To investigate the potential involvement of MLKL in Parkinson's disease (PD) stress conditions, we performed experiments to evaluate the effect of MLKL inhibition on the cytotoxicity of 6-hydroxydopamine (6-OHDA), a neurotoxin commonly used to model PD, in human neuroblastoma cell line SH-SY5Y, and primary mouse embryonic fibroblasts (MEFs) (Fig. 1a-1d). SH-SY5Y cell line is frequently chosen in current PD research and primary MEFs are sensitive to phospho-MLKL triggered necroptotic cell death 24,25. Also, an MLKL inhibitor, necrosulfonamide (NSA), was employed with 6-OHDA plus TNF-α-induced cell death. We observed that TNF-α enhanced the sensitivity of both cell types to necrotic cell death in the presence of 6-OHDA (Fig. 1a and 1c). However, the addition of NSA significantly impaired this effect in a dose-dependent manner (Fig. 1b and 1d). Moreover, the phosphorylated MLKL (p-MLKL) and inducible nitric oxide synthase (iNOS), which are key markers of cell necroptosis, were highly expressed in 6-OHDA/TNF-α-treated cells; however, their expression was significantly reduced when treated with NSA (Fig. 1g and S1a-S1b).
As a high expression level of α-synuclein (α-Syn) has been reported to produce PD-like cellular and axonal pathologies in the nigrostriatal region 26,27, we also investigated the effect of NSA on cell death induced by α-Syn aggregates under PD stress conditions. We transfected primary MEF cells with human A53T α-Syn-GFP and observed that NSA exhibited neuroprotective effects against 6-OHDA plus TNF-α, accompanied by downregulated expressions of p-MLKL and iNOS (Fig. 1e and 1h). Additionally, in A53T α-Syn-GFP-transfected MEF cells, knocking out MLKL caused a significant decrease in 6-OHDA plus TNF-α-triggered cell death (Fig. 1f). Consistently, the expression of p-MLKL and iNOS were detectable in wild-type (WT) MEF cells, but not in Mlkl−/− cells, even with 6-OHDA/TNF-α treatment (Fig. 1i). These results suggest that MLKL-mediated necroptosis may be positively correlated with oxidative stress responses.
Further enzyme-linked immunosorbent assay (ELISA) analyses revealed that 6-OHDA and TNF-α co-stimulation triggered the robust secretion of proinflammatory cytokines IL-6, IP-10, MCP-1, CCL3, CCL5, CXCL1, CXCL2, and CXCL5 (Fig. 1j). In line with these observations, the expression levels of chemokines, particularly IL-6 and MCP-1, were much higher in Mlkl+/+ MEF cells after 6-OHDA/TNF-α treatment but were lower in 6-OHDA/TNF-α-treated Mlkl−/− MEF cells (Fig. 1j). These findings suggest that MLKL-mediated inflammatory signaling is highly associated with the 6-OHDA or α-Syn-induced PD model.
MLKL deficiency protects dopaminergic neurons and attenuates neuroinflammation in the A53T transgenic mice
To investigate whether MLKL deficiency improves motor capability by regulating α-Syn function, we conducted immunoblotting, immunofluorescence, and immunohistochemical staining for phosphorylated α-Syn at serine 129 (p-α-Syn129S), a specific pathological form associated with α-Syn aggregation in PD 29. Our results demonstrated that p-α-Syn129S was present in high abundance in the cortex, striatum, and substantia nigra regions of Tg-Mlkl+/+ mice; however, they markedly decreased in the corresponding regions of Tg-Mlkl−/− mice (Fig. 3a-3b). Additionally, fluorescence immunohistochemistry (IHC) confirmed that MLKL deficiency significantly reduced phosphorylated α-Syn inclusions in the striaitum region of the A53T transgenic mice (Fig. 3c). In humans with Parkinson's disease, the substantial loss of dopaminergic (DA) neurons in the SN is closely related to motor dysfunction 30. Although there was no significant difference in dopaminergic neurodegeneration in the striatum among the three mouse groups (Fig. 3d-3e), we observed a higher level of TH-positive neurons in the SN of Tg-Mlkl−/− mice and WT mice, but not in Tg-Mlkl+/+ mice (Fig. 3d-3e). Furthermore, the DA neurons exhibited different morphologies in the SN regions of Tg-Mlkl+/+ and Tg-Mlkl−/− mice. DA neurons in the pars compacta of the SN in Tg-Mlkl−/− mice exhibited a high density of TH-positive fibers and contained a denser cell mass than those in Tg-Mlkl+/+ mice (Fig. 3d-3e). Immunoblot analysis also demonstrated that TH accumulation was significantly elevated, accompanied by a remarkable reduction of α-Syn, p-α-Syn, and iNOS in the cortex, striatum, and SN of Tg-Mlkl−/− mice (Fig. 3f-3g). Hence, these data indicate that MLKL-mediated signaling is closely related to dopaminergic neurodegeneration and α-Syn aggregation in mice.
