Identification of Proteins Differentially Expressed by Adipose-derived Mesenchymal Stem Cells Isolated from Immunodeficient Mice

Although cell therapy using adipose-derived mesenchymal stem cells (AdMSCs) regulates immunity, the degree to which cell quality and function are affected by differences in immunodeficiency of donors is unknown. We used liquid chromatography tandem-mass spectrometry (LC MS/MS) to identify the proteins expressed by mouse AdMSCs (mAsMSCs) isolated from normal (C57BL/6) mice and mice with severe combined immunodeficiency (SCID). The protein expression profiles of each strain were 98%–100% identical, indicating that the expression levels of major proteins potentially associated with the therapeutic effects of mAdMSCs were highly similar. Further, comparable levels of cell surface markers (CD44, CD90.2) were detected using flow cytometry or LC MS/MS. MYH9, ACTN1, CANX, GPI, TPM1, EPRS, ITGB1, ANXA3, CNN2, MAPK1, PSME2, CTPS1, OTUB1, PSMB6, HMGB1, RPS19, SEC61A1, CTNNB1, GLO1, RPL22, PSMA2, SYNCRIP, PRDX3, SAMHD1, TCAF2, MAPK3, RPS24, and MYO1E, which are associated with immunity, were expressed at higher levels by the SCID mAdMSCs compared with the C57BL/6 mAdMSCs. In contrast, ANXA9, PCBP2, LGALS3, PPP1R14B, and PSMA6, which are also associated with immunity, were more highly expressed by C57BL/6 mAdMSCs than SCID mAdMSCs. These findings implicate these two sets of proteins in the pathogenesis and maintenance of immunodeficiency.


Introduction
Stem cells play an important role in maintaining tissue homeostasis. Mesenchymal stem cells (MSCs) [1] have drawn particular attention for application to regenerative medicine [2].

Characteristics and Quality of C57BL/6 and SCID mAdMSCs
The mAdMSCs were cultured to 80% confluence in DMEM medium containing 10% FBS, with the medium exchanged every two days. Microscopy was performed to confirm that there were no abnormalities in the sizes or shapes of the C57BL/6 mAdMSCs ( Figure 1A, left panel) and SCID mAdMSCs ( Figure 1A, right panel) or in culture conditions. Flow cytometry was performed to detect markers of mAdMSCs (CD44 and CD90.2), hematopoietic stem cells (CD34), and leukocytes (CD45). mAdMSCs expressed CD44 and CD90.2 but not CD34 and CD45 ( Figure 1B, left panels: C57BL/6, right panels: SCID). We induced adipocyte differentiation of the C57BL/6 mAdMSCs ( Figure 1C, upper-left panel) and SCID mAdMSCs ( Figure 1C, upper-middle panel) and osteoblasts ( Figure 1C, upper-right panel). We induced osteocytic differentiation of the C57BL/6 mAdMSCs ( Figure 1C, lower-left panel) and SCID mAdMSCs ( Figure 1C, lower-middle panel) mAdMSCs and the differentiation of osteoblasts ( Figure 1C, lower-right panel) into osteocytes. Mature adipocytes and osteocytes were stained with Oil Red O and Alizarin Red S, respectively.

GO Classification (emPAI > 0)
The proteins were classified according to the subcategories of the Gene Ontology Consortium (available online: http://www.geneontology.org/) database as follows: biological processes, cellular components, and molecular functions. The proteins uniquely expressed by C57BL/6 mAdMSCs and SCID mAdMSCs are shown in Figure 3, left panel. The condition emPAI >0 identified 60 and 53 "viral process-related proteins" expressed by C57BL/6 mAdMSCs and SCID mAdMSCs, respectively. The condition emPAI >0 identified 154 and 137 "immune system process-related proteins" expressed by C57BL/6 mAdMSCs and SCID mAdMSCs, respectively. . GO analysis of proteins expressed by C57BL/6 mAdMSCs and SCID mAdMSCs. The left, center, and right panels indicate the number of proteins expressed by cultured C57BL/6 and SCID mAdMSCs with emPAI >0, >5, and >10, respectively. The proteins classified in Supplementary Tables S3, S5-S8 are listed according to the GO categories as follows: growth, immune system process, biological regulation, metabolic process, response to stimulus, and viral process. Fifty-seven proteins were classified as growth-related (Supplementary Table S3), 165 were classified as immune system process-related (Supplementary Table S5

Viral Process-Related Proteins Expressed by C57BL/6 mAdMSCs and SCID mAdMSCs
Viral process-related proteins were as follows: ). The 28 proteins were expressed by C57BL/6 mAdMSCs and SCID mAdMSCs. The emPAI values were corrected according to the expression of the HKG-encoded proteins. EEF1G, DDX39B, RPS15A, DDX1, NUP155, and SAMHD1 associated with the viral process were highly expressed in SCID mAdMSCs. The other proteins were detected at comparable levels (Supplementary Figure S2A).

