Progressive midbrain DAergic neuronal reduction in the SN of WDR45 cKOmice

To create a mouse model with selective deletion of WDR45 in the midbrain DAergic neurons, we used a TAM-inducible CreERT2/loxp gene-targeting system (Fig. S1a). WDR45-floxed mice were crossed with DATCreERT2 mice to generate WDR45Flox/Flox/DATCreERT2 mice. WDR45Flox/Flox mice were used as controls (WDR45cKO and WDR45cWT, respectively), and their genotype was confirmed by using conventional PCR analysis (Fig. S1b). When the mice reached 8 weeks of age, we administered intraperitoneal injections of TAM to both WDR45cWT and WDR45cKO mice. Tissues were collected 4 months after TAM administration (when the mice were 6 months old), and we used immunofluorescence (IFC) staining to detect WDR45 protein expression in the tyrosine hydroxylase (TH)-labeled DAergic neurons. The results showed a marked decrease in WDR45 expression in the DAergic neurons of WDR45cKO mice (Fig. S1c, d), indicating the successful deletion of WDR45 in these neurons.

WDR45cWT and WDR45cKO mice were subjected to behavioral tests at different ages, including 6–8 months (young), 11–13 months (middle-aged), and 17–19 months (aged). The results showed that the aged WDR45cKO mice had a significant motor impairment, as indicated by a decrease in the total distance traveled in the open-field test (Fig. S2a) and a notable reduction in stereotypic counts (Fig. S2b), suggesting an increased vulnerability to motor activity impairments with aging. However, no abnormalities were observed in the rotarod test (Fig. S2c). In addition to the locomotion deficits, the aged WDR45cKO mice also showed poor immediate spatial working memory performance, as evidenced by a decreased spontaneous alteration proportion in the Y maze test (Fig. S2d). Furthermore, the 3-chamber social performance test revealed that the aged WDR45cKO mice displayed less motivation in novelty (Fig. S2e-h). Moreover, WDR45cKO mice spent less time in exploration (Fig. S2i, j), suggesting that WDR45 dysfunction in midbrain DAergic neurons may lead to depression-like behavior in aging mice.

To further investigate the survival of DAergic neurons, we performed IFC staining for TH, a classic marker of DAergic neurons. We analyzed the number of DAergic neurons in the substantia nigra pars compacta (SNc) and ventral tegmental area (VTA) of WDR45cWT and WDR45cKO mice at young, middle-aged, and aged stages. We observed a reduction in the number of DAergic neurons in the VTA and SNc of middle-aged WDR45cKO mice compared to age-matched WDR45cWT mice, and this reduction was significantly exaggerated in the aged WDR45cKO mice (Fig. 1a, b). Additionally, we found a significant decrease in the DA content in the SN of aged WDR45cKO mice compared to that of aged WDR45cWT mice (Fig. 1c).

Fig. 1

DAergic neuronal reduction in the SN. a IFC staining was performed using an antibody against TH (red) in midbrains from young (6–8 months old), middle-aged (11–13 months old), and aged (17–19 months old) WDR45cWT and WDR45cKO mice. Scale bar, 250 μm. Scale bar for high-magnification images, 10 μm. b Quantifying TH-positive neurons in the VTA and SNc of WDR45cWT and WDR45cKO mice (N = 5 mice per genotype). c The dopamine concentration in the SN region was detected by high-performance liquid chromatography (N = 3–5 mice per genotype). d Representative TEM images of observed mitochondria in aged WDR45cWT mice and WDR45cKO mice. Scale bar, 500 nm. e Quantification of the perimeter of mitochondria in DAergic neurons (N = 154 mitochondria collectively counted from 9 slices of 3 WDR45cWT mice and 251 mitochondria from 9 slices of 3 WDR45cKO mice). f The proportion of mitochondria with damaged cristae was quantified (N = 15 slices from 3 mice per genotype). g The mean number of mitochondria observed in captured images was collected (N = 15 slices from 3 mice for each genotype). h Representative TEM images of observed RER. Scale bar, 500 nm. i The mean width of RER tubules is shown (N = 163 RER collectively counted from 9 slices of 3 WDR45cWT mice and 288 RER from 9 slices of 3 WDR45cKO mice). j The proportion of RER tubules (> 100 nm) was quantified (N = 15 slices from 3 mice per genotype). k The mean number of RER observed in captured images was collected (N = 15 slices from 3 mice for each genotype). l Double-label immunofluorescence of p-RIPK3 (Thr 231/Ser232) or p-MLKL (phosphor S345) (green) with TH (red) in the DAergic neurons of young, middle-aged, and aged WDR45cWT mice and WDR45cKO mice. Scale bar, 10 μm. m The proportion of TH-positive neurons with p-RIPK3 puncta was quantified. (N = 3 mice per genotype). n The proportion of TH-positive neurons with p-MLKL puncta was quantified. (N = 3 mice per genotype). Data were analyzed using two-way ANOVA followed by Sidak’s multiple comparisons tests (b, c, m, n) and Student’s t-test (e–g, i-k). Data are represented as the mean ± SEM. *p < 0.05, ***p < 0.001, ****p < 0.0001