Activation of microglia, as evidenced by increased expression of Iba-1, can indirectly indicate a neuronal abnormality in A53T transgenic (Tg-Mlkl+/+) mice. To investigate this, we assessed the expression of Iba1 and found that Mlkl knockout significantly reduced Iba1 immunoreactivity in the cortex and substantia nigra (SN) of Tg-Mlkl+/+ mice (Fig. 4a-4b). Additionally, MLKL deficiency resulted in a significant decrease in the expression of CD11b, a surface receptor that is upregulated on activated microglia and is involved in the neuroinflammatory response in the brain 31, in the cortex and ventricle regions of Tg-Mlkl+/+ mice (Fig. 4e-4f).
The pathological marker, glial fibrillary acidic protein (GFAP), was prominently accumulated in the cortex and SN regions of Tg-Mlkl+/+ mice. In contrast, it was significantly decreased in the cortex and SN regions of Tg-Mlkl−/− mice (Fig. 4c-4d). Notably, microglia and astrocytes showed enlarged somas in Tg-Mlkl+/+ mice, indicating morphological activation; however, microglia and astrocytes shrank when Mlkl was knocked out (Fig. 4b and 4d). These findings indicate that MLKL deficiency significantly attenuated microglia and astrocyte activation, ameliorating Parkinson's symptoms in A53T transgenic (Tg-Mlkl+/+) mice.
Moreover, as MLKL-mediated necroptosis exacerbates multiple neurodegenerative diseases by triggering cell death and neuroinflammation 32, we evaluated the production of multiple serum cytokines using the ELISA method. Our results demonstrated that many proinflammatory cytokines, including IL6 and MCP-1, were significantly reduced in Tg-Mlkl−/− mice compared to Tg-Mlkl+/+ mice (Fig. 4g), which was consistent with the aforementioned results (Fig. 1j).
Single-cell RNA sequencing (scRNA-seq) analysis reveals the upregulated synaptic-related neurons and downregulated microglia in the SN region of the Tg- Mlkl −/− mice
To investigate the role of MLKL in advanced PD, we conducted scRNA-seq on nuclei isolated from substantia nigra regions of Tg-Mlkl−/− and Tg-Mlkl+/+ mice (Fig. S2). In the SN region of Tg-Mlkl+/+ mice (n = 3), we generated 3,563 single nuclei gene expression profiles, with a median of 405 genes and 100,993 transcripts per nucleus. For the SN region of Tg-Mlkl−/− mice (n = 3), we generated 8,466 single nuclei gene expression profiles, with a median of 497 genes and 43,481 transcripts per nucleus (Fig. S3a-S3b). In addition, we utilized Uniform Manifold Approximation and Projection (UMAP) visualization to separate nuclei into distinct clusters (Fig. 5a). Next, we annotated these clusters using cell-type-specific markers to identify oligodendrocytes (e.g., Ptgds, Gm16233, Anln, Ndrg1, and Gng11), oligodendrocyte precursor cells (e.g., Vcan, Cspg5, Thr, Neu4, and Pdgfra), astrocytes (e.g., Atp1a2, Gm3764, Slc4a4, Slc1a2, and Rorb), neurons (e.g., Meg3, Snhg11, Ahi1, Ube3a, and Syt1), Bergman glial cells (e.g., Atp13a5, Pdgfrb, Kcnj8, Igfbp7, and Vtn), type II spiral ganglion neurons (e.g., H2-D1, H2-K1, Kif2, Cd52, and Ly6c1), and microglia (e.g., C1qb, Arhgap45, C1qc, C1qa, and Ctss) (Fig. 5b).
We observed significant differences in the cluster sizes between the two groups, with a marked increase in the proportions of neurons and astrocytes in the Tg-Mlkl−/− mice, while oligodendrocytes and microglia were more frequent in the Tg-Mlkl+/+ mice (Fig. 5a and 5c). Additionally, we identified three clusters (neuron, microglia, and astrocyte) that exhibited both cell-type-specific and common gene expression patterns between the Tg-Mlkl−/− and Tg-Mlkl+/+ mice (Fig. 5d).
Differential expression of genes (DEGs) was identified in neuron, microglia, and astrocyte clusters between Tg-Mlkl−/− and Tg-Mlkl+/+ mice, followed by gene ontology (GO) term enrichment analysis of biological processes (Fig. 5e-5f and S4a-S4b). Notably, Tg-Mlkl−/− neuronal cells exhibited upregulation of nervous system processes (e.g., Efnb3) and downregulation of nitrogen compound metabolic processes (e.g., Tsix) (Fig. 5e-5f). Furthermore, downregulated genes in Tg-Mlkl−/− neuronal cells were enriched in functions related to cytokine production and apoptotic signaling pathways, indicating reduced inflammation and cell death (Fig. 5e). Similarly, microglia from Tg-Mlkl−/− mice showed upregulation of DEGs related to neurogenesis and downregulation of DEGs that are positive regulators of inflammation and secretion (Fig. S4a). Additionally, upregulated genes were enriched in neuron projection development and morphogenesis, while DEGs related to innate immune response were downregulated in Tg-Mlkl−/− astrocytes (Fig. S4b).
