Proteomic profiling reveals mitochondrial metabolic alterations in dexamethasone-induced neuronal differentiation

Cell culture

Mouse CNS-derived CAD cell line (RRID:CVCL_0199) was obtained from European Collection of Authenticated Cell Cultures (ECACC, accession number: 08100805) and cultured in Dulbecco’s modified Eagle’s medium (DMEM): nutrient mixture F-12 (F12) (Gibco, USA) supplemented with 8% fetal bovine serum (FBS) (HyClone, USA) and 1% antibiotic-antimycotic 100X (Gibco, USA). The cells were maintained in a humidified incubator at 37 °C with 5% CO₂ and 95% air (v/v). The cultured cells were regularly tested for mycoplasma contamination via a Mycoplasma PCR detection kit (Abcam, UK).

Differentiation induction

Immortalized CAD cells cultured in complete DMEM/F12 supplemented with 8% FBS were considered undifferentiated or proliferating. Differentiation was conducted following 2 protocols previously reported. ² The complete medium was replaced with serum-free (0% FBS) DMEM/F12 to induce serum-free differentiated cells. To induce dexamethasone-induced neuronal differentiation, CAD cells were cultured in DMEM/F12 containing 1% FBS and dexamethasone (D4902, Sigma‒Aldrich, USA) at 100 μM. This concentration was selected on the basis of our previous work, in which methanol (vehicle control) had no detectable effect on the differentiation process. ² The culture media were replenished every 3 days. The cells were differentiated for 5 days and the pellets were collected for the proteomic study. Successful differentiation was indicated by morphological changes, including neurite outgrowth and arrested cell proliferation, and the cells were observed via a 10X objective lens on an inverted microscope (ECLIPSE TS100, Nikon, Japan).

Immunocytochemistry of beta-III tubulin and Ki67

CAD cells were plated at a density of 2 × 10⁴ cells/well on a coverslip placed on a 12-well plate. The plated cells were incubated under different conditions as indicated for 5 days. CAD cells were fixed and permeabilized with prechilled ethanol (95–100%) for 15 min at -20 °C. The cells were then processed as described in our previous study. ² The cells were incubated in 10% BSA (A2153, Sigma‒Aldrich, USA) dissolved in 1X PBS solution and incubated with mouse monoclonal anti-beta III (neuronal) clone 2G10 (T8578, Sigma‒Aldrich, USA) or rabbit polyclonal anti-Ki67 (cat. no. ab15580, Abcam, UK) in 1% BSA/PBS for 1 h at room temperature. The cells were incubated with goat anti-mouse IgG (H+L), a superclonal secondary antibody conjugated to Alexa Fluor 488 (cat.no. A28175, Invitrogen, USA), or donkey anti-rabbit IgG (H+L) highly cross-adsorbed secondary antibody, Alexa Fluor Plus 488 (cat.no. A32790, Invitrogen, USA). DAPI (Invitrogen, USA) was used for nuclear staining. The experiments were conducted with 3 biological replicates. The images were collected from 5 nonoverlapping areas via a 10-20X objective lens on a Fluoview FV10i laser scanning microscope (Olympus, Japan). The raw files were processed via Fiji version 1.54p (released February 17, 2025; https://imagej.net/software/fiji).

Cell proliferation assay

The relative number of cells was detected via an MTT assay (cat.no. M5655, Sigma‒Aldrich, USA) after 3 days of induction. Briefly, the immortalized cells were plated on 96-well plates at a density of 2,500 cells per well and maintained for 24 h before the media were removed. Different conditioning media (200 µl) were added to each well to induce proliferating or differentiated cells. The MTT stock reagent (20 µl, 5 mg/ml) was directly added to the conditioning media to obtain the final concentration (0.5 mg/ml) of the MTT reagent. The plates were further cultured for an additional 2.5 h. Subsequently, the MTT solution was carefully removed, and insoluble formazan was dissolved in 100 µ DMSO (cat.no. D8418, Sigma‒Aldrich, USA). The colorimetric determination was measured at 570 nm via a spectrophotometer (EZ Read2000 Microplate reader, Biochrom, USA). The data were normalized according to the relative fold change in the number of cells.

