**Murine Mesenchymal Stromal Cells Retain Biased Di**ff**erentiation Plasticity Towards Their Tissue of Origin**

#### **Ting Ting Ng 1, Kylie Hin-Man Mak 1, Christian Popp <sup>1</sup> and Ray Kit Ng 1,2,\***


Received: 29 February 2020; Accepted: 18 March 2020; Published: 19 March 2020

**Abstract:** Mesenchymal stromal/stem cells (MSCs) reside in many human tissues and comprise a heterogeneous population of cells with self-renewal and multi-lineage differentiation potential, making them useful in regenerative medicine. It remains inconclusive whether MSCs isolated from different tissue sources exhibit variations in biological features. In this study, we derived MSCs from adipose tissue (AT-MSC) and compact bone (CB-MSC). We found that early passage of MSCs was readily expandable ex vivo, whereas the prolonged culture of MSCs showed alteration of cell morphology to fibroblastoid and reduced proliferation. CB-MSCs and AT-MSCs at passage 3 were CD29+, CD44+, CD105+, CD106+, and Sca-1+; however, passage 7 MSCs showed a reduction of MSC markers, indicating loss of stem cell population after prolonged culturing. Strikingly, CB-MSC was found more efficient at undergoing osteogenic differentiation, while AT-MSC was more efficient to differentiate into adipocytes. The biased differentiation pattern of MSCs from adipogenic or osteogenic tissue source was accompanied by preferential expression of the corresponding lineage marker genes. Interestingly, CB-MSCs treated with DNA demethylation agent 5-azacytidine showed enhanced osteogenic and adipogenic differentiation, whereas the treated AT-MSCs are less competent to differentiate. Our results suggest that the epigenetic state of MSCs is associated with the biased differentiation plasticity towards its tissue of origin, proposing a mechanism related to the retention of epigenetic memory. These findings facilitate the selection of optimal tissue sources of MSCs and the ex vivo expansion period for therapeutic applications.

**Keywords:** mesenchymal stromal cell; differentiation; tissue of origin; prolonged culture; epigenetic memory

#### **1. Introduction**

Mesenchymal stromal cells (MSCs), also referred to mesenchymal stem cells [1], represent a heterogeneous population of cells that can be isolated from a wide range of tissues, including bone marrow, compact bone, placenta and adipose tissue [2–6]. MSC was first isolated from mouse bone marrow as fibroblast colony-forming units, which were distinguished by their ability to adhere to plastic culture dishes [2]. They display fibroblastic morphology and are capable of differentiation to chondrocytes, adipocytes, and osteoblasts in vitro [1,7]. Differentiation to other non-mesodermal cell types, such as neurons, muscles, endothelial cells, and hepatocytes, has also been reported [8–11]. MSCs are intensely studied in clinical research because of their multi-lineage potential and ease of isolation and culture [1,12]. In addition, their ability to evade the host immune system by suppressing T cells, B cells, and natural killer cells [13], and releasing anti-inflammatory proteins [14,15] have

made them an important tool for disease treatment. Clinical trials using MSCs for the treatment of osteoarthritis, degenerative disc disease, ischemic heart disease, and stroke are currently undergoing to explore their therapeutic applications [12].

Harvesting MSCs from the non-bone marrow tissue sources can be done by less invasive methods and the primary isolated MSCs can be expanded ex vivo to yield a larger number. Therefore, these non-bone marrow-derived MSCs are considered as an attractive repertoire for stem cell and regenerative medicine. In light of the broad potentials for therapeutic applications and the variety of sources for MSCs, it has been reported that MSCs, regardless of their tissue of origins, displayed similar characteristics in the differentiation to adipocytes, chondrocytes and osteocytes [16–18]. However, other studies comparing MSCs from human bone marrow, skin and adipose tissues showed considerable differences in their growth rate and differentiation potentials [19,20], supporting the hypothesis of preferential differentiation hierarchies [21,22]. It, therefore, remains inconclusive on the characteristics and the differentiation potentials of MSCs obtained from various tissue sources.