The differentially expressed genes (DEGs) within the neuron, microglia, and astrocyte clusters were combined to identify shared up-regulated and down-regulated genes. A heatmap displaying the clustering analysis of these 180 common DEGs is presented in Fig. 5e. Among them, the mt-Co1, mt-Co2, mt-Co3, mt-Cytb, mt-Atp6, Ptgds, AC149090.1, Tsix, Gm47283, and Xist genes exhibited the highest expression changes (Fig. 5e). Notably, five out of the ten DEGs (mt-Co1, mt-Co2, mt-Co3, mt-Cytb, and mt-Atp6) are critical components of the mitochondrial electron transport chain (ETC) and mitochondrial respiratory chain 33,34. Mitochondrial dysfunction has long been implicated in the pathogenesis of PD 35, and several studies have identified mutations in mt-Co2, mt-Co2, mt-Co3, and mt-Atp6 associated with various clinical brain disorders 36. Glutathione-independent prostaglandin D synthase (PTGDS), a prostaglandin involved in pain and sleep, was also identified as a unique blood-based signature capable of differentiating between idiopathic PD patients and controls 37. A recent study demonstrated that PTGDS was upregulated in PD patients and could serve as an optimal biomarker for PD diagnosis 37. In our case, quantitative real-time polymerase chain reaction (qRT-PCR) was conducted, confirming a 2.8-fold decrease in PTGDS mRNA expression in Tg-Mlkl−/− compared to Tg-Mlkl+/+ mice (Fig. 6a). IHC also confirmed that MLKL deficiency increased PTGDS protein levels in the cortex, striatum, and SN regions of the A53T transgenic (Tg) mice (Fig. 6b-6c).
Given the observed reduction of numerous proinflammatory cytokines in Tg-Mlkl−/− mice (Fig. 4g), we conducted a systematic examination of 16 PD-associated cytokines 38,39 within the identified microglia cluster. Our findings revealed that 11 cytokine genes, including Il1b, Il2, Il12a, Il6, Cxcl10, Il34, Il17d, Ccl2, Ccl5, Ccl4, and Tnf, were downregulated in Tg-Mlkl−/− mice (Fig. 5f). These results provide a detailed snapshot of global and cell-type-specific changes in gene expression and functional processes associated with MLKL deficiency in PD progression.
The subcluster-specific analysis identifies specific and functional cell transcriptomics subclusters after MLKL deficiency in the A53T transgenic mice
We utilized the Seurat algorithm to partition neurons, microglia, and astrocytes into subclusters based on their transcriptional characteristics. Through an unbiased analysis, we identified seven neuronal clusters, five microglial clusters, and seven astrocyte clusters, each enriched with specific functional categories and characterized by distinct molecular markers (Fig. 7a-7f). Interestingly, our findings revealed that MLKL deficiency did not significantly affect the proportion of microglial subclusters. In contrast, it led to the segregation of neurons and astrocytes into different clusters in Tg-Mlkl−/− mice compared to Tg-Mlkl+/+ mice (Fig. 7a-7f).
In particular, the n4 and n6 subclusters of Tg-Mlkl−/− mice showed higher proportions compared to those of Tg-Mlkl+/+ mice (Fig. 7a). These subclusters were enriched with genes associated with tissue growth and repair (Col6a2 and Col23a1), membrane proteins regulating trafficking, adipogenesis, or potassium absorption (Dlk2, Cpne2, and Atp12a), and neuronal RNA transporting (Stau2 and Elavl2) (Fig. 7b). Additionally, gene expression changes and enrichment analysis of the n4 subcluster in Tg-Mlkl−/− mice revealed an association with the response to cellular stress, including the cellular response to DNA damage stimulus, starvation, and nutrient levels (Fig. S5a). Furthermore, the upregulated genes of the n7 cluster in Tg-Mlkl−/− mice were linked to the negative regulation of the nitrogen compound metabolic process, consistent with our western blot data (Fig. 7b and S5a).
While only one microglial subcluster m4 exhibited a higher frequency in Tg-Mlkl−/− mice, most subcluster-specific genes across three microglial subclusters (m1-m3) were downregulated and enriched for processes including structural molecule activity, peptide metabolic and biosynthetic processes, and ribosome subunit (Fig. 7c-7d and S5b).
Regarding astrocyte subclusters, both a2 and a3 clusters were highly abundant in Tg-Mlkl−/− mice and expressed genes related to signal transduction and memory performance (Pde10a and Wwc1), biogenesis of mitochondrial ATP synthase (Tmem70), neurogenesis and synaptogenesis (Hes5 and Epha5) (Fig. 7e-7f). Notably, the cystine-glutamate exchanger, Slc7a11, which could inhibit ferroptosis and cell necrosis, and suppress inflammatory and oxidative responses, was upregulated in clusters a2 and a3 of Tg-Mlkl−/− mice (Fig. 7e-7f). Furthermore, the a3 subcluster of astrocytes in Tg-Mlkl−/− mice showed an upregulation of genes associated with the negative regulation of cytokine production, suggesting a non-inflammatory phenotype for this subcluster (Fig. S5c).