Effects of Culture Conditions on Protein Expression by AdMSCs
Immunofluorescence analysis revealed that SCID mAdMSCs expressed CD44 and CD90.2 at normal levels ( Figure 4A), consistent with the results of flow cytometry ( Figure 1B). The proliferation of C57BL/6 and SCID mAdMSCs isolated from female mice was assessed. The amounts of DNA obtained from C57BL/6 mAdMSCs and SCID mAdMSCs 1, 3, and 5 days after cell seeding were equivalent ( Figure 4B). Further, SCID mAdMSCs isolated from female and male mice expressed mRNAs encoding CD44 and CD90.2 at normal levels ( Figure 4C). The influence of the presence or absence of a scaffold material (gelatin-coat) on the function of mAdMSCs was investigated. The C57BL/6 mAdMSCs and SCID mAdMSCs expressed equivalent levels of mRNAs encoding the growth factors HGF, VEGF, and TGFB ( Figure 4D, upper panel) and the antioxidant factors HO-1 and iNOS ( Figure 4D, lower panel) when cultured with or without a gelatin coating.
The influence of the presence or absence of an inflammatory environment on the function of mAdMSCs was also investigated. The levels of growth factors HGF, VEGF, TGFB and antioxidant factors HO-1, iNOS expressed by C57BL/6 mAdMSCs and SCID mAdMSCs were not affected by the addition of LPS to the culture medium.

Correlation Analysis of Proteins Expressed by mAdMSCs
The normalized values of the proteins expressed by C57BL/6 mAdMSCs and SCID mAdMSCs are shown in Supplementary Figure S3. The average value of proteins expressed by C57BL/6 mAdMSCs was 106.0% of those expressed by C57BL/6 SCID mAdMSCs. (HKGs: YWHAZ, GAPDH, ATP5F1B, PGK1, PPIA, and TFRC) (Supplementary Figure S4). The values of ANXA2 and HSPA9 expressed by C57BL/6 mAdMSCs were higher compared with those of SCID mAdMSCs. The values of MYH9 and ACTN1 expressed by SCID mAdMSCs were higher compared with those of C57BL/6 mAdMSCs (Supplementary Figure S3).

Reliability of LC-MS/MS for Comparative Analysis of Protein Expression
Here, we employed LC-MS/MS to conduct a comparative analysis of proteins expressed by mAdMSCs isolated from SCID and normal mice. The ratio of the number of measured peptides to the number of theoretical peptides was linearly related to the logarithm of the protein concentration using the emPAI, which is defined by subtracting 1 from the index of the peptide-number ratio. The larger the emPAI value, the greater the number of proteins. The LC-MS/MS shotgun analysis detected peptides whose origins were determined using an online protein database. Here we combined inguinal-pad fat harvested from three C57BL/6 mice and three SCID mice to produce single cultures of mAdMSCs from each source. The reason for this approach is that unlike the case in humans, the amount of adipose tissue that can be extracted in mice is very small. We recently published a correlation between emPAI values (>0, >1, >2, >3, >5, >10) and the results of protein expression analyses. For example, using emPAI >10, a protein can be detected with high probability even when using n = 1 samples [20]. Thus, we concluded that protein expression data generated using a high emPAI (emPAI >10) is reliable. However, the data of these "n = 1" experiments was strongly affected by individual differences. Thus, the LC-MS/MS protein expression data presented here reveal that the individual differences among SCID mice were consistent with those of normal mice. This is a pilot study of protein expression, and it should be noted therefore that the reliability of the data is influenced by the research method.

