Immortalized Canine Adipose-Derived Mesenchymal Stem Cells as a Novel Candidate Cell Source for Mesenchymal Stem Cell Therapy

Mesenchymal stem cells are expected to be a cell source for stem cell therapy of various diseases in veterinary medicine. However, donor-dependent cell heterogenicity has been a cause of inconsistent therapeutic efficiency. Therefore, we established immortalized cells from canine adipose tissue-derived mesenchymal stem cells (ADSCs) to minimize cellular heterogeneity by reducing the number of donors, evaluated their properties, and compared them to the primary cells with RNA-sequencing. Immortalized canine ADSCs were established by transduction with combinations of the R24C mutation of human cyclin-dependent kinase 4 (CDKR24C), canine cyclin D1, and canine TERT. The ADSCs transduced with CDK4R24C, cyclin D1, and TERT (ADSC-K4DT) or with CDK4R24C and cyclin D1 (ADSC-K4D) showed a dramatic increase in proliferation (population doubling level > 100) without cellular senescence compared to the primary ADSCs. The cell surface markers, except for CD90 of the ADSC-K4DT and ADSC-K4D cells, were similar to those of the primary ADSCs. The ADSC-K4DT and ADSC-K4D cells maintained their trilineage differentiation capacity and chromosome condition, and did not have a tumorigenic development. The ability to inhibit lymphocyte proliferation by the ADSC-K4D cells was enhanced compared with the primary ADSCs and ADSC-K4DT cells. The pathway analysis based on RNA-sequencing revealed changes in the pathways mainly related to the cell cycle and telomerase. The ADSC-K4DT and ADSC-K4D cells had decreased CD90 expression, but there were no obvious defects associated with the decreased CD90 expression in this study. Our results suggest that ADSC-K4DT and ADSC-K4D cells are a potential novel cell source for mesenchymal stem cell therapy.


Introduction
Mesenchymal stem cells (MSCs) can be isolated from many different tissues, including bone marrow, adipose tissues, dental pulp, and umbilical cord [1][2][3]. Adipose tissuederived MSCs (ADSCs) have been investigated extensively for the treatment of various diseases because of the minimal invasiveness of their harvesting and high cell-proliferative capacity compared with bone marrow-derived MSCs in veterinary medicine [4][5][6][7][8][9]. The clinically expected functions of ADSCs are anti-inflammatory and immunomodulatory effects through secreted factors, including exosomes, miRNAs, cytokines, chemokines, and growth factors [10][11][12][13][14]. However, the clinical efficacy in multiple trials of ADSCs has been variable with results that are not consistent with pre-clinical findings in vitro [15,16]. A reason for this may be the heterogeneity of ADSCs [17][18][19]. The source of ADSCs is a major variable affecting their therapeutic efficacy. Therefore, donor-to-donor variability of the phenotype and growth kinetics causes significant inter-individual heterogeneity in the secreted factors of ADSCs. This potentially results in inconsistent outcomes in clinical trials and prevents the practical application of stem cell therapies [16,19]. A solution to avoid variations in ADSCs and stabilize the therapeutic effects is to restrict donors, but this is impossible because of the limited proliferative capacity of primary ADSCs. Therefore, in this study, we attempted to immortalize canine ADSCs.
There are several approaches to establish immortalized cell lines [20]. The classical immortalization method is using viral vectors to introduce ectopic expression of telomerase reverse transcriptase (TERT), human papillomavirus (HPV)-E6/E7 oncogenes, and simian virus 40 (SV40) large T antigen. HPV-E6/E7 and SV40 T antigen bind to negative cell cycle regulators, such as p53 and the retinoblastoma gene product, which act as tumor suppressor proteins in the genome [21]. Moreover, enhanced telomerase activity is essential for the immortalization of cells [22]. Therefore, combinations of human TERT (hTERT) and HPV-E6/E7 or SV40 T antigen expression have been reported to efficiently immortalize human primary cells [23]. However, the overexpression of these oncogenic proteins raises concerns about inducing genomic instability and affecting the properties of the original primary cells. To overcome these concerns, a novel immortalization method was developed by expression of R24C mutant cyclin-dependent kinase 4 (CDK4R24C), cyclin D1 (CCND1), and hTERT [24]. This method efficiently establishes immortalized human myogenic cell lines that maintain their original phenotype. Moreover, apart from humans, immortalization can be effectively induced in cells of various species [25][26][27]. Therefore, in this study, we aimed to establish immortalized canine ADSCs transduced with combinations of human CDK4R24C, canine CCND1, and canine TERT, and analyze the biological characteristics of these cells including whether they maintained MSC properties.

