-expression and can induce direct reprogramming into iDANs

To confirm whether ASCL1 expression persisted during direct reprogramming, iDANs expressing ASCL1 and non-reprogrammed astrocytes expressing ASCL1 were harvested. GFP protein was expressed at 27 kDa and ASCL1 fused to FLAG protein was expressed at 31 kDa in non-reprogrammed cells. After 4 weeks of iDAN reprogramming, the expression was similar to that in non-reprogrammed cells. ASCL1 expression was maintained after the direct reprogramming of astrocytes to iDANs (Fig. 3A). Thus, ASCL1 expression persisted for more than 4 weeks. We aimed to demonstrate that ASCL1 expression persists and induces direct reprogramming of iDAN.

Subsequently, we measured the mRNA expression levels at different stages of reprogramming. ASCL1 was overexpressed in CTX and VM astrocytes to induce direct reprogramming of iDANs, and the mRNA level was evaluated using reverse transcriptase-polymerase chain reaction (RT-PCR). mRNA was extracted from cells that differentiated after 2 and 4 weeks. Consistent with the immunocytochemistry (ICC) results, Tuj1 and TH mRNAs were expressed only in VM astrocytes overexpressing ASCL1. These expression levels showed no significant difference with respect to the iDAN differentiation period (Fig. 3B, C). Similar results were obtaineded using real-time quantitative reverse transcription PCR (RT-qPCR), which showed that the Tuj1 and TH mRNA levels were high in the VM astrocytes that overexpressed ASCL1. Tuj1 mRNA levels in this group were upregulated by approximately 18-fold or higher compared to those in the VM control, and TH mRNA expression was upregulated by approximately 720-fold or higher (Fig. 3D).

Furthermore, we confirmed that the dopaminergic neurons reprogrammed in astrocytes are functional. ELISA was conducted to verify dopamine release, which is an indicator of functional dopaminergic neurons. Supernatants containing dopamine were harvested after 48 h of cell incubation or 30 min of stimulation with 56 mM KCl. ASCL1-induced iDANs from VM astrocytes showed the highest dopamine release. Indeed, at 21 days after transduction, the VM iDANs released high levels of dopamine with and without 56 mM KCl (Fig. 3E).

3D culture is more effective for iDAN direct reprogramming

Previous research has demonstrated that passing through the neurosphere state helps maintain the intrinsic characteristics of NPCs (15, 16). ASCL1 was overexpressed in rat VM astrocytes, which were subsequently subjected to direct reprogramming to NPCs in the neurosphere state under a 3D culture environment in AggreWell^TM 800. Subsequently, these cells transdifferentiated into iDANs. Additionally, we attempted to confirm the possibility of forming neurospheres that mimic the midbrain environment. Experiments were conducted in which microglia, an essential brain cell type, were added to astrocytes at a ratio of 10-20%. After ASCL1 was overexpressed in rat VM astrocytes, the cells were seeded at a density of 1 × 10⁶ cells per well in AggreWell^TM 800. Neurospheres were cultured for a week in induced NPC reprogramming (iNPC) conversion medium, after which the medium was replaced with iDAN conversion medium (Fig. 4A, B). After 28 days of differentiation, the direct reprogramming of CTX and VM astrocytes into iDANs was induced. TH+ cells were not observed in the CTX astrocytes, consistent with previous findings. In contrast, numerous TH+ cells were detected in VM astrocytes, indicating a higher efficiency of direct reprogramming into iDAN than that in 2D cultures (Fig. 4C). Subsequently, we investigated the effect of microglia on the direct reprogramming of iDANs and observed that microglia had no significant effect on the generation of FOXA2+/TH+ dopaminergic neurons (Fig. 4D, E). These results suggest that 3D culture has a beneficial effect on the direct reprogramming of dopaminergic neurons but does not contribute to the generation of dopaminergic neurons that exhibit VM characteristics.

We demonstrated that rat VM astrocytes can be directly reprogrammed into iDANs by overexpressing ASCL1, whereas CTX astrocytes cannot. This implies that the results of direct reprogramming are dependent on region-specific characteristics of astrocytes. Therefore, it is important to consider the unique features of a particular astrocyte region when conducting direct reprogramming experiments. Differences in gene expression between cortical and ventral midbrain astrocytes have been reported (17, 18).

In this study, we used only astrocytes that were passaged once. Approximately 70% of astrocytes were maintained as GFAP+ cells, whereas the remaining 30% consisted of other cell types (Supplementary Fig. 1). Therefore, further verification is required to confirm whether direct reprogramming occurs in other cells.

A significant disparity in retroviral transduction efficiency was observed between CTX and VM astrocytes (Supplementary Fig. 2). Retroviruses are transduced during cell division, and the higher infectivity of VM astrocytes could be attributed to their relatively rapid growth rate. Therefore, it was crucial to confirm whether the observed differences in direct reprogramming resulted from ASCL1 overexpression or from differences in transduction efficiency (Figs. 1 and 2).