We conducted further analysis to investigate the impact of WDR45-deficiency on DAergic neurons at the subcellular level. We used TEM analysis to assess mitochondrial morphology in aged WDR45cKO mice. Our results showed the presence of vacuolized mitochondria in the neuronal soma from the SN region of aged WDR45cKO mice but not in age-matched WDR45cWT mice (Fig. 1d). The perimeter of mitochondria was significantly increased in WDR45cKO mice compared to WDR45cWT mice, and more than 40% of the mitochondria cristae of WDR45cKO mice were broken or disappeared (Fig. 1e, f), indicating mitochondrial may damage upon WDR45-deficiency in DAergic neurons. The ER-mitochondria contact sites captured in the TEM images were analyzed, without significant alterations in the total number of their contact sites (Fig. S3a, b). The mitochondria in the striatum were also analyzed, and the proportion of mitochondria with damaged cristae (broken or completely vacuolized) was significantly increased (Fig. S3f, g). Moreover, the BCL2 Interacting Protein 3 (BNIP3, as a mitophagy receptor) positive puncta were upregulated in the striatum (Fig. S3c-e), indicating the accumulation of damaged mitochondria in the striatal region, where receives the projections from midbrain DAergic neurons. We also analyzed the rough endoplasmic reticulum (RER) structure and found a significant change in RER morphology in aged WDR45cKO mice (Fig. 1h). The mean width of RER tubules was significantly expanded in the neurons from SN area of WDR45cKO mice compared to that in WDR45cWT mice (79.9 nm vs. 51.7 nm) (Fig. 1i). The proportion of RER tubules (> 100 nm) was increased from 6.4% in the aged WDR45cWT mice to 23.9% in the WDR45cKO mice (Fig. 1j), indicating that the DAergic neurons may suffer from severe RER tubular expansion due to WDR45 deletion. The proportion of mitochondria with broken cristae or completely vacuolized were significantly increased (Fig. 1f). Moreover, the mean number of mitochondria captured in TEM images was also dramatically increased in the WDR45cKO mice compared to that in the WDR45cWT mice (Fig. 1g), which was further verified by the significant increase of the intensity of translocase of outer mitochondrial membrane 20 (TOM20) in the soma of DAergic neurons detected by IF analysis (Fig. S4a, b). All these findings indicated an impaired mitochondrial homeostasis, possibly due to the damaged mitochondria accumulation during the WDR45 deficiency-induced impaired autophagic process. Similarly, in addition to the increased number of swollen RER tubules (Fig. 1i, j), the mean number of captured RER tubules in TEM images was also remarkably increased (Fig. 1k). The presence of lysine-aspartic acid-glutamic acid-leucine (KDEL) is necessary for ER retention and to be sufficient to reduce the secretion of proteins from the ER. We also found that KDEL expression dramatically increased in the DAergic neurons of the WDR45cKO mice (Fig. S4f). Additionally, to further clarify whether the associated functions of mitochondria and ER were impacted in the DAergic neurons, we detected several protein expressions in the DAergic neuronal soma by IFC staining and found BNIP3 expression in the nuclei was dramatically enhanced in the young WDR45cKO mice compared to that in the young WDR45cWT mice (Fig. S4b). However, in the aged mice, the BNIP3 positive puncta were mainly concentrated in the DAergic cytoplasm, with a significant increase in the WDR45cKO mice (Fig. S4b). Additionally, the mitochondrial fission processes were predominantly impacted under the WDR45 deficit since the intensity of fission mitochondrial 1 (FIS1), the key protein that mediates the mitochondrial fission process, was markedly increased in the DAergic neurons of both young and aged WDR45cKO mice (Fig. S4c-e). ER-associated proteins were also detected to examine the ER relative function, including SEC16 Homolog A, Endoplasmic Reticulum Export Factor (SEC16A, has a key role in the organization of the ER exit site), and SEC31 Homolog A, COPII Coat Complex Component (SEC31A, as the component of the coat protein complex II, which promotes the formation of transport vesicles from the ER). The expression of SEC16A in the DAergic neurons was significantly increased in the young WDR45cKO mice, while SEC31A expression down-regulated dramatically in both young and aged WDR45cKO mice, indicating the WDR45 deficit has an impact on the secretion and transport for proteins from ER (Fig. S4g, h). Overall, these findings suggest that WDR45 deletion may inhibit the clearance or turnover of damaged organelles, accelerating the degeneration of DAergic neurons.

