Quantitative Proteomics for the Development and Manufacturing of Human-Induced Pluripotent Stem Cell-Derived Neural Stem Cells Using Data-Independent Acquisition Mass Spectrometry

Abstract

Human-induced pluripotent stem cell (iPSC)-derived neural stem cells (NSCs) have several potential applications in regenerative medicine. A deep understanding of stem cell characteristics is critical for developing appropriate products for use in the clinic. This study aimed to develop approaches for characterizing iPSC-derived NSCs. Data-independent acquisition mass spectrometry (DIA-MS) was used to obtain temporal proteomic profiles of differentiating cells. Principal component analysis of the proteome profiles allowed for the discrimination of cells cultured for different periods. Cells were characterized by Gene Ontology analysis to annotate the upregulated proteins based on their functions. We found that trophoblast glycoprotein (TPBG), a membrane glycoprotein that inhibits the Wnt/β-catenin pathway, was elevated in NSC and that silencing TPBG promoted proliferation rather than neuronal differentiation. Treatment with Wnt/β-catenin pathway activators and inhibitors showed that modulating the Wnt/β-catenin pathway is crucial for differentiation into NSC. These results suggest that the level of TPBG is critical for differentiation into NSC, and TPBG is a potentially critical quality attribute of differentiating cells. In summary, DIA-MS-based proteomics is a promising multi-attribute method for characterizing stem cell-derived products.

Introduction

Induced pluripotent stem cells (iPSCs)¹ can differentiate into tridermic lineages and could be used in regenerative medicine,²⁻⁴ drug discovery research,⁵⁻⁷ and for developing disease models.^4,6,8−10 In the ectoderm lineage, iPSCs can differentiate into astrocytes,¹¹ oligodendrocytes,¹² and neurons, including dopaminergic,¹³ GABAergic,¹⁴ and glutamatergic neurons.¹⁵ Cells that differentiate into neuronal cells are potentially useful for treating spinal cord injury,^16,17 Parkinson’s disease,^9,18 and cerebral infarction.¹⁹ For using these iPS cell-derived products in the clinic, it is essential to evaluate and monitor the manufacturing process and quality of the final products. Currently, the iPS cell-derived products are evaluated based on their cell morphology, physiological function, and expression of molecular markers;²⁰ however, there are several issues regarding the specificity, reproducibility, and operability of these methods.

Cellular proteins contribute to the efficacy and safety of cell-derived products, and some cellular proteins function as critical quality attributes (CQAs) of final products and the intermediates in the manufacturing process.²¹ In addition, the proteins on the cell surface could be used to fractionate the target cells²² and monitor the manufacturing process. However, only a few proteins are currently being used for monitoring the manufacturing process, as the roles of several proteins in neural differentiation remain unclear.²³ A comprehensive knowledge of the temporal changes in the proteins during neuronal differentiation is required to understand the quality characteristics of iPSC-derived cells. Transcriptome analysis is commonly used to investigate such changes;^24,25 however, this approach is limited because the amounts of mRNA and proteins are not always correlated.²⁶ This is attributed to the different degradation rates of mRNA and proteins. This difference is likely even greater during the early stages of differentiation when various proteins appear and disappear within a short period.

Quantitative proteomics is a more direct and useful approach for studying temporal changes in proteins during neuronal differentiation.²⁷ Label-free quantification is generally performed using LC/MS/MS combined with data-dependent acquisition mass spectrometry (DDA-MS).²⁸ However, this approach lacks comprehensiveness as only the abundant peptide ions generated in MS are selected as precursor ions for MS/MS. Recently, the combination of LC/MS/MS and data-independent acquisition mass spectrometry (DIA-MS)²⁹ has gained popularity as a quantitative proteomics approach owing to its excellent quantitative and comprehensive performance. In theory, this approach obtains product ions from all peptide ions generated in MS and then quantifies the peptides based on the intensities of the product ions.³⁰ Furthermore, there has been an increased interest in expanding DIA-MS as a multi-attribute method (MAM) for the characterization of cell therapy products,²¹ such as CAR-T. The MAM approach using DIA-MS is a powerful tool for the characterization of cells in the process of neural differentiation when many new proteins are produced and degraded. However, only a few studies have used DIA-MS to study iPSCs.³¹

Our study aimed to develop an approach to characterize iPSC-derived neural stem cells (NSCs), the final product in the development of cells for regenerative medicine, or a critical intermediate in neuronal cell manufacturing, using DIA-MS as a tool for MAM. Using DDA-MS data from proteins derived from various cells at different stages of differentiation into NSCs, a mass spectral library containing abundant peptide data was constructed to obtain the temporal proteome profiles. The proteomic profiles enabled the evaluation of similarities and differences between cells, the characterization of cells, and the identification of differentiation markers. Combining DIA-MS with immunological, molecular biological, and pharmacological methods could be a useful strategy for identifying CQAs. We demonstrated that DIA-MS is a powerful MAM tool for developing and manufacturing iPSC-derived cells.

Experimental Methods

iPSC Culture and Differentiation into NSC

The human iPS cell lines iPSC 201B7 (HPS0063; RIKEN BRC, Ibaraki, Japan) and iPSC 610B1 (HPS0331; RIKEN BRC) were cultured on Matrigel (Corning, USA)-coated 6-well plates (TPP Techno Plastic Products AG Int., Klettgau, Switzerland) containing an Essential 8 medium (Thermo Fisher Scientific) at 37 °C and under 5% CO₂.