The expression of FOXA2, a midbrain-specific transcription factor, did not merge with that of TH-positive neurons that were directly reprogrammed by ASCL1 overexpression in VM astrocytes (Supplementary Fig. 3). This suggests that iDANs from VM astrocytes do not function as midbrain dopaminergic neurons. Therefore, to induce dopaminergic neurons with midbrain characteristics, VM astrocytes were 3D cultured to obtain the shape and characteristics of VM organoids. Brain organoids generated in vitro can be used for studying specific parts of the brain, as well as brain development and disorders. Human-derived iPSCs (19, 20) and human embryonic stem cells (hESCs) (21-23) are the most representative cell sources that initially emerged as brain organoids. NPCs differentiate into various cell types in the central nervous system to form small brain organoids, which are emerging as innovative treatments for degenerative brain diseases. In particular, the development of midbrain organoids is necessary for PD treatment. However, in the present study, when utilizing ASCL1 for 3D culture, dopaminergic neurons exhibiting midbrain characteristics were not observed. Because one gene was overexpressed using the retrovirus for each factor, it is possible that not all factors completely transduced the cells. Other limitations of previous attempts to develop organoids with midbrain dopaminergic neurons include the lack of small molecules that are commonly used in the production of VM organoids and the limited culture period (Fig. 4).

Specific expression of ASCL1 in rat brain astrocytes to induce iDANs was difficult to determine (Supplementary Fig. 4). To resolve these problems, a retrovirus that operates under an astrocyte-specific promoter needs to be injected. However, negative results have been published for astrocyte-specific promoters, such as the GFAP or ALDH1L1 promoter. Previous studies have verified that the GFAP promoter functions nonspecifically. The GFAP promoter has been reported to leak into endogenous neurons in vivo (24). Additional studies are required to overcome these limitations.

In conclusion, this study indicates the following: 1) Direct reprogramming of dopaminergic neurons is possible when ASCL1 is overexpressed in rat VM astrocytes. 2) These results were not observed in CTX astrocytes. 3) Direct reprogramming of ASCL1 into dopaminergic neurons has yielded different results in distinct brain regions. Based on these findings, direct reprogramming of non-neuronal cells into neurons presents innovative possibilities for cell regeneration and the treatment of degenerative brain diseases for which there is no definite cure.

Dopaminergic neuron differentiation

To generate iDANs from CTX and VM astrocytes, 2 × 10⁴ cells were seeded per well in a 24-well plate containing a 12 mm slide (Bellco, USA) coated with PLO/FN. The astrocyte expansion medium was utilized to stabilize the cells at 37°C and 5% CO₂ for 24 h. Retrovirus was mixed with astrocyte expansion medium, and 8 μg/ml polybrene was added. The cells were treated for 12 h, and subsequently, the treatment medium was replaced with the astrocyte expansion medium. After incubation for 24 h, this medium was replaced with DMEM:F12 (1 ml per well) containing 2% B-27, 1% glutaMax-I, 10 ng/ml bFGF, 10 μM Forskolin (FSK, Sigma-Aldrich), and 1% penicillin/streptomycin (P/S). After 4 days, 20 ng/ml brain-derived neurotrophic factor (BDNF; PeproTech, USA) was added. Three days later, the medium was replaced with N2 medium containing 2% B-27, 0.2 mM ascorbic acid (AA; Sigma-Aldrich), 250 μg/ml dibutyl cyclic-AMP (cAMP; Sigma-Aldrich), 20 ng/ml BDNF, and 20 ng/ml glial cell line-derived factor (GDNF; PeproTech) (iDAN conversion medium). The iDAN conversion medium was replaced every other day.

Induced neuronal progenitor cell reprogramming (iNPC)

iNPCs were generated from CTX and VM astrocytes using iNPC medium, DMEM: F12 containing 2% B-27, 1% glutaMax-I, 20 ng/ml bFGF, 10 μM Forskolin, and 1% P/S.

ELISA for dopamine release

A dopamine release assay was performed using the Dopamine Research ELISA Kit (Labor Diagnostika Nord, Germany) according to the manufacturer’s instructions. After 21 days of transduction, dopamine-released supernatants were collected following a 48 h incubation period (basal release) or 30 min of stimulation with 56 mM KCl (evoked).

Cell counting and statistical analysis

Cell counting was randomly performed in 10-15 microscopic fields/well and 3 wells/experimental conditions. Each experiment was independently performed at least three times. All data are expressed as mean ± standard error (S.E.). Using SigmaPlot for Windows version 10.0 (SystatSoftware GmbH, Erkrath, Germany), paired t-tests were performed for statistical comparison of data.

This research was supported by grants from the Basic Science Research Program (NRF-2019R1A2C2005681), the K-Brain Project of the National Research Foundation funded by Ministry of Science and ICT (RS-2023-00266171), and the Korea Dementia Research Project through the Korea Dementia Research Center funded by Ministry of Health and Welfare and MSIT (HU22C0143).

The authors have no conflicting interests.