Given that the WDR45cKO mice experience progressive reduction of DAergic neurons during aging, we aimed to investigate the mechanisms underlying cell reduction. Necroptosis is a regulated form of necrosis and is considered a new mode of cell death. When necroptosis is induced, receptor-interacting protein kinase-3 (RIPK3) becomes activated through phosphorylation and then phosphorylated RIPK3 activates mixed lineage kinase-like (MLKL) through phosphorylation [30, 31]. To determine if necroptosis was activated in the DAergic neurons of the WDR45cKO mice, we assessed the presence of phosphor-RIPK3 and phosphor-MLKL puncta in neuronal cytoplasm using our previously established methods [32]. Our data revealed a marked increase in the concentrated puncta of phosphor-RIPK3 and phosphor-MLKL in the cytoplasm of middle-aged and aged WDR45cKO mice (Fig. 1l, m, n). These results indicate that necroptosis was activated in the DAergic neurons, leading to progressive DAergic neuronal reduction in the WDR45cKO mice during aging.

Axonal degeneration in the striatum of WDR45cKO mice

In addition to the reduction of DAergic neurons in the SNc, WDR45 deletion also led to profound nerve fiber pathology in the striatum. Our longitudinal study revealed substantial changes in DAergic axonal terminals projected to the striatum of WDR45cKO mice. Specifically, we observed significant axonal enlargements in the young, middle-aged, and aged WDR45cKO mice, along with reduced fiber density during aging (Fig. 2a-c). It is worth noting that significantly more enlargements were accumulated in the nucleus accumbens (NAc), which receives the projection of DAergic neurons from the VTA, than in the caudate putamen (CPu) from SNc DAergic neurons, in the young, middle-aged, and aged WDR45cKO mice (Fig. 2d, e). These findings demonstrate that WDR45cKO mice develop severe DAergic axonal degeneration in the striatum prior to neuronal loss and reveal the differential axonal vulnerability of DAergic neuronal subtypes in response to WDR45 deletion.