With such variability in mind, a better understanding of the differences between MSCs from different tissue origins can help identify the most suitable cell source for specific clinical purposes. Here we compared and characterized murine MSCs obtained from the adipose tissues (AT-MSC) and compact bone (CB-MSC) using standard isolation methods and expanded ex vivo at early and late cell passages. Both MSCs cultured for an extended period of time showed morphological changes and a decline in cell proliferation. We also demonstrated the tissue origin of MSC is associated with the alterations of cell surface marker patterns and differentiation potential towards osteogenic and adipogenic lineages. Removal of DNA methylation by pharmacological agent can alter the biased differentiation potential of MSCs dependent on the tissue source.

#### **2. Materials and Methods**

#### *2.1. MSC Isolation and Culture*

MSCs were harvested from 8-week old C57BL/6 mice (Laboratory Animal Unit, The University of Hong Kong). Written informed consent to use the animals was approved by the Committee on the Use of Live Animals in Teaching and Research of the University of Hong Kong (Reference no.: 2416-11). Three mice were used for the MSC isolation experiment and a total of five experiments were performed. The male to female ratio was 2:1. Mice were sacrificed by over-dosage of isoflurane inhalation. The MSC isolation procedures were described previously [5,23]. For CB-MSC isolation, muscles from femur, humerus, and tibia were removed. The epiphyseal ends of the bone were cut and discarded. Bone marrow was released by gently crushing the bones in cold PBS with 2% FBS (Gibco, Invitrogen, Grand Island, NY, USA) and 1 mM EDTA (Sigma-Aldrich, St. Louis, MO, USA). The cleaned bone fragments were digested in 0.25% collagenase I (Gibco) with 20% FBS for 5 min at 37 ◦C. The bone fragments were further chopped into 1–2 mm bits and digested for another 45 min at 37 ◦C. CB-MSCs were separated from the bone fragments by filtering the cell suspension through a 70 μm cell strainer (BD Biosciences, Franklin Lakes, NJ, USA). For AT-MSC isolation, adipose tissue dissected from inguinal and subcutaneous sites was digested in 5% collagenase I (Gibco) for 1 h at 37 ◦C. Cells were released from the adipose tissue by centrifuging for 5 min at 500× *g*. The released cells were treated with ammonium-chloride-potassium (ACK) lysing buffer (Sigma-Aldrich) for 3 min at room temperature to lyse red blood cells. The cell suspension was washed twice with α-MEM and tissue debris was removed by filtering through a 70 <sup>μ</sup>m cell strainer. The isolated cell suspension (5 <sup>×</sup> 106) were cultured in α-MEM with 10% fetal bovine serum (FBS) (HyClone, Logan, UT, USA) and 1× penicillin-streptomycin-glutamine (Gibco) on a 100 mm culture dish at 37 ◦C with 5% CO2. MSCs were adhered within 48 h. The culture medium was changed every 2 days. MSCs were passaged in a 1:4 ratio when reaching 80% confluence by Accutase (Gibco) for 5 min at 37 ◦C.

#### *2.2. Cell Proliferation and Immunophenotypic Analysis*

AT-MSCs and CB-MSCs (2 <sup>×</sup> 104 cells) were seeded into 6-well plates on day 0. The total number of expanded cells were counted at day 1, 3 and 5. Cell doubling time (DT) was calculated by the following formula: DT = T × ln2/ln × (Xe/Xb), where T is the incubation time in any units; Xb is the cell number at the beginning of the incubation time; Xe is the cell number at the end of the incubation time.

Cultured cells were stained with FITC-conjugated anti-CD29 or anti-c-kit, PE-conjugated anti-CD44, anti-CD45, anti-CD106 or anti-Sca-1, APC-conjugated anti-CD105, and APC-Cy7-conjugated anti-CD11b (all from Biolegend, San Diego, CA, USA) at a concentration of 0.5 μg/mL for 30 min at 4 ◦C. The corresponding fluorophore-conjugated isotype controls were used for the gating of the positive-stained cells. Immunophenotypic analysis of 5000–10,000 cells of each sample was performed using FACSCanto II flow cytometer (BD Biosciences). The flow cytometry data were analyzed using Flowjo software (Tree Star, Ashland, OR, USA, ver. 10.0.7). Both assays were performed three times with duplicated samples.