Quantitative real-time-PCR

The mRNA expression of TP53 gene was determined by real-time quantitative PCR. Briefly, undifferentiated and the Dex-or SF-differentiated CAD cells were cultured for 5 days before total RNA extraction using TRIzol (Invitrogen, USA). RNA concentration and purity were evaluated using a NanoDrop One^C Microvolume UV-Vis Spectrophotometer (Thermo Fisher Scientific, USA). The cDNA was synthesized using iScript cDNA synthesis (Bio-Rad, USA) according to the manufacturer’s protocol with Applied Biosystems, SimpliAmp Thermal Cycler (Thermo Fisher Scientific, USA). Quantitative real-time PCR was performed on an equal amount of cDNA (100 ng) from each condition using iTag Universal SYBR Green supermix (Bio-Rad, USA) with the qTOWER 3 touch real-time PCR Thermal cycler and analyzed with qPCRsoft version 4.1 (Analytik Jena, Germany). The mRNA level of GAPDH was used as a reference gene. Ct values of the sample were calculated, and transcript levels were analyzed by the 2^-ΔΔCt method. The primers used in this study are listed in Table 1. Final concentration for each primer was 400 nM.

OPEN IN VIEWER

Table 1.

Primers used for this study.

Gene name	Forward primer	Reverse primers
TP53	TCACAGTCGGATATCAGCC	ATCTTCTGGAGGAAGTAGTTTCC
GAPDH	TGTGTCCGTCGTGGATCTGA	CCTGCTTCACCACCTTCTTGA

Protein extraction and in-solution digestion for proteomic study

The cells were seeded at a density of 2 × 10⁵ cells per Petri dish (70 × 15 mm) and maintained for 24 h. Differentiation induction was carried out for 5 days. The cell pellets were collected, washed 3 times with cooled 1X PBS, and stored at -80 °C until use. Four biological replicates were prepared for each group. The samples were subjected to LC‒MS/MS analysis at the National Omics Center, National Science and Technology Development Agency, Thailand.

The samples were lysed with lysis buffer solution (0.5% Triton X-100, 10 mM DTT, and 10 mM NaCl in 50 mM HEPES-KOH, pH 8.0). The proteins extracted were precipitated with ice-cold acetone and stored at -20 °C for 16 hours. After precipitation, the protein pellet was reconstituted in 0.25% RapidGest SF (Waters, UK) in 10 mM ammonium bicarbonate. The protein concentrations were measured with a Bradford Reagent assay kit (Sigma‒Aldrich, USA) using bovine serum albumin as the standard. A total of 20 μg of protein was subjected to trypsin digestion. The sulfhydryl bonds were reduced by using dithiothreitol (DTT, 10 mM) in ammonium bicarbonate (10 mM) at 62 °C for 20 minutes, and the sulfhydryl groups were alkylated with iodoacetamide (IAA, 20 mM) at room temperature for 25 minutes in the dark. The solution was cleaned using a desalting column (Thermo Fisher Scientific, USA). The flow-through solution was enzymatically digested with trypsin (Thermo Fisher Scientific, USA) at a ratio of 1:50 (enzyme:protein) and incubated at 37 °C for 6 hours. The digested peptides were reconstituted in 0.1% formic acid and transferred to a TruView LCMS vial (Waters, UK).