Fig. 2

Axonal degeneration in the striatum of WDR45cKO mice. a IFC staining for striatal axons, including the NAc and CPu, was performed using an antibody against TH (red) in young, middle-aged, and aged WDR45cWT and WDR45cKO mice. Scale bar, 200 μm. For high-magnification images: 50 μm. b, c Quantifying the fiber density in the NAc and CPu, respectively (N = 3 mice per genotype). d, e The calculation of densities of DA axonal enlargements (area > 5 μm2) per 0.045 mm2 perspective in the NAc and CPu from WDR45cWT and WDR45cKO mice, respectively (N = 3–7 slices from 3 mice per genotype). f Representative TEM images of the observed PSD. Scale bar, 100 nm. The red arrowhead indicates PSD. g, h The PSD width and PSD area were quantified (N = 33 PSD collectively counted from 3 WDR45cWT mice and 35 PSD from 3 WDR45cKO mice). i IFC analysis of synapse-related proteins in the striatum of aged WDR45cWT mice and WDR45cKO mice. Scale bar, 20 μm. j Quantifying PSD95' fluorescence density (N = 7–8 slices from 3 mice per genotype). k Quantifying SYT1' fluorescence density (N = 11 slices from 3 mice per genotype). l Quantifying SYN1' fluorescence density (N = 5–6 slices from 3 mice per genotype). m Quantifying HOMER1' fluorescence density (N = 5 slices from 3 mice per genotype). n Quantifying BSN' fluorescence density (N = 7–8 slices from 3 mice per genotype). Data (b-e) were analyzed using two-way ANOVA followed by Sidak’s multiple comparisons test and Student’s t-test (g, h, j-n). Data are represented as the mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Since large enlargements were observed in the axonal terminals, we decided to examine whether the striatal synapses were affected. To study the effect of WDR45 deletion on excitatory synapses of DAergic projections, we conducted TEM analysis for postsynaptic density (PSD), which contributes to information processing and memory formation by changing synaptic strength in response to neural activity [33]. The results showed that PSD density was significantly reduced (Fig. 2f), and PSD width and average area were significantly reduced (Fig. 2g, h). The data suggest that the synaptic structures in the striatum of WDR45cKO mice have undergone alterations. Furthermore, we assessed the levels of some synaptic proteins in the striatum, including PSD95, a membrane protein of presynaptic vesicles called synaptotagmin 1 (SYT1), a synaptic vesicle protein called synapsin-1 (SYN1), postsynaptic density scaffolding protein called homer scaffold protein 1 (HOMER1), and presynaptic cytomatrix protein bassoon (BSN). We observed a significant decrease in the expression of presynaptic-related proteins, BSN and SYT1, in the striatum of young WDR45cKO mice (Fig. S5e), and a significant decrease in the fluorescence density of PSD95, SYT1, SYN1, HOMER1, and BSN in the striatum of aged ones (Fig. 2i-n), indicating the progressive dysfunction of synapses under WDR45 deficit in aging. Additionally, we detected the associated proteins of DAergic axonal terminals, including DAT, dopamine receptors D1 and D2 (DRD1, DRD2), and vesicular monoamine transporter member 2 (vMAT2). The expression of striatal DRD1 was significantly decreased, while DRD2 expression was dramatically increased in the aged WDR45cKO mice (Fig. S5a-d), which may be associated with depression-like behavior. These results further support that synaptic signaling transmission is disrupted under the longtime dysfunction of WDR45.

Accumulation of increased fragmented tubular ER constitutes a pathological feature of swollen axons in the WDR45cKO mice

Axonal swellings (also called axonal beading, bubbling, or spheroid) are hallmarks of degenerating axons, almost universal in neurodegenerative diseases [34, 35]. In our study, WDR45 depletion in the DAergic neurons resulted in axonal swellings in the striatum. To gain insights into the molecular basis of axonal degeneration, we evaluated potential candidates by investigating their presence at the axonal enlargements. First, we examined the ER proteins KDEL and cytoskeleton-associated protein 4 (CKAP4, also called Climp-63) [36], the tubular ER protein ATL3, and the tubular ER-shaping proteins RTN3 and RTN4 [37, 38]. RTN4, KDEL, Climp-63, and ATL3 were not observed in the axonal enlargements (Fig. S6). By contrast, RTN3 was highly concentrated at the striatal axonal enlargements in young, middle-aged, and aged WDR45cKO mice (Fig. 3a, d), suggesting that RTN3 is one of the enlargement components and may contribute to the formation of axonal swellings as an early pathogenic event. As a typical tubular ER-shaping protein, the accumulation of RTN3 implies that the tubular ER shape may be affected in the striatum. We then investigated the molecular composition of axonal enlargements by determining other tubular ER-shaping proteins, REEP2 and REEP5 [39, 40]. We found that REEP2 and REEP5 also colocalized with TH-positive enlargements (Fig. 3b, c, e, f), further indicating that the shape of tubular ER in the axons was disrupted upon WDR45 depletion. These findings highlight the crucial role of ER-shaping proteins in forming axonal enlargements, providing further evidence for the importance of maintaining a normal tubular ER shape in regulating distal axonal homeostasis. The above findings prompted us to determine whether the tubular ER shape is abnormal in the WDR45cKO mice. We then examined the tubular ER ultrastructure by TEM in the striatal samples from aged WDR45cWT and WDR45cKO mice. Compared to the normally distributed tubular ER in WDR45cWT mice, a remarkably large accumulation of fragmented tubular ER was noticed in the axons of WDR45cKO mice (Fig. 3g-k), supporting the notion that the fragmented tubular ER cluster is a major pathological abnormality associated with axonal degeneration in WDR45cKO mice.