Neural differentiation of iPSCs was based on the method described by Chambers et al.³² Differentiation into NSC was performed on 80% confluent iPSCs cultured in medium A (47.5% DMEM/F12 (FUJIFILM Wako Pure Chemical Inc., Osaka, Japan), 47.5% Neurobasal medium (Thermo Fisher Scientific), 1% N2 supplement (Thermo Fisher Scientific), 2% B-27 supplement (Thermo Fisher Scientific), 1% nonessential amino acids (NEAA; FUJIFILM Wako Pure Chemical), 2 mM L-alanyl-l-glutamine (FUJIFILM Wako Pure Chemical), 100 μM 2-mercaptoethanol (2-ME; FUJIFILM Wako Pure Chemical), 100 nM LDN193189 (FUJIFILM Wako Pure Chemical), and 10 μM SB431542 (FUJIFILM Wako Pure Chemical); Table S1) at 37 °C under 5% CO₂ for 0–7 days. The medium was changed every 24 h. Prior to proteomic analysis, the cells were washed twice with Dulbecco’s phosphate-buffered saline without calcium and magnesium (DPBS (−); Nacalai Tesque Inc., Kyoto, Japan), and then, 1 mL of Accutase (Innovative Cell Technologies, Inc., USA) was added to each well to detach the cells, which were collected by pipetting.

The potential for neuronal differentiation of cells on day 7 of differentiation was confirmed by culturing the cells in medium B (97% KnockOut DMEM/F12 (Thermo Fisher Scientific), 2% StemPro Neural Supplement (Thermo Fisher Scientific), 20 ng/mL fibroblast growth factor 2 (Oriental yeast), 20 ng/mL epidermal growth factor (EGF; FUJIFILM Wako Pure Chemical), and 2 mM GlutaMAX (Thermo Fisher Scientific); Table S1) dispensed in CELLstart (Thermo Fisher Scientific)-coated 6-well plates. The cells were grown at 37 °C under 5% CO₂ to a density of approximately 5 × 10⁴ cells/cm². The culture medium was changed every 24 h. To induce differentiation into neurons, cells on day 7 of differentiation at 5 × 10⁴ cells/cm² were transferred to poly-l-ornithine/laminin-coated plates and cultured at 37 °C under 5% CO₂ for 3 days. Then, 40% confluent NSCs were cultured in medium C (96% Neurobasal medium, 2% NS21,³³ 1% GlutaMAX supplement, 1% CultureOne supplement (Thermo Fisher Scientific), and 200 μM ascorbic acid (FUJIFILM Wako Pure Chemical); Table S1) for 14 days.

The effect of activators/inhibitors on the Wnt/β-catenin signaling pathway was studied by adding 1 μL of the respective small molecular compounds dissolved in DMSO to cells in medium A on days 5 and 6. The compounds added were as follows: DMSO alone (1 μL), 5 or 10 μM CHIR99021 (FUJIFILM Wako Pure Chemical) in DMSO, 5 or 10 μM IWR-1 (FUJIFILM Wako Pure Chemical) in DMSO, and 5 or 10 μM Wnt-C59 (Cayman Chemical, Ann Arbor, Michigan, USA) in DMSO. On day 7, the cells were transferred to medium C and cultured without the compounds for 14 days.

Immunofluorescence

The differentiated cells were washed twice with DPBS (−) and incubated in 99.8% methanol at −20 °C for 10 min. The solution was then replaced with DPBS (−) containing 3% BSA (Nacalai Tesque) and incubated at 25 °C for 1 h. The cells were centrifuged at 700×g for 5 min, the supernatant was removed, and the cells were incubated with the following primary antibodies at 4 °C for 16 h: rBC2LCN-FITC (1/500; 180-0299; FUJIFILM Wako Pure Chemical) for iPSC; anti-Nestin antibody (1/100; MA1-110; Thermo Fisher Scientific), anti-PAX6 antibody (1/150; MA1-109; Thermo Fisher Scientific), and anti-SOX2 antibody (1/100; MA1-014; Thermo Fisher Scientific) for NSCs, and anti-tubulin β3 antibody (Tuj1; 1/100; MMS-435P; BioLegend, California, USA) for neurons. The cells were centrifuged at 700×g for 5 min, the supernatant was removed, and the cells were rinsed three times with DPBS (−) containing 0.1% Tween 20 (Sigma-Aldrich Int., St. Louis, Missouri, USA) (DPBST) and incubated with goat anti-mouse IgG Alexa Fluor 488 (1/2000, A-11001; Thermo Fisher Scientific) at 25 °C for 1 h. The cells were then rinsed three times with DPBST, stained with 5 μg/mL 4′,6-diamidino-2-phenylindole (DAPI; FUJIFILM Wako Pure Chemical), and washed with DPBST. The stained cells in the 24-well plates were examined under an epifluorescence microscope (Eclipse Ts2, Nikon, Tokyo, Japan) at 488 and 385 nm to visualize differentiation-specific markers and nuclei, respectively.