Enhanced Proliferation of ADSC-K4DT and ADSC-K4D Cells
We performed sequential passaging to determine the cell proliferation ability (Figure 1). Although the primary ADSCs had almost ceased proliferation at approximately population doubling level (PDL) 15, the ADSC-K4DT and ADSC-K4D cells continued cell proliferation to >PDL 100 during the 3 months of observation. The proliferation rate of the ADSC-K4D cells was higher than that of the ADSC-K4DT cells. However, similar to the primary ADSCs, the ADSC-TERT cells did not maintain proliferation and had almost ceased proliferation at PDL 15.

CD90 Expression Is Altered in ADSC-K4D and ADSC-K4DT Cells
The primary ADSCs expressed the MSC markers CD29, CD44, and CD90, and very few expressed CD34, CD45, and HLA-DR. Even after numerous cell divisions, the expression pattern of the cell surface markers, except for CD90 in the ADSC-K4DT and ADSC-K4D cells, was similar to the primary ADSCs ( Figure 3). As shown in Figure 4 and Table 1, CD90 expression in the ADSC-K4D cells was consistently slightly lower, whereas that in the ADSC-K4DT cells was decreased markedly as division progressed compared with that in the primary ADSCs.

Maintenance of the Trilineage Differentiated Ability of ADSC-K4DT and ADSC-K4D Cells
We determined whether the ADSCs maintained their trilineage differentiation ability after transduction with canine TERT, canine CCND1, or human CDK4R24C. As shown in Figure 5, the ADSC-K4DT and ADSC-K4D cells were able to differentiate into adipocytes, osteocytes, and chondrocytes similarly to the primary ADSCs.

Lack of Cellular Senescence in ADSC-K4D and ADSC-K4DT Cells
The primary ADSCs and ADSC-TERT cells showed enlarged cytoplasm at approximately PDL 15, but the ADSC-K4DT and ADSC-K4D cells showed no morphological changes ( Figure 6). Most of the primary ADSCs and ADSC-TERT cells that had almost ceased proliferation at approximately PDL 15 were positive for senescence-associated βgalactosidase (SA-β-gal). However, few cells that had undergone repeated cell division of approximately PDL 100 among the ADSC-K4DT and ADSC-K4D cells were positive for SA-β-gal ( Figure 6).

TERT, CCND1, and CDK4R24C Expression in Primary ADSCs, and ADSC-K4DT and ADSC-K4D Cells
In the primary ADSCs, canine CCND1 expression was observed, but canine TERT was expressed even in the cells in proliferative periods. Amplification products specific to canine TERT, canine CCND1, and human CDK4R24C were detected in the ADSC-K4DT cells ( Figure 7). Canine TERT expression was observed in the ADSC-K4D cells despite not being transduced with canine TERT.

Maintenance of the Chromosome Condition in ADSC-K4DT and ADSC-K4D Cells
We performed karyotype analysis of 50 mitotic primary ADSCs (passage 3), ADSC-K4DT cells (PDL 105), and ADSC-K4D cells (PDL 102). All cells in each cell line had 2 n = 78, indicating that the ADSC-K4DT and ADSC-K4D cells had maintained the original number of chromosomes ( Figure 8). Next, we performed G-banding analysis of 20 mitotic cells, which allowed for the identification of individual chromosomes by the banding patterns. There were no chromosome abnormalities in the ADSC-K4DT or ADSC-K4D cells.