Fig. 3

Increasing fragmented tubular ER constitutes a pathological feature of axons in WDR45cKO mice. a IFC analysis for RTN3 in the NAc of WDR45cWT mice and WDR45cKO mice was performed using antibodies against RTN3 (green) and TH (red). The nuclei were labeled with DAPI (blue). Scale bar, 10 μm. b IFC staining for REEP2 in the NAc of aged WDR45cWT mice and WDR45cKO mice was performed using antibodies against REEP2 (green) and TH (red). The nuclei were labeled with DAPI (blue). Scale bar, 10 μm. c IFC staining for REEP5 in the NAc of aged WDR45cWT mice and WDR45cKO mice was performed using antibodies against REEP5 (green) and TH (red). The nuclei were labeled with DAPI (blue). Scale bar, 10 μm. d Analysis of relative density of RTN3- and TH-positive enlargements in the NAc of aged WDR45cWT mice and WDR45cKO mice (N = 5–9 slices from 3 mice per genotype). e Analysis of relative density of REEP2- and TH-positive enlargements in the NAc of aged WDR45cWT mice and WDR45cKO mice (N = 9 slices from 3 mice per genotype). f Analysis of relative density of REEP5- and TH-positive enlargements in the NAc of aged WDR45cWT mice and WDR45cKO mice (N = 9 slices from 3 mice per genotype). g-j Samples from aged WDR45cWT mice and WDR45cKO mice were examined by TEM, and representative TEM images of observed tubular ER at the axons of the striatum are shown. The tubular ER is highlighted in black. Scale bar, 500 nm. For enlarged images, 250 nm. k The mean length of tubular ER was analyzed from aged WDR45cWT mice and WDR45cKO mice (N = 8–9 slices from 3 mice for each genotype). Data were analyzed by using Student’s t-test. Data are represented as the mean ± SEM. ****p < 0.0001. White arrows indicate axonal enlargements

Disrupted autophagic flux in the DAergic neurons may contribute to the accumulation of tubular ER in axons

To understand what contributes to the accumulation of tubular ER at axons, we first examined whether autophagic flux was disrupted in the DAergic neurons of WDR45cKO mice. We stained midbrain sections to detect the expression and distribution of autophagic substrates SQSTM1 (p62) and Ub, as well as LC3 (a classical marker for autophagic vesicles). Compared with WDR45cWT mice, we observed distinct p62-positive puncta accumulated in the soma of DAergic neurons in young WDR45cKO mice, and this accumulation was aggravated in the aged WDR45cKO mice (Fig. 4a, b). Additionally, we found that Ub expression was significantly increased in the nucleus of DAergic neurons of young WDR45cKO mice and in the cytoplasm of DAergic neurons of aged WDR45cKO mice, of which the Ub staining was not entirely colocalized with p62-positive puncta (Fig. 4a, c). Similarly, LC3-positive puncta were also concentrated in the cell body of DAergic neurons in the aged WDR45cKO mice (Fig. 4a, d). These data suggest that WDR45 depletion induced an early impairment of autophagic flux in the DAergic neurons, likely triggering axonal and cell body degeneration.