T Cell/B Cell-Associated Proteins
Here we show that cultured AdMSCs expressed the B cell-and T cell-associated proteins (MAPK1 and CTPS1). MAPK1, also known as extracellular signal-regulated kinase 2 (ERK2), phosphorylates proteins in the cytoplasm and nucleus and is involved in diverse cell functions, ranging from cell cycle control to higher brain functions. ERK2 promotes T cell activation [29] and the proliferation and survival of CD8-positive T cells [30]. However, the increase in the levels of MAPK1, which acts in the intracellular signaling pathway of AdMSCs, does not mean that AdMSCs directly increase the expression of MAPK1 in B and T cells. For example, human bone marrow-derived MSCs are likely to differentiate into adipocytes when ERK1 and ERK2 are activated. However, in our experiments, induction of the differentiation of AdMSCs into adipocytes did not differ between AdMSCs derived from SCID mice or C57BL/6 mice ( Figure 3C).
In contrast, the loss of function of Cts1 is selectively impaired by T cell proliferation after antigen stimulation. Thus, CTS1 is an important checkpoint in adaptive immunity [31]. Further, a loss-of-function homozygous mutation of Ctps1 inhibits the proliferation of T and B cells in response to activation by an antigen receptor [32]. We were unable to identify studies of the expression or function of CTS1 by MSCs. MSCs and B cells, when cocultured, regulate IgG secretion [33], the inhibition of apoptosis [34], proliferation [35], inhibition of plasma cell differentiation [36], and chemotaxis of B cells [37]. However, the mechanisms of many of these activities, such as those that involve cell adhesion as well as the role of humoral factors, are unknown [38].

Global Comparison of Proteins Expressed by mAdMSCs
After normalizing the levels of proteins expressed by mAdMSCs isolated from the two mouse strains to those encoded by HKGs, we found that the average level of proteins expressed by C57BL/6 mAdMSCs was 1.06-times higher compared with that of the SCID mAdMSCs. Thus, the mean normalized value of total protein levels of the SCID mAdMSCs was 122.0% that of the C57BL/6 mAdMSCs ( Figure 5).
The growth rates of the mADMSCs isolated from the two mouse strains were similar ( Figure 4B) and did not change when the culture dishes were coated with gelatin or when the cultures were treated with LPS. Thus, it is unclear why the average amount of protein expressed in SCID mAdMSCs was approximately 20% higher compared with that of normal mouse mAdMSCs. This may be explained by differences in the genotypes of the two mouse strains.

Proteins Related to the Immune Response
The 165 proteins associated with the GO term "immune system process" of biological processes were investigated (Figure 3). After normalization using HKGs, the emPAI value of the myosin subunit MYH9 was higher for SCID mAdMSCs compared with that of the C57BL/6 mAdMSCs (Supplementary Figure S3). Moreover, the mean level of immunity-associated proteins after normalization was 121.0% that of the C57BL/6 group (Supplementary Figure S3). However, proteins involved in specific immune responses were not detected.

Analysis of Proteins Associated with the Viral Response
Next, we focused on the expression of proteins related to the GO terms viral response (Supplementary Figure S2A) and immune-related proteins (Supplementary Figure S2B). The levels of viral process-related protein factors (SAMHD1) and immune-related protein factors (CTPS1, HMGB1 and SAMHD1) were high in SCID mAdMSCs. SAMHD1 protects cells from viral infection and is involved in the development of cancer and chronic inflammation. Further, the mechanism was recently identified [39]. Further, loss-of-function mutations in Ctps1 are associated with the severe and selective impairment of T cell proliferation after antigen stimulation, indicating that CTPS1 serves as a checkpoint in adaptive immunity [31]. The cytokine HMGB1, which is produced in response to injury, infection, and inflammatory stimuli, is secreted by activated macrophages, mature dendritic cells, and natural killer cells [40].