Flow Cytometric Analysis

On day 7 of differentiation (2.0 × 10⁶), cells were washed twice with DPBS (−) and incubated in DPBS (−)/EDTA at 37 °C for 5 min to detach. The cells were then centrifuged at 700×g for 5 min, the supernatant was removed, and the cells were incubated with 4% paraformaldehyde/DPBS (−) at 25 °C for 10 min. The solution was then replaced with 0.1% saponin/DPBS (−), and the mixture was incubated at 25 °C for 5 min. After removing the solution, the cells were incubated in 0.5% BSA/DPBS (−) with anti-TPBG (1/400; 30874; Cell Signaling Technology Inc., Massachusetts, USA) and anti-Nestin (1/200; ab18102; Abcam) antibodies at 4 °C for 16 h. The cells were rinsed twice with DPBST and incubated with goat anti-mouse IgG H&L (FITC) (1/500, ab6785; Abcam) or Goat Anti-Rabbit IgG H&L (APC) (1/1250; ab130805; Abcam) at 25 °C for 1 h. The cells were washed twice with DPBST, and aggregates were removed by filtration through a cell strainer (BD Falcon, New York, USA). The cells were examined using a CytoFLEX-S flow cytometer (Beckman Coulter, California, USA) at 488 and 638 nm at a flow rate of 60 μL/min; 20,000 events were recorded for each sample.

Enzyme Digestion

Cells on days 0, 1, 2, 3, 5, and 7 of differentiation (2.0 × 10⁶) were washed twice with DPBS (−) and then mixed with 100 μL dissolution buffer D consisting of 12 mM sodium deoxycholate, 12 mM N-lauroylsarcosine, 100 mM Tris–HCl pH 9.0, 1% O-GlcNAcase inhibitor, and 1% Halt protease and phosphatase inhibitor cocktail (Nacalai tesque).³⁴ The harvested cells were incubated at 95 °C for 5 min and sonicated to reduce viscosity. Protein concentration was determined using a bicinchoninic acid protein assay,³⁵ and 100 μg of protein was diluted to 2.5 μg/μL with dissolution buffer D. The diluted proteins were reduced by adding 10 mM 2-mercaptoethanol and incubating the mixture at 37 °C for 30 min, followed by another incubation with 20 mM acrylamide at 37 °C in the dark for 60 min. One microgram of mass spectrometry grade lysyl endopeptidase (Lys-C; FUJIFILM Wako Pure Chemical) was added to the solution and incubated at 37 °C for 60 min. Next, 2 μg of trypsin (FUJIFILM Wako Pure Chemical) was added to the solution and incubated at 37 °C for 6 h, followed by the addition of equal amounts of ethyl acetate and 0.1% formic acid (FA; Nacalai Tesque). After stirring, the supernatant was removed by centrifugation at 14,000×g. The sample solution was desalted with an Oasis PriME HLB 1 cc Extraction Cartridge (Waters Corporation, Milford, California, USA) following the manufacturer’s instructions. A portion (20 μg of protein) of the desalted peptide solution was used for library preparation by strong cation polymer (SCX) fractionation, and the remainder was dried using a Speed Vac concentrator (Sakuma Int., Tokyo, Japan). The dried sample was reconstituted with 0.1% FA/3% acetonitrile (ACN) to a concentration of 0.1 μg/μL proteins.

Fractionation of Peptides by SCX

The desalted peptide solutions from each differentiation date for library preparation (n = 8) were pooled together, and the solution was applied to a peptide fraction tip packed with SCX (GL-Tip SCX; GL Sciences, Tokyo, Japan) that had previously been washed with 80% ACN/0.1% trifluoroacetic acid (TFA), followed by Solution A (0.1% TFA). The SCX tip was centrifuged at 3000×g for 5 min, and then, the peptides were fractionated by sequential elution and centrifugation at 3000×g for 5 min with Solution B (80% ACN/0.1% TFA), Solution C (30% ACN/0.5% TFA), Solution D (30%ACN/1%TFA), Solution E (30% ACN/2% TFA), Solution F (30% ACN/3% TFA), Solution G (30% ACN/3% TFA/100 mM ammonium acetate), Solution H (30% ACN/4% TFA/500 mM ammonium acetate), and Solution I (30% ACN/500 mM ammonium acetate). Each fraction was desalted using a peptide fraction tip packed with styrene-divinylbenzene polymer (GL-Tip SDB; GL Sciences). The peptide solution was then diluted with 0.1% FA/5% ACN (50 μL).

LC/MS/MS

Both DDA-MS and DIA-MS were performed using an LC/MS/MS instrument comprising a Q-Exactive (Thermo Fisher Scientific) and an EASY-nLC1000 (Thermo Fisher Scientific) equipped with an Acclaim PepMapC18 (3 μm, 0.075 mm × 10 mm; Thermo Fisher Scientific) and an NTCC-360/75-3-125 (C18, 3 μm, 0.075 × 125 mm; Nikkyo Technos Int., Tokyo, Japan) column. To construct the spectral library by DDA-MS, 0.5 μg of the tryptic digest was injected into LC/MS/MS to acquire as much peptide data as possible. The injection volume of the SCX-fraction was set to 1 μL as the peptide concentration was unknown. In the quantitative analysis by DIA-MS, 0.1 μg of the tryptic digest was used to avoid peptide saturation. Solvent A consisted of 0.1% FA, and solvent B consisted of ACN with 0.1% FA. The separation was performed using a 0–35% linear gradient of solvent B from 0 to 150 min, 35–100% of B from 150 to 151 min, and 100% of B from 151 to 155 min at a flow rate of 300 nL/min. The mass spectrometer was operated in either DDA-MS or DIA-MS mode, using a positive ion mode of data acquisition with a spray voltage of 1800 V and fragmentation by higher energy collision dissociation. Table S2 lists the parameters of DDA-MS and DIA-MS.