DDA-based spectral generation

A Sciex triple TOF-6600+ mass spectrometer (SCIEX, USA) coupled to an EASY-nLC1000 nanoliquid chromatography (LC) system (Thermo Fisher Scientific, USA) equipped with a nanoanalytical column (75 μm i.d. × 15 cm, packed with Acclaim PepMap C18) (Thermo Fisher Scientific, USA) was used for proteomic analysis. For spectral library generation, digested peptide samples (1.0 μg) were analyzed using data-dependent acquisition (DDA). Peptides were separated using mobile phase A (0.1% formic acid (FA) in water) and mobile phase B (80% acetonitrile with 0.1% FA). The samples were loaded onto the nanoanalytical column, and the first linear gradient was separated according to 95 min from 3–35% B from the nano-LC system at a constant flow rate of 300 nL per min. The analytical column was regenerated at 90% B for 10 minutes and re-equilibrated at 5% B for 15 minutes. The eluted peptides were analyzed via LC‒MS/MS. The MS acquisition time was set from gradient time zero to 120 min, and the MS1 spectra were collected in the mass range of 400 to 1,500 m/z with 250 ms in “high sensitivity” mode. Further fragmentation of each MS1 spectrum occurred, with a maximum of 30 precursors per cycle. The switch criteria used were a charge of 2+ to 5+, a 500 cps intensity threshold, and dynamic exclusion for 15 s.

DIA (SWATH- MS) quantitative proteomic analysis

SWATH-MS data for individual samples were acquired via LC‒MS/MS using the workflow described above. Swath acquisition was carried out in data-independent acquisition (DIA) mode. The MS1 spectra were collected in the 400 to 1,250 m/z mass range in “high sensitivity” mode. The variable Q1 isolation windows are optimized on the basis of the spectral library via the SWATH Acquisition Variable Window Calculator. The collision energy was different for each window. Single injections of biological triplicates were performed.

Protein identification and quantification

The protein library was constructed via the Paragon algorithm via ProteinPilot software version 5.0 (SCIEX, USA) and searched against the Mus musculus UniProt database (UniProt release 2024_03) downloaded on 19 June 2024. The search parameters included trypsin as the digestion enzyme, iodoacetamide (IAA) for cysteine alkylation, biological modification, and a threshold of [Unused ProtScore (Conf)] ≥ 0.05 with a 1% false discovery rate (FDR) with ≥ 10 peptides/protein as the filter. This was followed by reviewing the protein molecule markers via PeakView software version 2 (SCIEX, USA). The processing parameters were set as a peptide confidence threshold of 99% and an FDR > 1%. All group files were extracted using an MS tolerance of 25 ppm and an MS/MS tolerance of 50 ppm. The following parameters were considered: number of peptides: 10: number of transitions: 6: and shared peptides excluded for SWATH analysis.

The resulting raw files (Supplementary File 1) were imported into Perseus version 2.1.1.0 software (https://maxquant.net/perseus). ²⁹ The data were log₂-transformed and filtered such that 70% of the values were valid. Missing values were imputed by creating a normal distribution of random numbers with a width of 0.3 relative to the standard deviation of the measured values and a 1.8 standard derivation downshift of the mean. Data normalization was performed by subtracting the median. Significant differences in global protein abundance were assessed via Student’s t test, with a permutation-based FDR of 5% and an S0 value of 0.1. The cutoff criteria for up- or downregulated proteins were set at a log₂(fold change) >1 or <-1, respectively. According to Benjamini and Hochberg, statistically significant changes were regarded as adjusted p values < 0.05 or log₁₀(FDR) >1.3.

Bioinformatics for functional and pathway annotation analysis

Gene Ontology (GO) annotations describing differentially abundant proteins were retrieved from bioinformatic tools, including DAVID Bioinformatics (https://davidbioinformatics.nih.gov)^30,31 or ShinyGO 0.8 (https://bioinformatics.sdstate.edu/go80). ³² For all enrichment analyses, the background was set to the default reference Mus musculus proteome provided by the tools. Term enrichments were considered significant when the FDR was < 0.05. All the databases used Mus musculus as a reference organism. A common and unique list of proteins was identified via InteractiVenn (https://www.interactivenn.net). ³³ Protein‒protein interaction (PPI) analysis was conducted via STRING version 12.0 (https://string-db.org). ³⁴