#### *2.3. MSC Di*ff*erentiation Assays*

MSCs at passage 3 or 7 (6 <sup>×</sup> 10<sup>4</sup> cells) were seeded into 24-well plates. One group of MSCs was treated with 0.5 μM 5-azacytidine (5-aza, Sigma-Aldrich) for 48 h prior to differentiation. For osteogenic differentiation, MSCs were differentiated in α-MEM with 10% FBS and StemXVivo Mouse/Rat Osteogenic Supplement (R&D Systems, Minneapolis, MN, USA) for 18 days. Differentiated cells were fixed in 4% formaldehyde (Sigma-Aldrich) for 10 min at room temperature and stained with 2% Alizarin Red solution (Chemicon, Merck Millipore, Billerica, MA, USA) for 15 min at room temperature. For adipogenic differentiation, MSCs were differentiated in α-MEM with 10% FBS and StemXVivo Adipogenic Supplement (R&D Systems) for 14 days. Differentiated cells were fixed in 4% formaldehyde (Sigma-Aldrich) for 10 min at room temperature and stained with 0.5% Oil Red O solution (Sigma-Aldrich) for 15 min at room temperature. For chondrogenic differentiation, MSCs were centrifuged for 5 min at 200× *g* in a 1.5 mL tube and differentiated in DMEM /F-12 with 1× Insulin-Transferrin-Selenium (Gibco) and StemXVivo Human/Mouse Chondrogenic Supplement (R&D Systems) for 21 days. Chondrocyte spheroids were fixed in 4% formaldehyde (Sigma-Aldrich) for 1 h at room temperature and stained with Alcian Blue 8GX solution (Sigma-Aldrich) for 30 min at room temperature. MSCs cultured in the differentiation medium without supplements were served as controls. The differentiation assay was performed three times with duplicated samples.

#### *2.4. RNA Extraction and Quantitative RT-PCR (qRT-PCR)*

Total RNA was extracted from the differentiated MSCs using MiniBEST Universal RNA Extraction Kit (Takara, Kusatsu, Japan). Genomic DNA eraser column and DNaseI treatment were used to remove genomic DNA. cDNA was synthesized using PrimeScriptTM RT reagent kit with gDNA Eraser (Takara) according to the manufacturer's protocol. qRT-PCR was performed with the 7900HT Fast Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) using SYBR Premix Ex TaqTM (Takara) with the oligo primers listed in Supplementary Table S1. *Gapdh* and β*-Actin* served as house-keeping genes for normalization of gene expression. All samples were analyzed in triplicate. Three independent experiments were performed and relative gene expression was calculated using 2−ΔΔCT method.

#### *2.5. Statistical Analysis*

A statistically significant difference was calculated by two-tailed unpaired Student's *t*-test.

#### **3. Results**

#### *3.1. Ex Vivo Expansion of MSC Isolated from Compact Bone and Adipose Tissue*

Murine MSCs isolated from the adipose tissue (AT) and compact bone (CB) were expanded ex vivo. Both AT-MSCs and CB-MSCs displayed spindle-like to fibroblastoid cell morphology [4,5,23]. It was observed that cells at passage one (P1) contain a small number of cells with spherical shape, which were presumably dividing cells or non-MSCs (Figure 1a,e). However, continuous passaging of cells to the third passage (P3), which is one week of culture, gradually eliminated the non-MSC populations and enriched for MSCs (Figure 1b,f). From passage three (P3) onwards, cell morphology changed from elongated to fibroblastoid in both cultures (Figure 1b–d,f–h).