Spectral Library Construction

The UniProt human database (July 2018) was searched using Proteome Discoverer 1.4.0 with Sequest HT (Thermo Fisher Scientific). Peptide sequencing was performed on fully trypsin-digested proteins with a maximum of two missed cleavages and peptide lengths of 6-144 using a 6 ppm mass tolerance for precursor ions and a 0.02 Da fragment ion tolerance. Propionamidation of Cys was used for static modification. Dynamic modifications of proteins included the oxidation of methionine and deamidation of asparagine. At least one peptide was used for the identification of proteins. The false discovery rate (FDR) was set at <0.05 to avoid missing possible proteins. Decoy proteins were added to allow protein-level FDR estimation. Decoy sequences were created using inverse sequences.

Quantitative Analysis of Proteins

Proteins were identified and quantified based on raw data acquired in DIA-MS mode using Skyline software (https://skyline.ms/project/home/software/Skyline/begin.view) and UniProt (https://www.uniprot.org/) (human, 2019/2), as reported previously.³⁶ The MS data were filtered according to the following criteria: isotope peaks included count; precursor mass analyzer, Orbitrap; peak, 3; resolving power, 35,000; scan range, >m/z 200; and isotope labeling enrichment, default. The MS/MS data were filtered according to the following criteria: acquisition method, DIA-MS; product mass analyzer, Orbitrap; resolving power: 17500, scan range, >m/z 200; isolation scheme, m/z 500–900; window width, 20; retention time, within 3 min of MS/MS IDs; precursor charges, 2 and 3; ion types, b and y from ion 3 to last ion −1, and peptides with proline at the N-terminus were set as special ions. All matching transitions were selected with an ion-match tolerance of 0.05 m/z. At least one fragment was used for the identification of peptides. Duplication and redundancy were eliminated during protein identification. Duplicates and redundancies were allowed. The search parameter was trypsin [KR/P], with a maximum of two missed cleavages. The oxidation of methionine, deamidation of asparagine, and propionamidation of cysteine were set as variable modifications.

RNAi

Opti-MEM (Thermo Fisher Scientific) and Lipofectamine RNAiMAX (Thermo Fisher Scientific) were mixed in a ratio of 50:3. Opti-MEM and 10 μM siRNA (Silencer Select siRNA TPBG, Silencer Select Negative Control, or BLOCK-iT Alexa Fluor Red Fluorescent Control) were mixed in a ratio of 50:1. Equal amounts of two mixtures were combined to prepare siRNA solutions. iPSCs were seeded in 6- or 24-well plates and cultured in medium A for 5 days. The cells were transfected with siRNA by adding siRNA solution at a final concentration of 10 nM. After 24 h, the fluorescence of BLOCK-iT was confirmed using fluorescence microscopy. The culture medium was changed to medium A without siRNA, and the cells were cultured for 14 days.

Statistical Analysis

iPSCs 201B7 were cultured in medium A for 1, 2, 3, 5, and 7 days (n = 4 each day; four biological repeats). Two technical replicates were generated by performing DIA-MS twice with the samples prepared by the exact condition for quantification. To create the spectral library, 54 samples from 201B7 cell lines on days 0, 1, 2, 3, 5, and 7 of differentiation into NSC were subjected to DDA-MS. Specifically, six samples were tryptic peptides, and 48 were tryptic peptides fractionated by SCX. None of the DDA-MS samples were replicated. All DIA-MS data were used as an input for the Skyline. Relative quantification was performed using MSstat.³⁷ Data were analyzed using a two-tailed Student’s t-test. Data are expressed as mean ± standard error. Volcano plots were created using our script with pandas 1.2.3, NumPy 1.20.1, scikit-learn 0.24.2, seaborn 0.11.1, matplotlib 3.3.4, and bioinfokit 2.0.4. PCA was performed in R (version 4.2.0) using the RGui (64-bit). GO analysis was performed using Metascape (https://metascape.org/gp/index.html#/main/step1). Proteins matching at least one peptide were used for PCA, volcano plots, and GO analysis. Graphs were created using Microsoft Excel.

Data Availability

All proteomic data were deposited into the Japan Proteome Standard Repository/Database (PXD034925). The list of proteins identified is provided in Table S3 (proteins identified by DIA-MS).

Results

Quantitative Proteomics in the Differentiation Process from iPSCs into NSCs Using DIA-MS

Consistent with a previous report, iPSCs 201B7 were differentiated into NSCs by culturing in medium A for 7 days. Separately, iPSCs were cultured for 21 days in a neuronal differentiation medium to confirm that they have neuronal differentiation potential. The proteins extracted from the cells on days 0, 1, 2, 3, 5, and 7 of differentiation (n = 8 each day) were digested with Lys-C and trypsin, and the digests were divided into two portions (Figure 1A). A part of the digest from each differentiation day was combined in a tube, and the six samples were fractionated by SCX into eight fractions each (a total of 48 fractions). DDA-MS provided peptide data for 5923 proteins from unfractionated samples and 8434 proteins from the SCX-fractionated samples. The SCX-fractionation provided more peptide data. Overall, peptide data, including m/z values and retention times for precursor and product ions from 9539 proteins, were used to construct a mass spectrometry library (Figure 1B). The library was used to compare daily changes in protein levels using Skyline (Figure 1C). When DIA-MS was performed on the remaining digests (n = 8, each day), the number of proteins identified per run improved from about 3500 to 5560, with 5544 commonly identified proteins on days 0–7 (Table S3). We used the proteins to characterize cells on each day and extracted marker proteins that could be used to develop manufacturing processes and quality control strategies (Figure 1D).