Non-Tumorigenicity of ADSC-K4DT and ADSC-K4D Cells
Tumor development was observed at sites injected with Hela cells (all 15 sites: Figure 9B), whereas no tumor formation sites were observed in the sites injected with primary ADSCs or ADSC-K4DT and ADSC-K4D cells in groups 1-3 after 30 days ( Figure 9A). Furthermore, no tumors were generated at all sites injected with primary ADSCs or ADSC-K4DT and ADSC-K4D cells even after 16 weeks in groups 4 and 5 ( Figure 9C).

ADSC-K4DT and ADSC-K4D Cells Inhibit PBMC Proliferation
The proliferation rate of concanavalin A (ConA)-stimulated peripheral blood mononuclear cells (PBMC) cocultured with primary ADSCs was significantly decreased compared with that of ConA-stimulated PBMCs (stimulated PBMCs: 83.1 ± 2.5%; cocultured with primary ADSCs: 71.3 ± 1.4%). The ADSC-K4DT cells inhibited the ConA-stimulated PBMCs similarly to the primary ADSCs, but the ADSC-K4D cells strongly inhibited the ConA-stimulated PBMCs compared with the primary ADSCs and ADSC-K4DT cells ( Figure 10). The suppressive effect of the ADSC-K4DT and ADSC-K4D cells on the PBMCs remained as cell division progressed.

Differences in Pathways among Primary ADSCs, and ADSC-K4DT and ADSC-K4D Cells
Using differentially expressed genes, we carried out the pathway analysis between groups. The pathway analysis using RaNA-seq data was performed with the KEGG, REACTOME, and WikiPathways databases. By comparing the ADSC-K4DT and ADSC-K4D cells with the primary ADSCs, some pathways related to the cell cycle and telomeres were revealed (Supplementary Tables S2-S4). The cell cycle in the KEGG pathway database contains 133 genes, 113 of which were detected with RNA-seq. The expression levels of these 113 genes are presented as a heatmap, and the differentially expressed genes were mapped to the KEGG cell cycle pathway map ( Figure 11). Cellular senescence in the KEGG pathway database contains 159 genes, 138 of which were detected with RNA-seq. The expression levels of these 138 genes are presented as a heatmap, and the differentially expressed genes were mapped to the KEGG cellular senescence pathway map ( Figure 12).