**Figure 1.** Cell morphology of AT-MSC and CB-MSC. Morphologies of MSCs at (**a**,**e**) passage 1, (**b**,**f**) passage 3, (**c**,**g**) passage 5, and (**d**,**h**) passage 7 were shown. Cell morphology changed gradually from spindle-like to flat and fibroblastoid with increasing passage number. Representative images were taken at 20× magnification. Scale bars: 100 μm.

Ex vivo culture of AT-MSC and CB-MSC at P3 or P7 for 5 days demonstrated cell number expansion. P3 and P7 AT-MSC showed limited expansion by 2.2- and 1.5-fold, respectively; whereas P3 and P7 CB-MSC were expanded 4.3- and 3.3-fold, respectively (Figure 2a,b). Besides, it was found that the doubling time of CB-MSC was comparable between both passages (2.4 days and 2.8 days for P4 and P7, respectively); however, AT-MSC demonstrated a significant increase in doubling time from P4 (4.3 days) to P7 (8.9 days) (Figure 2c). These results demonstrated different cell proliferation patterns between MSCs isolated from different tissue origins. Nevertheless, prolonged culture of both types of MSCs gradually reduced proliferation rate beyond passage 7.

#### *3.2. Alterations of MSC Immunophenotypes by Prolonged Culture*

Previous studies have shown that prolonged culture of MSC altered their immunophenotypes [24]. This prompt us to examine the expression of a panel of mesenchymal stromal cell surface markers, including CD29, CD44, CD105, CD106, and stem cell antigen-1 (Sca-1) [25–28], in the ex vivo expanded cells. Hematopoietic markers c-kit, CD11b, and CD45 were served as negative markers for the detection of contamination of hematopoietic cells from the MSC isolation procedures [27,29]. c-kit<sup>+</sup> and CD11b<sup>+</sup> populations were generally low in both types of MSCs, particularly for the late passage culture (Figure S1). It was observed that 38.4% of CD45<sup>+</sup> populations were present in P3 CB-MSC, suggesting a low degree of hematopoietic cell contamination from compact bone during MSC isolation.

**Figure 2.** Ex vivo expansion of MSCs at different passage numbers. Cell proliferation assay was performed to determine the growth rate of early (Passage 3) and late passage (Passage 7) of (**a**) AT-MSC and (**b**) CB-MSC. Cells were counted on day 1, 3 and 5 (*n* = 3). (**c**) Doubling times of MSCs were calculated over 5 days of culture. CB-MSCs demonstrated a higher cell proliferation rate than AT-MSCs. The doubling time of AT-MSC was significantly increased at late passage. Experiments were performed with three replicates. Data represent mean ± SD; \**p* < 0.05, \*\* *p* < 0.01 and \*\*\* *p* < 0.001.

Nevertheless, the CD45<sup>+</sup> hematopoietic cells were gradually lost when cells passaging to P7. Both AT-MSCs and CB-MSCs demonstrated high expression of most of the MSC markers at passage 3. It was noted that CD29+, CD44+, and CD106<sup>+</sup> populations showed further increased in passage 7 (Table 1, Figure 3). However, CD105<sup>+</sup> population was reduced significantly at late passage MSCs. While a significant portion of the AT-MSC population retained as CD105<sup>+</sup> (33.6 <sup>±</sup> 4.3%) at P7, the CD105<sup>+</sup> population in CB-MSC reduced drastically from 34.2% at P3 to 7.5% at P7. In contrast, CB-MSC consisted of over 83% Sca-1<sup>+</sup> cells at P3 and P7, whereas the Sca-1<sup>+</sup> population dropped from 98.5% to 26.3% in AT-MSC from P3 to P7. These immunophenotypic results demonstrated the alteration of MSC surface marker pattern during ex vivo culture, suggesting that prolonged culture of MSC is accompanied by the loss of MSC identity.


**Table 1.** Percentage of cell populations in AT-MSC and CB-MSC.

\* *p* < 0.05 and \*\* *p* < 0.01 in the comparison between P3 and P7.