Western blot analysis

Total proteins were extracted with RIPA lysis buffer (cat. no. 89900, Thermo Scientific, USA) containing 10% (v/v) Triton X-100. The concentration of the isolated proteins was measured via a Pierce BCA protein assay kit (cat. no. 23225, Thermo Scientific, USA). An identical concentration of total extracted protein was prepared in 1X LaemmLi SDS sample buffer (Thermo Scientific, USA), heated at 95°C for 15 min, separated by SDS‒PAGE, and transferred to a PVDF membrane (1620177, Bio-Rad Laboratories, USA). The completeness of the blotting process was confirmed by Ponceaus S staining.The membranes were blocked in 2.5 % nonfat dry milk in TBST for 30 min, washed once with 1x PBS, and incubated with primary antibodies against an SDHB (cat.no. 92649, Cell Signaling Technology, USA), or GAPDH (cat.no. ABS16, Merck Millipore, USA). HRP-conjugated secondary antibodies for anti-rabbit IgG (cat. no.7074, Cell Signaling, USA) were used for detection with enhanced chemiluminescence (ECL) reagents (GE Healthcare). The signals were imaged with a Fusion FX chemiluminescence system (Vilber, France). Densitometric quantification of SDHB was normalized to GAPDH, and relative protein expression levels were expressed as fold changes compared with PRO (proliferating) cells. Uncropped images are provided in Supplementary File 7.

Statistical analysis

Statistical analysis was performed, and graphs were created via GraphPad Prism version 8.0.1. Statistical significance was determined via one-way ANOVA for multiple comparisons. Statistical significance was interpreted at a p value < 0.05. All experiments were performed with at least three biological replicates.

CAD cells exhibited morphological differentiation under two validated protocols

CAD cells are CNS-originating neuronal cell lines and are immortalized (proliferating, PRO) when maintained in high-serum-containing media for routine maintenance. The cells, however, exhibited differentiation characteristics under two validated protocols.^1,2 The conventional method for inducing neuronal differentiation involves the use of serum-free media (SFs). Without mitogenic factors, the cells undergo neurite outgrowth and stop proliferating. However, cells ultimately die due to a lack of surviving nutrients primed by serum (Supplementary Fig. 1). Our previous findings demonstrated that dexamethasone dissolved in 1% serum-supplemented medium (DEX) is an alternative differentiating method that promotes morphological differentiation and prolongs differentiated cell survival. Proliferating (PRO) cells were polygonal in shape, while both SF and DEX clearly promoted neurite elongation (Figure 1(a)).OPEN IN VIEWER

As presented in Figure 1(b), PRO cells actively proliferated as the cell number progressively increased. The cells were slowly proliferated upon differentiation, suggesting that differentiating cells retain a low proliferative state, which is consistent with our previous observations. ² Immunocytochemical staining for Ki67, a nuclear marker of cell proliferation that is absent in G0-arrested cells, revealed a significant reduction in Ki67-positive cells in both differentiated cell populations, suggesting a transition toward a post-mitotic state (Figure 1(c)–(d)). Furthermore, to demonstrate the molecular mechanism underlying growth arrest, we observed the expression of TP53, a tumor suppressor gene that controls many cell cycle regulators and promotes growth arrest or death in response to cellular stress.^35–37 The level of TP53 mRNA was significantly increased in the differentiated cells. SF induced a two-fold higher TP53 mRNA expression level compared to DEX, which may contribute to the pronounced cell death observed in SF-differentiated cells (Figure 1(e)).

Even if morphological differentiation is evident, the molecular profile of the protein expression of these cells is unknown. To investigate the signature profile of protein expression in differentiated and proliferating cells, we applied a label-free quantitative mass spectrometry (MS)-based proteomic approach in SWATH acquisition mode. The workflow of the proteomic study and subsequent functional enrichment analysis is illustrated in Figure 2. Following cell plating, CAD cells were induced to differentiate using either DEX or SF for five days. An undifferentiated group (PRO) was also included. After the treatment period, cell lysates were collected, and proteins were extracted for downstream proteomic analysis. The primary objectives of the analysis were to: (1) identify DEPs between the experimental groups, (2) determine commonly expressed proteins shared between the two differentiation conditions, and (3) perform functional enrichment and pathway annotation to interpret the biological significance of the identified protein sets.OPEN IN VIEWER

SWATH -based proteomics identified various proteins associated with the neuronal differentiation and phenotypes of CAD cells