Figure 1.

Workflow of DIA-MS and the number of proteins identified. (A) Sample preparation. The cells on days 0, 1, 2, 3, 5, and 7 of differentiation were harvested and digested with lysyl endopeptidase/trypsin. (B) Spectral library construction. Portions of the tryptic digests from the cells on days 0, 1, 2, 3, 5, and 7 of differentiation were subjected to DDA-MS, whereas the remaining portion of the digests was fractionated into eight fractions using strong cation exchange polymer, followed by DDA-MS. The DDA-MS data were integrated and stored into the spectral library by Skyline software. (C) Quantitative proteomics. The remaining portions of digests from the cells on days 0, 1, 2, 3, 5, and 7 of differentiation were subjected to DIA-MS, and the spectral library was used for quantifying each protein using Skyline software. (D) Characterization of the cells by proteome profiles obtained by DIA-MS and their use in manufacturing and quality control.

Characterization of Cells Using Proteome Profiles

The cells on days 0, 1, 2, 3, 5, and 7 of differentiation were stained with the undifferentiated marker rBC2LCN (Figure 2A) and the NSC marker Nestin (Figure 2B). rBC2LCN-positive cells were observed on days 0 and 1 of differentiation, whereas the number of Nestin-positive cells increased on day 5 of differentiation. On the seventh day, the NSC status was confirmed by PAX6 and SOX2 staining (Figure S1A). On day 21, the cells were confirmed to be stained with an anti-Tuj1 antibody, a neuronal marker (Figure S1B). These results suggested that undifferentiated iPSCs disappeared within 2 days, and the cells differentiated into NSCs after 5 days.

Figure 2.

Characterization of differentiating cells by immunocytochemistry and DIA-MS. (A) Immunocytochemical analysis on days 0, 1, 2, 3, 5, and 7 for an undifferentiated marker of H type3 (Fucα1-2Galβ1-3GalNAc) with FITC-labeled rBC2LCN (green) on the membrane. Blue indicates 4′,6-diamidino-2-phenylindole (DAPI). (B) Immunocytochemical analysis on days 0, 1, 2, 3, 5, and 7 for NSC marker Nestin (green) in the cytoplasm on days 0, 1, 2, 3, 5, and 7. Blue: DAPI. (C) PCA of the temporal proteome profiles. PCA was created with the average of all replicates. (D) Proteins with a high contribution in the positive direction of PC1. (E) Proteins with a high contribution in the negative direction of PC1. (F) Proteins with a high contribution in the positive direction of PC2. (G) Proteins with a high contribution in the negative direction of PC2.

The relative amounts of 5544 proteins in cells on days 1–7 of differentiation were calculated relative to the amount of proteins in cells on day 0 of differentiation (Table S3). On day 7, NSC marker proteins such as Nestin (four peptides; fold change: 2.3; P-value: 0.06), PAX-6 (one peptide: fold change: 3.4; P-value: 0.04), and SOX2 (eight peptides; fold change: 2.2; P-value: 0.005) were found to be elevated.

PCA was performed to reduce complexity and summarize the proteome profile (Figure 2C, Table S4). Cells on days 1 and 2 were clearly segregated from cells on days 5 and 7 in component 1(PC1), while cells on days 1 and 7 showed similarity in component 2 (PC2). Components 1 and 2 were able to distinguish cells during the differentiation process. Proteins that strongly correlate with PC1 and PC2 were extracted, and their daily fold changes are shown in Figure 2E–H. High-mobility group protein B2 (HMGB2), glutathione S-transferase LANCL1 (LANCL1), alpha-aminoadipic semialdehyde dehydrogenase (ALDH7A1), neutral alpha-glucosidase AB (GANAB), U5 small nuclear ribonucleoprotein 200 kDa helicase (SNRNP200), and RNA-binding protein RO60 (RO60) correlated with PC1 in the positive region (Figure 2D), and six proteins, including four mitochondrial proteins, correlated with PC1 in the negative region (Figure 2E). In PC2, histone acetyltransferase p300 (EP300) and methionine aminopeptidase 1 (METAP1) correlated in the positive region (Figure 2F), whereas uridine phosphorylase 1 (UPP1) and protein sel-1 homolog 2 (SEL1L2) correlated in the negative region (Figure 2G). These proteins could be used to estimate the progression of cell differentiation into NSCs.

Extraction of Differentiation Marker Proteins

Volcano plots show proteins whose abundances changed significantly after differentiation. The proteins that were increased and decreased on each day were identified using cutoff levels of >2-fold change or <0.5-fold change and p < 0.05 (Figure 3A). GO analysis was conducted to annotate the increase and decrease in proteins based on their molecular function to characterize the proteins (Figure 3B and Figure S2). The proteins that were elevated on day 1, day 2, day 5, and day 7 were predicted to be associated with positive regulation of morphogenesis of the epithelium, regulation of epithelial to mesenchymal transition, replication proteins, and forebrain development, respectively (Figure 3B). Some proteins that were decreased on day 7 were annotated as mitochondrial proteins, consistent with the PCA results. GO analysis can be described as the representation of cellular attributes during NSC differentiation.

Figure 3.