Summary
The results of studies using mice with severe immune disorders that are deficient in T and B cells, such as SCID mice, can be used as references for evaluating the function of MSCs derived from patients with immunological abnormalities. In the present study, the constellations of proteins and their levels were highly similar between MSCs isolated from normal SCID mice. Moreover, we compared the levels of proteins that mediate diverse biological functions between mAdMSCs generated from three SCID mice compared with those expressed by mAdMSCs generated from three C57BL/6 mice. The levels of representative growth factors and cell surface markers were comparable between SCID and C57BL/6 mAdMSCs. Further, 28 GO-category viral response-related factors and 70 GO-category immune-related factors were expressed by SCID mAdMSCs and C57BL/6 mAdMSCs. However, MYH9, ACTN1, CANX, GPI, TPM1, EPRS, ITGB1, ANXA3, CNN2, MAPK1, PSME2, CTPS1, OTUB1, PSMB6, HMGB1, RPS19, SEC61A1, CTNNB1, GLO1, RPL22, PSMA2, SYNCRIP, PRDX3, SAMHD1, TCAF2, MAPK3, RPS24, and MYO1E, which are associated with immunity, were more highly expressed by SCID mAdMSCs compared with those of C57BL/6 mAdMSCs. In contrast, ANXA9, PCBP2, LGALS3, PPP1R14B, and PSMA6, which are associated with immunity, were more highly expressed in C57BL/6 mAdMSCs than in SCID mAdMSCs. In particular, SCID mAdMSCs expressed the immunity-associated protein SAMHD1 as well as CTPS1, HMGB1, and SAMHD1 at higher levels compared with those of C57BL/6 mAdMSCs. However, no studies are available that identify the roles these proteins play in the therapeutic effects of AdMSCs. The results of the present study provide important clues that may help guide the selection of targets when investigating the underlying mechanisms of the regulation of the immune response by AdMSCs.

Reagents and Materials
Fetal bovine serum (FBS) was obtained from BioWest (Nuaille, France). D-MEM (high glucose) with L-glutamine, phenol red, and sodium pyruvate (DMEM) were obtained from Fujifilm Wako Pure Chemical Corporation (Osaka, Japan). Mouse osteoblasts from cranial bone (Code No. OBC12C) were obtained from Cosmo Bio Co., Ltd. (Tokyo, Japan). Plastic dishes were obtained from TPP (Trasadingen, Switzerland). All other materials used were of the highest commercial grade.

Animal Care
Experimental protocols were conducted in accordance with the guidelines for the care and use of laboratory animals published by the Research Laboratory Center, Faculty of Medicine and the Institute for Animal Experiments, Faculty of Medicine, University of the Ryukyus (Okinawa, Japan). The experimental protocol was approved by the Committee on Animal Experiments of University of the Ryukyus (permit number: R2017081. 08/05/2018). C57BL/6 female mice (8 weeks of age; Japan SLC, Shizuoka, Japan) and CB17/Icr-Prkdc scid /CrlCrlj male and female mice (8 weeks of age; Charles River Laboratories Japan, Inc., Kanagawa, Japan) were maintained under controlled temperature (23 ± 2 • C) and illumination conditions (lights on from 08: 30-20:30). Animals were fed standard rodent chow pellets and had free access to water. All efforts were made to minimize the suffering of the animals.

Isolation of AdMSCs from Mouse Adipose Tissue Via the Inguinal Pad Fat
Adipose tissue was obtained from the inguinal pad fat of three 8-week-old mice (female, n = 3). The method for isolating AdMSCSs from adipose tissue was in accordance with the adipose tissue-derived stem cell product standard document (RMRC-A 01: 2015) of the Ryukyus Regenerative Medicine Research Center. Tissues were stored in cold HBSS and vigorously washed three times in HBSS before starting digestion. Next, the tissues were cut into small fragments with a scalpel for enzymatic digestion (2 mg collagenase type IV/ml; HBSS) in 50-ml tubes rotated at 20 RPM, 37 • C for 60 min) using a shaker (BioShaker BR-42FM; TAITEC, Saitama, Japan). The tubes were then centrifuged at 800× g for 5 min. The SVF [41] containing multiple cell types, including AdMSCs, was collected from the bottom of the tube after centrifugation. The AdMSCs were collected as a cell pellet and then washed with fresh DMEM medium containing 10% FBS to remove the enzyme after digestion. The digested tissue was then incubated in a T25 flask.

Real-Time PCR (qPCR)
Cells (1 × 10 4 ) were seeded onto 24-well plates. Subconfluent cells were dissociated with trypsin -EDTA and recovered. RNA was prepared for a qPCR using an RNeasy Mini Kit and QuantiTect Reverse Transcription Kit according to the manufacturer's instructions (QIAGEN, Hilden, Germany). Reactions were performed using a LightCycler 96 Real-Time PCR system (Roche,

Immunofluorescence Analysis
Immunofluorescence analysis was performed according to a published protocol [44]. The antibodies were the same as those described in Section 4.7.