Principal component analysis (PCA) and unsupervised hierarchical clustering based on global protein expression profiles revealed distinct proteomic patterns among the experimental conditions (Supplementary Fig. 2b-c). Although biological replicates within each group were not always positioned immediately adjacent to one another in the dendrogram, samples belonging to the same treatment condition displayed comparable overall expression profiles. On the protein axis, the heatmap revealed distinct co-expression modules, representing groups of proteins with coordinated expression patterns across the sample set. These structured patterns suggest that the dataset effectively captures underlying biological signals rather than random variability, supporting its suitability for downstream analyses. Therefore, downstream differential expression and pathway enrichment analyses were employed to identify biologically meaningful changes associated with neuronal differentiation.

A total of 1,114 proteins were identified. These proteins were directly subjected to pathway annotation analysis via DAVID bioinformatics to identify their functions (Supplementary File 2). We identified proteins involved in neuronal differentiation processes (Figure 3(a)). GO terms associated with the cell cycle, DNA replication, regulation of the cell cycle, cell division, and various terms for neuronal differentiation, which include microtubule cytoskeleton organization, regulation of cell shape, axonogenesis, axon guidance, positive regulation of cell fate, regulation of stem cell differentiation, synaptic vesicle endocytosis, dendrite morphogenesis, and anterograde axonal protein transport, were identified. Various cellular stress responses involved in the DNA damage response, DNA repair, the response to oxidative stress, cellular senescence, and the negative regulation of the apoptotic process have also been shown. In addition, some proteins play roles in the response to steroid hormones and nuclear receptor-mediated glucocorticoid signaling pathways. These pathways indicate cellular and biological alterations that might underlie the neuronal differentiation of CAD cells.OPEN IN VIEWER

The GO terms of cellular component (CC) revealed that the identified proteins were localized mainly in the cytoplasm, followed by the nucleus, cytosol, and mitochondria (Figure 3(b)). Importantly, proteins categorized in the neuron-related compartments were also identified, including those in the synapse, postsynapse, presynapse, glutamatergic synapse, neuronal cell body, axon, dendrites, dendrite spine, myelin sheath, growth cone, filopodium, and other types of cytoskeletons. These data elucidate the neuronal phenotypic characteristics of CAD cells.

Quantitative analysis identified DEPs in dexamethasone-differentiated cells and highlighted alterations in mitochondrial metabolism

To identify the proteins significantly expressed between each comparative group, we set stringent significance thresholds of a P value < 0.05 and an absolute fold change of 2 or greater. These cutoff criteria identified 65 (DEX vs PRO), 25 (DEX vs SF), and 63 (SF vs PRO) proteins as DEPs. The lists of these proteins are provided in Supplementary File 3. As shown in Figure 4(a), 49 and 16 proteins were upregulated and downregulated, respectively, in DEX compared with PRO. In addition, 14 (upregulated) and 11 (downregulated) proteins were quantified as DEPs between DEX and SF (Figure 4(b)). The five proteins with the highest fold changes were highlighted, with succinate dehydrogenase subunit B (SDHB) emerging as the most prominently upregulated protein in dexamethasone-differentiated cells relative to the other conditions.OPEN IN VIEWER

We identified associated functional pathways (BP, MF, CC) of the upregulated proteins (49) in DEX vs PRO by subjecting the list to ShinyGO. The complete list of outputs is available in Supplementary File 4. As shown in Figure 5(a), these proteins were annotated with various BP terms related to mitochondrial metabolic processes, including cellular respiration, the electron transport chain, and oxidative phosphorylation (OXPHOS). The key MF terms enriched in the DEX group included increased oxidoreductase and electron transfer activity (Figure 5(b)). Moreover, network visualization of CC terms (Figure 5(c)) revealed that the upregulated proteins were localized mainly in the mitochondria. These data suggest that mitochondrial metabolism is altered upon DEX differentiation in CAD cells.OPEN IN VIEWER

Additionally, pathway enrichment analyses of DEPs for the DEX vs SF and SF vs PRO comparisons were performed and are presented in Supplementary Figure 3 and 4, respectively. Upregulated DEPs in the DEX vs SF comparison were associated with succinate dehydrogenase activity and the respiratory chain complex. In the SF vs PRO comparison, upregulated DEPs were enriched in oxidative phosphorylation and were predominantly localized to mitochondria, whereas downregulated DEPs were associated with translation.