Fig. 4

Disrupted autophagic flux in the DAergic neurons may contribute to the accumulation of tubular ER in axons. a Left panel: IFC staining for p62 (purple) and Ub (green) in the TH-positive neurons (red) of WDR45cWT mice and WDR45cKO mice. The nuclei were labeled with DAPI (blue). Scale bar, 10 μm. Right panel: IFC staining for LC3 (green) in the TH-positive (red) DAergic neurons. The nuclei were labeled with DAPI (blue). Scale bar, 10 μm. b The proportion of TH-positive neurons with p62 puncta (> 0.5 μm2) is presented (N = 3 mice per genotype). c The proportion of TH-positive neurons with Ub-positive puncta (> 0.5 μm2) is presented (N = 3 mice per genotype). d The proportion of TH-positive neurons with LC3-positive puncta (> 0.5 μm2) is presented (N = 3 mice per genotype). e–h LC3 (green), Lamp1 (red), Ub (green), and p62 (green) were detected in the NAc of aged WDR45cWT mice and WDR45cKO mice. The nuclei were labeled with DAPI (blue). Scale bar, 10 μm. i-k IFC staining of RTN3, REEP2, and REEP5 in the NAc of 12-month-old VMP1cWT mice and VMP1cKO mice was performed using antibodies against RTN3 or REEP2 or REEP5 (green) with TH (red), respectively. The nuclei were labeled with DAPI (blue). Scale bar, 10 μm. l-n Analysis of the relative density of RTN3- and TH-positive enlargements, REEP2- and TH-positive enlargements, and REEP5- and TH-positive enlargements (> 5 μm2), respectively (N = 5–9 slices from 3 mice per genotype). Data were analyzed using two-way ANOVA followed by Sidak’s multiple comparisons tests (b-d) and Student’s t-test (l-n). Data are represented as the mean ± SEM. ****p < 0.0001, **p < 0.01

To assess whether LC3-labeled autophagosomes are present in axonal enlargements, we stained the striatal sections and found that LC3-positive puncta were absent in the TH-positive axonal enlargements (Fig. 4e), indicating that the LC3-labeled autophagosomes did not directly contribute to the formation of axonal enlargements. Furthermore, the lysosome marker Lamp1 was also absent in the axonal enlargements (Fig. 4f). The autophagic substrates Ub and p62 were also not colocalized with axonal enlargements (Fig. 4g, h). Therefore, the accumulations of autophagic proteins were mainly observed in the soma but not in the axons of DAergic neurons deficient in WDR45. We speculate that disrupting autophagic flux in the DAergic neurons may lead to axonal enlargements by promoting tubular ER accumulation at axons. To test this hypothesis, we employed another mouse model with damaged autophagic flux, the VMP1cKO mice that conditionally knocked out autophagic gene VMP1 in the DAergic neurons upon TAM treatment postnatally [32]. VMP1cKO mice also displayed severe damage to autophagic flux and large axonal enlargements in the striatum [32]. RTN3, REEP2, and REEP5 were highly accumulated at the TH-positive axonal enlargements in the striatum of 12-month-old VMP1cKO mice (Fig. 4i-n), indicating that defective autophagy may induce axonal accumulation of tubular ER. Together, these results suggest that the abnormal clustering of tubular ER in axons may have pathological effects on the brain. Furthermore, our findings provide additional evidence that autophagy plays a critical role in maintaining axonal homeostasis by regulating the shape and accumulation of tubular ER.