Cell Proliferation
Cells were seeded onto 24-well plates (1.0 × 10 4 cells/well). On days 1, 3, and 5, the cells were recovered after detachment using trypsin-EDTA, and DNA was extracted using a DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany). The final volume of the DNA extraction solution was 100 µL/well.

LC-MS/MS Analysis of Protein Expression
We used an EzRIPA Lysis Kit (Atto Corporation, Tokyo, Japan) for cell lysis, according to the manufacturer's instructions. The protein concentrations of the lysates solutions were 3105 µg/mL for C57BL/6 mAdMSCs and 6506 µg/mL for SCID mAdMSCs, and 6.0 µg protein was used for sample preparation, and 0.4 µg protein was analyzed using nanoLC-MS/MS. Nanoflow LC-MS/MS and a database search (Mascot analysis) were conducted by Ikuko Sagawa of the Support Center for Advanced Medical Sciences, Tokushima University Graduate School of Biomedical Sciences. Briefly, proteins were reduced using 10 mM DTT in 8 M urea and Tris buffer with 2 mM EDTA (pH 8.5). Proteins were alkylated using 25 mM iodoacetamide in 8 Murea and Tris buffer with 2 mM EDTA (pH 8.5). The alkylation reactions were subsequently diluted in pig-derived trypsin and incubated at 37 • C overnight. Solid-phase extraction was performed using ZipTip µC18 pipette tips (Merck Millipore, Darmstadt, Germany) to concentrate the peptides. Nano LC-MS/MS was performed using an UltiMate 3000 RSLC Nano System (Thermo Fisher Scientific, Tokyo, Japan) The reconstituted peptides were fractionated using an Acclaim PepMap C18 Trap C18 column (75 µm × 15 cm, 2 µm) (Merck Millipore, Darmstadt, Germany). Solvent A contained 0.1% formic acid and Solvent B contained 80% acetonitrile/0.08% formic acid. Peptides were eluted using a 229-min gradient (4% solvent B in solvent A to 90% solvent B in solvent A), 300 mL/min. The ionization settings of the Orbitrap Elite were as follows: Nanoflow-LC ESI, positive; and capillary voltage, 1.7 kV. Tandem mass spectrometry was performed using Proteome Discoverer software (version 1.4, Thermo Fisher Scientific). We did not perform charge state deconvolution or deisotoping.
Database search-Mass spectra were extracted using Proteome Discoverer software (version 1.4). We did not perform charge state deconvolution or deisotoping. The analyses of the MS/MS samples were performed using Mascot software (version 2.5.1; Matrix Science, London, UK), which was configured to search the SwissProt_2018_08 database (version unknown, 558125 entries). Mascot was searched using a 0.60-Da fragment-ion mass tolerance and a 5.0-PPM parent ion tolerance. Deamidation of asparagine and glutamine, and oxidation of methionine and carbamidomethyl of cysteine were specified as variable modifications in Mascot.
Protein identification criteria: Proteins were identified using a previously reported method [50]. Briefly, the relative abundance of proteins was estimated by determining the protein abundance index (PAI) and the emPAI. Visualization and validation of the data were performed using Scaffold software to compare samples and identify their biological relevance. We used Scaffold software to validate peptide and protein identifications. Peptide identifications were accepted if it was established with >83.0% probability that they would achieve a false discovery rat (FDR) of <1.0% according to the FDR algorithm of Scaffold. We accepted protein identifications with >95.0% probability that they contained ≥1 identified peptide. Protein probabilities were determined using the Protein Prophet algorithm [51]. Proteins that contained similar peptides, which could not be differentiated based on MS/MS alone, were grouped to satisfy the principles of parsimony. Proteins that shared significant peptide evidence were grouped into clusters. Proteins were annotated with GO terms (goa_uniprot_all.gaf, downloaded 10/14/2016 [26]. Go analysis was performed using the GO analysis function of Scaffold 4 software with data imported from the external GO Annotation Source database (goa_uniprot_all.gaf [downloaded 2016/10/14]) [26].

Statistical Analysis
Statistical analysis was performed using Student's t-test to compare two samples. Statistical significance was set at P < 0.05 for all tests.

Conclusions
The present study identified proteins differentially expressed by SCID mAdMSCs that may contribute to dysfunctions of the immune system.