Discussion
We attempted the immortalization of canine ADSCs by transduction with combinations of human CDKR24C, canine CCND1, and canine TERT genes. After repeated passaging, TERT transduction alone did not immortalize the canine ADSCs, whereas K4DT or K4D transduction successfully immortalized the cells. In a previous report, rat ADSCs were successfully immortalized by transduction with hTERT alone [28], whereas immortalization was not achieved with hTERT alone in human ADSCs but rather by transduction in combination with either SV40 or HPV-E6/E7 [29]. Flow cytometry confirmed that canine TERT was transduced into >90% of the primary ADSCs, and integration of canine TERT into the ADSCs was confirmed with PCR, but cell proliferation was arrested at approximately PDL 15. Conversely, the ADSC-K4DT and ADSC-K4D cells maintained proliferation for PDL >100. In a report of immortalization of human dental pulp stem cells, K4DT cells did not show decreased proliferation, whereas K4D cells had a slower proliferation rate at approximately passage 4 [30]. Additionally, senescence of the K4D cells was indicated by an enlarged cytoplasm, and the cells stained positively for SA-β-gal. However, in the immortalization of Tsushima leopard cat fibroblasts, proliferative capacity was maintained by K4D and K4DT cells [25]. The results of enzymatic activity of telomere elongation led to the conclusion that the combination of K4DT transduction was an effective immortalization method. In our study, the proliferation of the ADSC-K4D cells was faster than that of the ADSC-K4DT cells, and the ADSC-K4D and ADSC-K4DT cells did not show cellular senescence with SA-β-gal staining. Furthermore, canine TERT expression was detected in the ADSC-K4D cells that were not transduced with canine TERT, and the expression levels of canine TERT determined with RNA-seq were similar between the ADSC-K4D and ADSC-K4DT cells.
Several combinations of transduced genes have been reported for ADSC immortalization, namely hTERT and SV40 T antigen, hTERT and HPV-E6/E7, and murine Bmi-1 and hTERT [28,29,31]. Chromosomal aberrations and unbalanced translocations were detected using the combinations of hTERT and SV40 as well as hTERT and HPV-E6/7 [28,29], whereas the karyotype of ADSCs immortalized by transducing with the combination of hTERT and murine Bmi-1 was normal [31]. Our established canine ADSC-K4DT and ADSC-K4D cells showed no change in chromosome number in 50 mitotic cells or chromosome aberrations in 20 mitotic cells. A previous study with detailed chromosome analysis of immortalized human fibroblasts transduced with K4DT showed that the incidence of chromosome abnormalities in the K4DT cells was similar to that in wild-type cells [32]. Therefore, cellular immortalization with K4DT is more advantageous by maintaining the original condition of chromosomes compared with oncogenic immortalization methods. Neither the ADSC-K4DT nor ADSC-K4D cells formed tumors at 16 weeks after subcutaneous injection into nude mice, indicating that these cells can be applied to transplantation therapy in vivo.
Except for CD90, the cell surface marker expression in the ADSC-K4DT and ADSC-K4D cells was similar to that in the primary ADSCs. CD90 expression was decreased as the cells proliferated, especially in the ADSC-K4DT cells. CD90 is highly expressed in MSCs regardless of the tissue source [33]. The detailed function of CD90 in MSCs remains unclear, but several studies have suggested that CD90 functions in MSC self-renewal, differentiation, and immunosuppression [34][35][36]. A study of MSCs isolated from various tissue sources, including adipose tissue with CD90 knockdown by a CD90-targeted small hairpin RNA lentiviral vector, showed that the reduction in CD90 expression did not affect the maintenance of MSC morphology, colony-forming ability, or proliferation [35]. However, CD90 knockdown MSCs had reduced CD44 and CD166 expression that enhanced osteogenic and adipogenic differentiation. Therefore, the authors suggested that a reduction in CD90 expression indicates a shift in the stemness state of MSCs towards a state more susceptible to differentiation [35]. We did not examine the expression of CD166, but highly expressed CD44 was maintained even when CD90 expression was reduced in the ADSC-K4DT and ADSC-K4D cells. There is no consensus on whether the characteristic immunosuppressive function of MSCs is affected by decreased CD90 expression [35,36]. A study cocultured human MSCs and PBMCs stimulated with PHA and then measured CD90 expression [36], which showed a negative correlation between CD90 expression and the lymphoproliferative response to PHA activation. Thus, it was suggested that human MSCs with suppressed CD90 expression had a reduced immunosuppressive effect by affecting T cell proliferation. However, another study also cocultured human CD90 knockdown MSCs and CFSE-labeled PBMCs stimulated with PHA and then analyzed lymphocyte proliferation [35], which showed no differences in lymphocyte proliferation between MSCs cocultured with CD90 knockdown or control MSCs. Thus, a reduction in CD90 expression did not affect the immunosuppressive effect of MSCs on lymphocyte proliferation. In our study, the inhibitory effect of the ADSC-K4DT and ADSC-K4D cells on PBMCs was not reduced compared with that of the primary ADSCs. Moreover, as CD90 expression decreased with increasing PDL, the inhibitory effect remained in the ADSC-K4DT and ADSC-K4D cells.
The pathway analysis revealed many pathway changes, especially in those related to the cell cycle and telomeres in the ADSC-K4DT and ADSC-K4D cells compared with the primary ADSCs. The changes in the differentially expressed genes in the cell cycle and cellular senescence pathways were similar between the ADSC-K4DT and ADSC-K4D cells compared with the primary ADSCs, suggesting that canine TERT transduction did not have a significant effect on cellular immortalization. However, further studies are needed to determine how the differences in gene expression and pathways in ADSC-K4DT and ADSC-K4D cells affect the properties and functions of ADSCs and to investigate the therapeutic effects of ADSC-K4DT and ADSC-K4D cells using in vivo experiments.

Animals
Three beagles (males; mean age: 1.5 years; mean body weight: 10.3 kg) were used to isolate ADSCs and PBMCs. Twenty-five male nude mice (5-6 weeks old) were used for the in vivo tumorigenic assays. The mice were housed in a temperature-and light-controlled room (12-h light/dark cycle) and had free access to water and standard laboratory food.