Changes in proteome profiles during differentiation of iPSCs into NSCs. (A) Volcano plots of the proteome in the cells on days 1, 2, 3, 5, and 7 of differentiation. The points are relative to day 0. (B) Results of GO analysis with a group of proteins increased during NSC differentiation.

Proteins representing cellular attributes of NSC were extracted from the proteins that were significantly increased on day 7. Proteins in which more than four peptide sequences were identified by DIA-MS and which were more abundant than Nestin (four peptides; fold change: 2.3; P-value: 0.06) were trophoblast glycoprotein (encoded by TPBG; four peptides; fold change: 13.8; P-value: 2.16E-09), choline transporter-like protein 2 (encoded by SLC44A2; six peptides; fold change: 7.3; P-value: 2.43E-08), and mesoderm-specific transcript homolog protein (encoded by MEST; five peptides; fold change: 7.0; P-value: 3.07E-06) (Figure 3A). The FDR values of these proteins were less than 0.01. TPBG was selected as a candidate marker for NSCs as it was elevated on the day 7 volcano plot. This protein is a membrane protein associated with signaling by the Wnt pathway. Although not highly significant, the Wnt signaling pathway was enriched in the GO analysis.

Evaluation of TPBG as a Marker for NSC by Flow Cytometry and RNA Interference

Changes in the TPBG levels were monitored on days 0–7 (Figure 4A). The levels increased rapidly from day 5 to day 7. The increase in TPBG abundance on day 7, as quantified by DIA-MS, was confirmed by flow cytometry after staining for Nestin and TPBG. The percentages of Nestin- and TPBG-positive cells on day 7 of differentiation were 93.2% (n = 3) and 92.6% (n = 3), respectively (Figure 4B, left, middle). The percentage of cells positive for both Nestin and TPBG was 93.3% (n = 3) (Figure 4B, right), suggesting that TPBG was expressed in most Nestin-positive cells.

Figure 4.

Levels of TPBG in the differentiating cells and effect of TPBG gene knockdown on differentiation of iPSCs into NSCs. (A) Daily changes in levels of TPBG as measured by DIA-MS. (B) Flow cytometry on day 7 for Nestin (left), TPBG (middle), and Nestin and TPBG (right). Blue: sample; gray: control. (C) Flow cytometric analysis of TPBG content in control cells (blue) and siRNA-treated cells (red) using an anti-TPBG antibody. Unstained cells are in gray. (D) Timeline of the knockdown experiment. On days 5 and 6 of NSC differentiation, TPBG was knocked down by siRNA, after which the cells were transferred to medium C and cultured for 14 days. Cells on day 21 of differentiation were stained with anti-Tuj1 antibody (green) and DAPI (blue). (E) Tuj1 in the cytoplasm wild-type (left), siRNA-negative control (middle), and siRNA-transfected (right) cells on day 21. Green: Tuj1; blue: 4′,6-diamidino-2-phenylindole (DAPI).

Next, we investigated the significance of TPBG in the differentiation of iPSCs into NSCs and neurons using RNA interference. On day 5 of differentiation, cells were transfected with short interfering RNA (siRNA) designed to silence the TPBG gene. The efficiency of siRNA transfection was verified using fluorescent-labeled control siRNA (BLOCK-iT Alexa Fluor Red Fluorescent Control, Thermo Fisher). Transfection of siRNA was confirmed by red fluorescence in the cytoplasm of NSCs after 24 h, as determined by fluorescence microscopy. The decrease in TPBG at the protein level with no effect on Nestin levels was confirmed by flow cytometry (Figure 4C, left and right). After TPBG-siRNA transfection, the cells were cultured for 2 days in medium A and then further cultured in neuronal differentiation medium C for another 14 days (i.e., a total culture time of 21 days) (Figure 4D). Compared with untransfected and control-siRNA-transfected cells, more Tuj1-positive cells were observed after 16 days of TPBG-siRNA transfection (Figure 4E). These results suggest that the knockdown of TPBG on day 5 promoted proliferation rather than differentiation of NSCs.

Evaluation of the Wnt/β-catenin Signaling Pathway as a Critical Attribute to NSC Differentiation

TPBG is a membrane glycoprotein that inhibits the Wnt/β-catenin pathway by binding to transmembrane-spanning low-density lipoprotein receptor 6 (LRP6), a coreceptor of Frizzled, involved in the canonical Wnt pathway. We confirmed the criticality of the Wnt/β-catenin pathway in the differentiation of iPSC into NSCs by examining the effects of activators and inhibitors of this pathway. We used CHIR99021, an inhibitor of GSK-3β, as an activator, and IWR-1, a tankyrase inhibitor, and Wnt-C59, a PORCN inhibitor, as Wnt/β-catenin pathway inhibitors. These compounds were added to the cells on day 5 of differentiation when DIA-MS detected an increase in the TPBG levels. The cells were cultured with these compounds for another 2 days, followed by culturing for 14 days in neuronal differentiation medium C lacking these reagents. DMSO without the test compounds was used as a control. To confirm that drug reactivity is not specific to the 201B7 cell line, the 610B1 cell line was also used in this experiment. The presence of Tuj1, a neuronal marker, was determined by staining with antibodies.