Upregulated proteins in DEX that play roles in oxidative phosphorylation and pathways of mitochondrial metabolism showed the strong protein-protein interaction

Furthermore, we identified proteins associated with OXPHOS and mitochondrial metabolism. Their abundance in each group is illustrated in Figure 6(a) and Supplementary File 5. These proteins include succinate dehydrogenase (Q9CQA3, Sdhb), malate dehydrogenase (P08249, Mdh2), aconitate hydratase (Q99KI0, Aco2), cytochrome c oxidate subunit 2 (P00405, mt-Co2), NADH dehydrogenase (Q9CQH3, Ndufb5), electron transfer flavoprotein subunit alpha (Q99LC5, Etfa), dihydrolipoyl dehydrogenase (O08749, Dld), trifunctional enzyme subunit alpha (Q8BMS1, Hadha), very-long-chain enoyl-CoA reductase (Q9CY27, Tecr), phosphoenolpyruvate carboxykinase (GTP) (Q8BH04, Pck2), prostaglandin E synthase 3 (Q9R0Q7, Ptges3), delta(3,5)-delta(2,4)-dienoyl-CoA isomerase (O35459, Ech1), poly [ADP-ribose] polymerase 1 (P11103, Parp1), and mitochondrial fission 1 protein (Q9CQ92, Fis1). In addition to parp1, all proteins are located in the mitochondrion. Protein‒protein interaction (PPI) analysis revealed strong connections among these proteins, especially sdhb, mt-Co2, and Ndufb5, which are key DEPs that play roles in OXPHOS (Figure 6(b)).OPEN IN VIEWER

Validation of SDHB, a key DEPs associated with OXPHOS, was performed using immunoblotting (Figure 7(a)–(b)). The results demonstrated a significant upregulation of SDHB in DEX-differentiated cells, confirming the activation of oxidative phosphorylation during dexamethasone-induced neuronal differentiation. In contrast, SDHB expression was reduced in serum-free differentiated cells, which may be attributable to pronounced cell death observed after five days of differentiation under serum-free conditions (SF). To minimize potential bias arising from differential expression of housekeeping protein, particularly under SF-induced cell death, the target protein was normalized to total protein rather than a single loading control (Supplementary Fig. 5). This total protein normalization approach yielded a similar expression pattern of SDHB across experimental groups.OPEN IN VIEWER

Proteins commonly expressed via the two neuronal differentiating methods play roles in mitochondrial metabolism-associated pathways

Next, we aimed to identify the list of proteins commonly expressed in the two validated differentiation protocols. The list of significant DEPs from DEX and SF was input into InteractiVenn. There are 15 upregulated and 2 downregulated proteins identified as overlapping sets of proteins in the differentiation protocols. The list of commonly expressed proteins is presented in Figure 8(a). The functional annotation (BP, MF, CC) of these proteins was performed, and the results are provided in Supplementary File 6. Interestingly, annotation of the typical upregulated genes revealed that both methods resulted in the upregulation of various BP pathways, which were further categorized into two clusters, first associated with mitochondrial metabolism, including the respiratory electron transport chain, and another cluster involving DNA repair mechanisms, such as nucleotide-excision repair and double-strand break repair (Figure 8(b)). Moreover, the results of the PPI analysis of 15 commonly upregulated proteins are shown in Figure 8(c). The data revealed the interactions of the 6 proteins. Among these, 3 proteins were associated with the electron transport chain, as annotated from UniProt keywords (KW-0249).OPEN IN VIEWER

Overall, these data emphasize that mitochondrial metabolism is tightly regulated during CAD cell differentiation, in which a set of proteins functioning in the electron transport chain, OXPHOS, and DNA repair mechanisms are upregulated compared with those in the proliferating stage.