Isolation and Expansion of Canine Adipose-Derived Mesenchymal Stem Cells
Canine primary ADSCs were isolated and expanded as described previously [37]. In brief, after collecting falciform ligament adipose tissue from the anaesthetized dogs, the cells were digested with collagenase type I (Sigma-Aldrich, St. Louis, MO, USA), filtered through a 100-µm nylon mesh, and cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Capricorn, Hessen, Germany) and a 1% antibiotic-antimycotic solution (Thermo Fisher Scientific, Waltham, MA, USA) at 37 • C in a humidified incubator containing 5% CO 2 . At 80-90% confluence, the cells were detached with a trypsin-EDTA solution (Sigma-Aldrich) and passaged repeatedly.

Population Doubling Analysis
To measure the proliferation rate in the long-term culture, the cells were evaluated by population doubling (PD) with sequential passages. The PD value was calculated with the following formula: PD = log2 (a/b), where "a" is the total number of cells recovered at the end of each passage, and "b" is the number of cells seeded at the start of each passage [30]. The PD level was the sum of the PD values obtained from each passage. The cells were cultured continuously by passaging for approximately 3 months.

Cell Cycle Analysis
The cell cycle analysis of the primary ADSCs (n = 3), ADSC-K4DT (n = 3) cells, and ADSC-K4D (n = 3) cells was performed and repeated in three independent experiments with flow cytometry using Cell Cycle Assay Solution Deep Red (DOJINDO, Kumamoto, Japan) in accordance with the manufacturer's instructions.

Trilineage Differentiation Assay
For adipogenic differentiation, the cells were seeded on 12-well plates (4 × 10 4 cells/well) and cultured in DMEM supplemented with 10% FBS and a 1% antibiotic-antimycotic solution until confluency. Then, the medium was changed to StemPro Adipogenesis Differentiation Kit (Thermo Fisher Scientific). The medium was changed twice weekly. Adipogenesis was analyzed with oil red O staining after 28 days.
For osteogenic differentiation, the cells were seeded on 12-well plates (2 × 10 5 cells/well) and cultured in DMEM supplemented with 10% FBS and a 1% antibiotic-antimycotic solution. The next day, the medium was changed to StemPro Osteogenesis Differentiation Kit (Thermo Fisher Scientific). The medium was changed twice weekly. Osteogenesis was analyzed with Alizarin red staining after 21 days.
For chondrogenic differentiation, a MesenCult-ACF Chondrogenic Differentiation Kit (STEMCELL Technologies, Vancouver, BC, Canada) was used. In brief, 1 × 10 6 cells suspended in 0.5 mL medium were placed in a 15 mL polypropylene tube and centrifuged at 300× g for 5 min. Then, the tube cap was loosened, and the cells were incubated at 37 • C with 5% CO 2 . The medium was changed every 3 days. After 21 days, the cell pellet was fixed in 10% formalin and embedded in paraffin. After fixation, the cell pellet was cut into 6-µm-thick sections and stained with Alcian blue.

Cellular Senescence Staining
The cells were seeded in a six-well plate (8 × 10 4 cells/well). After 24 h, β-galactosidase expression was detected using a SA-β-gal staining Kit (Cell Signaling Technology, Danvers, MA, USA) in accordance with the manufacturer's instructions.

Karyotype Analysis
Karyotype analysis was carried out in the primary ADSCs, and the ADSC-K4DT and ADSC-K4D cells by Nihon Gene Research Laboratories Inc. In brief, the cells were treated with colcemid overnight to increase the number of cells in metaphase. After trypsinization, the cells were treated with a hypotonic solution, fixed, stained with a Giemsa solution, and analyzed for detailed chromosomal patterns with G-banding.