**Figure 3.** Immunophenotypes of MSCs. Cell surface markers for MSCs, CD29, CD44, CD105, CD106, and Sca-1 were used to characterize (**a**) AT-MSC and (**b**) CB-MSC at passage 3 (P3) and 7 (P7), respectively. Representative flow cytometry patterns were shown. Shaded peaks represent antibody-labeled population; blank peaks represented isotype controls.

#### *3.3. Biased Di*ff*erentiation Towards the Tissue Origin*

A defining feature of MSC is their ability to differentiate into multiple mesodermal lineages. To examine the multi-lineage differentiation potentials of MSCs derived from different tissue origins, we induced in vitro differentiation of early and late passage (P3 and P7) AT-MSC and CB-MSC into the osteogenic, adipogenic, and chondrogenic lineages. We observed that both AT-MSCs and CB-MSCs were able to differentiate into the three lineages (Figure 4), indicating that MSCs from adipose tissue or compact bone are multipotent in nature. However, although both types of MSCs were able to form the positive Alcian Blue stained chondrocyte spheroids efficiently (Figure 4i–l), we observed that fewer cells stained positive with Alizarin Red in the AT-MSC sample (Figure 4a–d) and lower number of Oil Red O stained cells from CB-MSC sample (Figure 4e–h). Besides, the P7 MSCs of both types appeared to have weaker positive staining patterns when compared to the early P3 samples. These results suggest that MSCs derived from different tissue origins exhibit differentiation bias and their differentiation capacities reduce after prolonged culture.

To further elucidate the differentiation bias associated with the tissue origin of MSCs, we examined the expression of osteogenic (*Ocn* and *Opn*), adipogenic (*Adipoq* and *Pparg*) and chondrogenic (*Sox9* and *Col2a1*) markers in the differentiated MSC samples. Induction of *Ocn* and *Opn* were high in the osteogenic differentiation of P3 CB-MSC when compared to the P7 CB-MSC (over 4-fold for both genes). Importantly, osteogenic differentiated AT-MSC demonstrated significantly lower expression of these two osteogenic markers, regardless of the length of culture (Figure 5a, Figure S2a).

**Figure 4.** Osteogenic and adipogenic differentiation of MSCs. AT-MSCs and CB-MSCs underwent 18-day of osteogenic, 14-day of adipogenic, or 21-days of chondrogenic differentiation conditions. (**a**–**d**) Alizarin red staining, (**e**–**h**) Oil Red O staining, and (**i**–**l**) Alcian blue staining were used to assess osteogenic, adipogenic and chondrogenic differentiation, respectively. Passage 3 of CB-MSCs displayed stronger staining for Alizarin red; whereas passage 3 of AT-MSCs displayed stronger Oil-Red-O staining. Late passage MSCs showed weaker staining in both lineage differentiations. Chondrogenic differentiation is comparable in both types of MSCs. The white arrows indicate the stained chondrocyte spheroids. Representative images were taken at 20× (**a**–**h**) or 5× (**i**–**l**) magnification. Scale bars: (**a**–**h**) 100 μm; (**i**–**l**) 1 mm.

**Figure 5.** *Cont*.

**Figure 5.** Expression of lineage marker genes in the differentiated MSCs. (**a**) Osteogenic markers, *Ocn* and *Opn*, (**b**) adipogenic markers, *Adipoq* and *Pparg*, and (**c**) chondrogenic markers, *Sox9* and *Col2a1*, were used to determine the multi-lineage differentiation of AT-MSCs and CB-MSCs. MSCs cultured in basic medium without differentiation agents for the same period of time were served as controls. Gene expressions were normalized with housekeeping gene *Gapdh*. Experiments were performed with three replicates. Data represent mean ± SD; \**p* < 0.05 and \*\* *p* < 0.01.

In contrast, AT-MSCs were able to express a high level of *Adipoq* and *Pparg* when compared to the CB-MSCs in adipogenic differentiation (over 4-fold for both genes) (Figure 5b and Figure S2b). The expression of adipogenic markers was less pronounced between different passages of MSCs. We also noticed that the expression of *Sox9* and *Col2a1* was high in both types of differentiated MSCs (Figure 5c and Figure S2c), which implies comparable chondrogenic differentiation efficiency. Taken together, the differential expressions of osteogenic and adipogenic markers are in agreement with the Alizarin red and Oil Red O staining patterns (Figure 4), suggesting that MSCs, although harboring multi-lineage differentiation potential, have a preference to differentiate towards their tissues of origin.