Axon-like protrusions were observed in 201B7 cells in the control treatment after 13 days and in the 610B1 cells that received the control treatment after 11 days. Tuj1-positive cells were detected in the controls of both cell lines on day 21 (Figure 5A,B). After CHIR99021 (5 μM) treatment, axon-like shapes were detected on days 12 and 9 in 201B7 and 610B1 cells, respectively. We observed more Tuj1-positive cells on day 21 with CHIR99021 treatment compared to that in the controls in 201B7 cells (Figure 5C, left) and 610B1 cells (Figure 5D, left). Moreover, treatment with 10 μM CHIR99021 increased the number of Tuj1-positive cells with axon-like protrusions in 201B7 (Figure 5C, right) and 610B1 cells (Figure 5D, right).

Figure 5.

Effect of activators and inhibitors of the Wnt/β-catenin pathway on differentiation of iPSCs into NSCs. Green: Tuj1; blue: 4′,6-diamidino-2-phenylindole (DAPI). (A) 201B7 cells treated with DMSO. (B) 610B1 cells treated with DMSO. (C) 201B7 cells treated with 5 μM CHIR99021 (left) and 10 μM CHIR99021 (right). (D) 610B1 cells treated with 5 μM CHIR99021 (left) and 10 μM CHIR99021 (right). (E) 201B7 cells treated with 5 μM Wnt-C59 (left) and 10 μM Wnt-C59 (right). (F) 610B1 cells treated with 5 μM Wnt-C59 (left) and 10 μM Wnt-C59 (right). (G) 201B7 cells treated with 5 μM IWR-1 (left) and 10 μM IWR-1 (right). (H) 610B1 cells treated with 5 μM IWR-1 (left) and 10 μM IWR-1 (right).

On day 21 of Wnt-C59 (5 μM) treatment, Tuj1-positive cells undergoing axonogenesis were observed in 201B7 cells (Figure 5E, left), and Tuj1-positive cells with axon-like protrusions were observed in 610B1 cells (Figure 5F, left). In the cells treated with10 μM Wnt-C59, the number of Tuj1-positive 201B7 cells (Figure 5E, right) and Tuj1-positive 610B1 cells (Figure 5E, right; Figure 5F, right) with axon-like protrusions decreased on day 21 compared with the control (Figure 5F, right).

In 201B7 cells treated with 5 μM IWR-1, more cells undergoing axonogenesis were observed than Tuj1-positive cells with axon-like projections on day 21 (Figure 5G, left). Axon-like protrusions were observed on day 12 in IWR-1 (5 μM)-treated 610B1 cells, and immunostained 610B1 cells showed fewer Tuj1-positive cells than in control on day 21 (Figure 5H, left). In the IWR-1 (10 μM)-treated 201B7 and 610B1 cell lines, no or a low number of Tuj1-positive cells (Figure 5G, right; Figure 5H, right) were observed on day 21. These results suggest that Wnt/β-catenin pathway activators promote the abnormal proliferation of NSC and slightly shorten the time needed to differentiate into neurons, and Wnt/β-catenin pathway inhibitors suppress proliferation and prolong the process of differentiation into neurons. Regulation of the Wnt/β-catenin pathway is important in neuronal differentiation, and our results suggest that TPBG expression at appropriate levels is critical for regulating the Wnt/β-catenin pathway in differentiation into neuronal cells.

Discussion

DIA-MS-Based MAM for Characterization of iPSC-Derived Cells

DIA-MS-based proteomics allows quantitative analysis of many different proteins simultaneously and is expected to be used as a MAM for the characterization of materials and quality of cell therapy products, such as CAR-T products. Knowledge of cellular characteristics is critical for process development and establishing quality control strategies to ensure the safety and efficacy of cell therapy products. This study demonstrated that quantitative proteomics using DIA-MS is a promising MAM for the characterization of stem cell-derived products for developing manufacturing processes and quality control.

To enhance the DIA-MS library, we utilized a variety of iPSC-derived cells that were in the process of differentiating into NSCs. Many types of proteins from the cells were fractionated by SCX and then subjected to DDA-MS. This strategy resulted in a library consisting of 9539 proteins. Meanwhile, using DIA-MS, we were able to quantify the changes in the abundance levels of 5544 different proteins in cells undergoing differentiation into NSCs. Proteome profiles were utilized to characterize the cells and develop marker molecules for manufacturing and quality control (Figure 1).

In general, iPSCs and NSCs were roughly distinguished by immunological methods using molecular markers found empirically (Figure 2A,B). We demonstrated that PCA with proteome profiles acquired by DIA-MS enabled us to evaluate the similarities and differences between cells from day 0 to day 7 of differentiation (Figure 2C). These clustering methods could potentially be used for an identification test and a similarity test to evaluate the comparability of cellular products. In the development and control of manufacturing processes, it would be more efficient to use some proteins that correlate with differentiation as a tool for process analytical technology (PAT). The proteins that highly correlated with PC1 in the positive area were HMGB2, LANCL1, ALDH7A1, GANAB, SNRNP200, and RO60, whereas some mitochondrial proteins showed a correlation with PC1 in the negative area (Figure 2D,E). HMGB proteins are essential for DNA unraveling and conformational changes and are associated with the transition of adult neural stem cells from a quiescent state to an activated state.³⁸ Mitochondrial dynamics are also related to neuronal differentiation.³⁹ Proteins correlated with PC2 in the positive area were EP300 and METAP1, whereas SEL1L2 and UPP1 were correlated in the negative area (Figure 2F,G). Current NSC markers such as Nestin appear only on days 5 and 7 of differentiation, whereas the levels of these proteins vary throughout days 0–7 of differentiation. The role of these proteins is still unknown, but future studies could use them as a PAT tool to estimate and monitor the progression of cell differentiation into NSCs.