In Vivo Tumorigenic Assay
Nude mice were used to assess the tumorigenicity of the primary ADSCs, and the ADSC-K4DT and ADSC-K4D cells. The cells suspended in PBS were implanted in nude mice by subcutaneous injection. HeLa cells were used as a positive control. Twentyfive nude mice were equally divided into five groups. Group 1 was subcutaneously injected with primary ADSCs (left trunk) and Hela cells (right trunk). Group 2 was subcutaneously injected with ADSC-K4DT cells (left) and Hela cells (right). Group 3 was subcutaneously injected with ADSC-K4D cells (left) and Hela cells (right). Group 4 was subcutaneously injected with primary ADSCs (left) and ADSC-K4DT cells (right). Group 5 was subcutaneously injected with ADSC-K4D cells (left) and 0.1 mL PBS (right). A total of 1 × 10 6 cells was implanted at each injection site. Groups 1-3 were euthanized with CO 2 asphyxiation at 30 days after implantation. Groups 4 and 5 were euthanized with CO 2 asphyxiation at 16 weeks after implantation. Tumor nodules were removed and fixed in 4% paraformaldehyde. After fixation and embedding in paraffin, the tissues were cut into 4-µm-thick sections and stained with hematoxylin and eosin.

Lymphocyte Proliferation Assay
The primary ADSCs, and the ADSC-K4DT and ADSC-K4D cells (2 × 10 5 cells/well) were seeded in flat-bottomed 24-well plates in RPMI 1640 medium supplemented with 10% FBS and a 1% antibiotic-antimycotic solution. Canine blood was collected from the jugular vein of healthy beagles into heparinized tubes. The PBMCs were immediately isolated with density-gradient centrifugation using Histopaque-1077 (Sigma-Aldrich) and SepMate-15 (VERITAS, Tokyo, Japan). The PBMCs were prelabeled with a 5-µM CFSE solution using a CFSE Cell Division Tracer Kit (BioLegend) before seeding in accordance with the manufacturer's instructions, and 1 × 10 6 PBMCs were added to the wells with or without primary ADSCs, and ADSC-K4DT and ADSC-K4D cells. To evaluate the lymphocyte proliferation in the presence of primary ADSCs, and ADSC-K4DT and ADSC-K4D cells, the lymphocytes were activated with 5 µg/mL ConA and maintained at 37 • C with 5% CO 2 for 3 days. After coculture, the PBMCs were collected and washed with FACS buffer, and the PBMC proliferation was measured with flow cytometry.

RNA-Sequencing
We performed RNA-seq of total RNA samples isolated from the primary ADSCs, and high PDs (approximately PDL 50) of the ADSC-K4DT and ADSC-K4D cells. The cDNA library construction was carried out with 1 µg total RNA using a NEBNext Ultra II RNA Library Prep Kit from Illumina in accordance with the manufacturer's instructions followed by paired-end sequencing (2 × 150 bp) using a Novaseq6000. For each library, an average of 16-20 million read pairs were generated. Quality control checks of the sequencing raw data were conducted with FastQC ver 0.23.2. Adapter trimming was performed with Trim Galore. The relative expression of the transcripts was quantified in each sample using featureCount (ver 2.0.1). The Fastq files were mapped to the reference genome for Canis lupus familiaris using STAR software ver 2.7.10a. The differentially expressed genes (upregulated or downregulated genes) were determined using edgeR ver 3.22.2 with an adjusted p-value of < 0.05 and fold change of >2 or <0.5. Gene set enrichment analysis was performed with RaNa-seq (https://ranaseq.eu/ accessed on 28 September 2022).

Statistical Analysis
All data are presented as the mean ± standard deviation. The differences among multiple groups were assessed with one-or two-way analysis of variance. The differences were compared using the Tukey-Kramer post-hoc test. p < 0.05 was considered statistically significant. The statistical analyses were performed using Excel 2019 with add-in software Statcel 3.

Conclusions
We established two types of immortalized canine ADSCs, ADSC-K4DT and ADSC-K4D cells, by transduction with combinations of human CDK4R24C, canine CCND1, and canine TERT. These cells had high proliferative and anti-senescence abilities in addition to the fundamental characteristics of primary ADSCs except for CD90 expression. Although further functional analysis is needed, ADSC-K4DT and ADSC-K4D cells have the potential to be a novel cell source for stem cell therapy, which may reduce the number of donors and achieve stable therapeutic effects.

Data Availability Statement:
The data supporting the findings of this study are available from the corresponding author upon reasonable request.