Exploring the proteome landscape of striatum in the WDR45 cKO mice

To gain a general landscape of the pathological abnormalities in the DAergic axons, we dissected the striatal samples from both young and aged WDR45cKO mice, age- and sex-matched WDR45cWT mice for proteomic analysis. Principal component analysis (PCA) revealed the distinct proteomic profiles of WDR45cWT mice and WDR45cKO mice (Fig. S7a, b). Further proteomic analyses registered 6,647 targets in the young mice and 6,290 targets in the aged mice, of which 31 differentially expressed proteins (DEPs) in the young mice and 167 DEPs in the aged mice were identified (> 1.3-fold or < 1/1.3-fold change cutoff, p < 0.05). Among the DEPs, 17 from young mice were upregulated, 14 were downregulated, 115 DEPs from aged mice were upregulated, and 47 were downregulated (Fig. S7c, d, and Supplementary Table 1). In the young mice, the majority of the top 10 up-regulated proteins are involved in the regulation of carbohydrate metabolic process, such as Beta-enolase3 (Eno3), and in the regulation of lipid metabolic process, like lysophosphatidylcholine acyltransferase 1 (Lpcat1), as well as regulating nitrogen compounds' metabolic process, like mannosyl-oligosaccharide 1,2-alpha-mannosidase IA (Man1a1) (Fig. 5a, b). The top 10 downregulated DEPs are most associated with nucleobase-containing compound metabolic process, regulation of cytokine production, and nitrogen compounds' metabolic process (Fig. 5a, b). In the aged mice, the majority of the top 20 up-regulated proteins were enzymes that are involved in the regulation of lipid metabolic process, such as Lpcat1, ethanolamine-phosphate phospho-lyase (Etnppl), abhydrolase domain containing 4, N-acyl phospholipase B (Abhd4), and phytanoyl-CoA dioxygenase domain-containing protein 1 (Phyhd1), and in the regulation of nitrogen compounds' metabolic process, such as ElaC ribonuclease Z1 (Elac1), aminomethyltransferase (Amt), lactate dehydrogenase D (Ldhd), and cold-inducible RNA binding protein (Cirbp), as well as in regulating anatomical structure morphogenesis, like secreted protein acidic and cysteine-rich (Sparc), angiotensinogen (Agt) (Fig. 5a, b). The top 20 downregulated DEPs are most associated with nitrogen compounds' metabolic process, such as complex integrator subunit 4 (Ints4), keratin 2 (Krt2), and strawberry notch homolog 2 (Sbno2), and proteins with cell morphogenesis, such as protein cordon-bleu (Cobl), amyloid β-A4 precursor protein-binding family B member 1-interacting protein (Apbb1ip) (Fig. 5a, b).

Fig. 5

The proteome landscape of striatum in the WDR45 cKO mice. a The heatmap of the DEPs from young and aged WDR45cWT mice and WDR45cKO mice. b Volcano plots and top 20 up- or down-regulated DEPs organized by fold change in the striatum of young and aged WDR45cKO mice vs. WDR45cWT mice (DEPs marked by red and blue circles). The top 10 CC, MF, and BP terms in GO annotation analysis for DEPs from (c) young and (d) aged WDR45cWT mice and WDR45cKO mice. The top 20 terms related to BP in the GO enrichment analysis for DEPs from (e) young and (f) aged WDR45cWT mice and WDR45cKO mice. The top 20 terms related to MF in the GO enrichment analysis for DEPs from (g) young and (h) aged WDR45cWT mice and WDR45cKO mice. Data were analyzed by using the Student’s t-test. Data are represented as the mean ± SEM. ****p < 0.0001

To support the biological significance of these DEPs, we performed Gene Ontology (GO) annotation analysis, which depicts protein functions in three categories: biological processes (BP), cellular components (CC), and molecular functions (MF). The most correlated BP of the DEPs from both young and aged mice is the regulation of the biological process, metabolic-related process, including organic substance, cellular, primary, and nitrogen compounds metabolic process, and that regulation of anatomical structure development (Fig. 5c, d). Regarding CC, these DEPs are mostly found in the intracellular anatomical structure, cytoplasm, and organelle (Fig. 5c, d). The analysis results of MF show that these DEPs are primarily associated with protein binding, ion binding, organic cyclic compound binding, and hydrolase activity, etc. (Fig. 5c, d). To further investigate the functions and signaling pathways of the DEPs, we performed GO enrichment analyses. In the young mice, the top 20 enriched BP pathways are most related to positive regulation of receptor internalization, protein localization to lysosome and vacuole, regulation of phospholipid metabolic process, and positive regulation of protein catabolic process (Fig. 5e). While in the aged group, amino acids' catabolic and biosynthetic processes, positive regulation of the lipid catabolic process, and protein depalmitoylation are the most enriched pathways (Fig. 5f). Notably, many DEPs from young mice are involved in regulating ion channel activity and transmembrane transporter activity, like outward rectifier potassium channel activity and voltage-gated potassium channel activity, cation channel activity, and ion transmembrane transporter activity (Fig. 5g). As aging progresses, the regulation of DEPs is mostly involved in the enzyme activity, like oxidoreductase activity, Amt, and steroid hydroxylase, that are enriched in the amino acid metabolic process pathway, including lysine, L-cysteine, serine family amino acids, glycine, aspartate family amino acids, and in the tricarboxylic acid metabolic process, lipid catabolic process (Fig. 5h).