Identification of CQAs of NSC

Differentiated cells were characterized by GO analysis to annotate the upregulated proteins based on their functions. The cells on days 1, 2, and 7 produced proteins involved in regulating epithelial morphogenesis, regulation of epithelial to mesenchymal transition, and forebrain development, respectively. These classifications were indicative of the differentiation stages of the cells (Figure 3B). Cells on day 7 of differentiation are often defined as NSC, which are expected to be used as therapeutic cells or critical intermediate cells in the manufacturing process for regenerative medicine. NSCs are generally evaluated for the presence of Nestin, PAX6, and SOX2 proteins using antibodies. These markers are intracellular proteins and are unsuitable for monitoring the manufacturing process and fractionation of desired cells. We sought out TPBG as a marker protein on day 7 by visualizing the protein profile using volcano plot analysis (Figure 3A). TPBG inhibits the Wnt/β-catenin pathway by binding to LRP6. The Wnt/β-catenin pathway is associated with stem cell differentiation. The term Wnt/β-catenin was enriched in GO analysis of cells on day 7 (Figure 3B). DIA-MS revealed daily changes in TPBG levels and a 14-fold change on day 7 compared to day 0 (Figure 4A). We verified the increase in TPBG protein expression on day 7 by flow cytometry and confirmed the colocalization of TPBG-positive and Nestin-positive cells (Figure 4B). These results suggest that increased TPBG levels are dependent on differentiation into NSCs. TPBG gene silencing by siRNA remarkably enhanced proliferation and slightly accelerated neuronal differentiation (Figure 4E).

The Wnt/β-catenin pathway is widely known to be involved in various biological processes, including the maintenance of pluripotency and differentiation of stem cells.⁴⁰ The conflicting results of many reports suggest that controlling the Wnt/β-catenin pathway is crucial for stem cell maintenance and differentiation.^41,42 In this study, we confirmed that CHIR99021, a Wnt/β-catenin pathway activator, promoted the proliferation of NSCs, while Wnt/β-catenin pathway inhibitors (Wnt-C59 and IWR-1) delayed neuronal differentiation of iPSCs in a dose-dependent manner (Figure 5A–H). Recently, it has been reported that TPBG is upregulated in the floor plate of the developing human ventral midbrain from 6- to 11-week-old embryos and a subset of some dopaminergic neurons, based on single-cell RNA sequence data sets.⁴³ Additionally, TPBG is involved in the development and function of midbrain dopaminergic neurons⁴⁴ and can be used as a diagnostic marker of Parkinson’s disease.⁴⁵ Our results suggest that TPBG may regulate iPSC differentiation into NSCs by regulating the Wnt/β-catenin pathway via binding to LRP6.^46,47 The TPBG can be a candidate NSC marker and could be developed as a CQA of NSC as an intermediate in the manufacturing process. Further studies are needed to fully assess the impact on safety and efficacy to identify TPBG as a CQA for the cellular product.

Conclusions

In this study, we developed a characterization strategy for iPSC-derived cells using DIA-MS as a tool for MAM. As a pilot study to demonstrate the utility and benefits of this strategy, we evaluated NSC differentiation and demonstrated that DIA-MS provides a comprehensive and quantitative protein list, which is useful in developing manufacturing processes and quality control methods, such as identification tests, similarity assessment, and PAT tools for monitoring. Furthermore, combining DIA-MS with immunological methods such as flow cytometry, molecular biological methods including RNAi, and pharmacological methods using activators and inhibitors is a useful strategy for understanding the molecular mechanisms of neuronal differentiation. As an example, we demonstrated that TPBG could be identified as a CQA candidate. Our strategy could be applied to the development and manufacture of various stem cell-derived products for medical use.

Glossary

Abbreviations

ACN

acetonitrile

CQA

critical quality attributes

DAPI

4′,6-diamidino-2-phenylindole

FA

formic acid

FDR

false discovery rate

GO

Gene Ontology

MAM

multi-attribute method

NSC

neural stem cells

PAT

process analytical technology

PCA

principal component analysis

SCX

strong cation exchange polymer

TFA

trifluoroacetic acid

TPBG

trophoblast glycoprotein

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jproteome.2c00841.

• Solutions used for differentiation experiments and fractionation by SCX (Table S1), parameters in DDA-MS and DIA-MS modes (Table S2), immunocytochemical analysis of cells on day 7 (Figure S1), and results of GO analysis with a group of proteins decreased in NSC differentiation (Figure S2) (PDF)

• Proteins identified by DIA-MS and used for the spectral library (Table S3) (XLSX)

• Proteins used for PCA and their loadings (Table S4) (XLSX)

Author Contributions

N.K. designed the study, project administration, and funding acquisition; T.U., T.K., K.K., and Y.O. performed the experiments; T.U. did the statistical data analysis and visualization; and T.U. and N.K. wrote the manuscript. All the authors commented on the manuscript.

This study was supported by JSPS KAKENHI Grant Number JP21H02617 (to N. K.), a grant-in-aid (IBUNNYAYUGO) provided by Kanagawa Prefectural Government for Integration of Advanced Multidisciplinary Research Activity (to N.K.) and a grant for 2021–2023 Strategic Research Promotion (SK3001) of Yokohama City University (to N.K.).

The authors declare no competing financial interest.