Overall, the proteomic data indicate that after WDR45 deficiency in the DAergic neurons, proteins that regulate ion transmembrane transporter activity and ion channel activity change earlier. The main changes brought by these proteins focus on regulating receptor internalization, protein localization and catabolism, and chemical synaptic transmission. The impact of WDR45 deficiency becomes more profound with aging. The proteins that undergo changes mainly focus on those that regulate enzyme activity, such as oxidoreductase activity, Amt activity, and glutathione transferase activity, mainly affecting amino acid and lipid metabolism, and there may be severe oxidative stress reactions in the brain. The activated catabolism of these molecules may also indicate an energy supply deficit in the striatum, supported by the enrichment of DEPs in the tricarboxylic acid metabolic process that produces adenosine triphosphate (ATP) for cellular energy (Fig. 5h).

The connection of the phospholipid metabolism with the striatal pathology

According to the proteomic data, 17.6% of up-regulated DEPs in young mice and 18% in aged mice were found to regulate lipid metabolism. Additionally, among down-regulated DEPs, 21% from young mice and 15% from aged mice were involved in lipid metabolism (Fig. 6a, b). To validate these findings from proteomics, we performed qRT-PCR and confirmed the role of lipid metabolism in striatal pathology (Fig. S7e, f). Notably, the expression of Lpcat1 (also called AYTL2), a phospholipid biosynthesis/remodeling enzyme that facilitates the conversion of palmitoyl-lysophosphatidylcholine (PPC) to dipalmitoyl-phosphatidylcholine (1,2-dipalmitoyl-sn-glycero-3-phosphocholine, DPPC), was significantly increased both in young and aged WDR45cKO mice (Fig. 6c). To gain further insights into the association between striatal pathology and lipids, particularly phospholipids, we performed a comprehensive lipidomic analysis with a specific focus on phospholipid metabolites. The striatal lipid profile of both young and aged WDR45cKO mice exhibited distinct separation from that of WDR45cWT mice in both positive and negative ionization modes, indicating a significant alteration in the striatal lipid profile due to the loss of WDR45 in DAergic neurons (Fig. 6d, e).

Fig. 6

The connection of the phospholipid metabolism with the striatal pathology. GO analysis for BP displays up-regulated DEPs (pink) (a) and downregulated DEPs (green) (b) that participate in the regulation of lipid metabolism. Lower blue bars represent the magnitude of p values. Percentages indicate the fraction of each category of total up- or down-regulated DEPs. c The Venn diagram of the number of DEPs in young and aged groups. d PLS-DA score plot of the lipidomic profile of striatum tissues from both young and aged WDR45cKO mice and WDR45cWT mice in positive mode. R2X = 0.454, R2Y = 0.492, Q2 = 0.219. e PLS-DA score plot of the lipidomic profile of striatum tissues from both young and aged WDR45cKO mice and WDR45cWT mice in negative mode. R2X = 0.518, R2Y = 0.510, Q2 = 0.271. f Volcano plot of DELs between young WDR45cKO mice and WDR45cWT mice. g Volcano plot of DELs between aged WDR45cKO mice and WDR45cWT mice. h Venn diagram of the number of DELs between WDR45cKO mice and WDR45cWT mice in young and aged groups. Heat maps of DELs between WDR45cKO mice and WDR45cWT mice in (i) young groups, (j) aged group, and (k) both young and aged groups. Percentage of DELs in each lipid class between WDR45cKO mice and WDR45cWT mice in (l) young groups, (m) aged group, and (n) both young and aged groups

Further lipidomic analyses identified 91 differentially expressed lipids (DELs) in young mice and 77 DELs in aged mice. Among the DELs from young WDR45cKO mice, 56 were down-regulated and 35 were up-regulated compared to young WDR45cWT