#### *3.4. Inhibition of DNA Methylation Alters MSC Multipotency*

MSCs derived from different tissues could be modulated by the microenvironment which confers a differential epigenetic state associated with stem cell multipotency. To determine whether DNA methylation, a well-known epigenetic modification, is involved in the differentiation bias of CB-MSC and AT-MSC, we treated MSCs with the DNA methylation inhibitor, 5-azacytidine (5-aza), for 48 h prior in vitro differentiation. The 48-h treatment period was chosen based on the rationale that the epigenetic function of 5-aza as a DNA methylation inhibitor is dependent on cell division [30], which takes roughly 2 days (determined by the doubling time in Figure 2c) for both types of MSCs. Both types of MSCs under 5-aza treatment were able to differentiate into osteogenic and adipogenic lineages (Figure 6a). Interestingly, we observed that there were more osteogenic differentiated cells stained with Alizarin red from the 5-aza-treated CB-MSCs when comparing to the untreated sample. However, the number of osteogenic differentiated AT-MSCs remains low by the 5-aza treatment. This observation is in agreement with the qRT-PCR results of the osteogenic marker expression, which showed a significant increase in *Ocn* and *Opn* expression in the 5-aza-treated CB-MSCs (Figure 6b and Figure S3a). Unexpectedly, although the AT-MSCs are more competent to undergo adipogenic differentiation, the 5-aza treatment resulted in a lower number of Oil Red O stained cells (Figure 6a), with a significant decrease in the expression of adipogenic markers *Adipoq* and *Pparg* (Figure 6b and Figure S3b). By contrast, the treated CB-MSCs showed enhanced adipogenic differentiation with a comparable level of adipogenic marker gene expression to the untreated AT-MSCs. These results suggest that inhibition of DNA methylation can restore the biased differentiation capacity of CB-MSC to the adipogenic lineage, whereas AT-MSC loses its multipotency under the same epigenetic condition. It thus implies a differential epigenetic effect of 5-aza on the MSCs derived from different tissue sources.

**Figure 6.** Osteogenic and adipogenic differentiation of the 5-aza-treated MSCs. (**a**) AT-MSCs and CB-MSCs were pre-treated with 5-aza for 48 h prior to osteogenic (Alizarin red staining) or adipogenic differentiation (Oil Red O staining). Representative images were taken at 20×. Scale bars: 100 μm. (**b**) The expression of osteocyte markers (*Ocn* and *Opn*) and adipocyte markers (*Adipoq* and *Pparg*) were determined by qRT-PCR. MSCs cultured in basic medium without differentiation agents for the same period of time were served as controls. Gene expressions were normalized with housekeeping gene *Gapdh*. Experiments were performed with three replicates. Data represent mean ± SD; \* *p* < 0.05 and \*\* *p* < 0.01.

#### **4. Discussion**

In this study, murine MSCs were isolated from compact bone and adipose tissue. Our results showed that prolonged culture of MSC leads to changes in cell morphology and cell surface marker patterns, and cell proliferation rate. Importantly, the tissue origins of MSC have impact on their differentiation capacity towards the corresponding cell lineages, indicating the presence of epigenetic memory in the MSCs. The multipotency of CB-MSC, but not AT-MSC, can be enhanced through inhibition of DNA methylation prior differentiation, which suggests a possible strategy to erase the epigenetic memory in certain tissue-derived MSCs. Although we were using murine MSCs in our study, cross-species comparisons of MSC corroborated that the surface markers [31] and biological functions [32] of MSC are similar, even though not identical, between different species. There are numerous studies of the therapeutic applications of MSCs using mouse models for the investigation of the molecular mechanisms and the safety concerns prior to human clinical trials [33]. With the findings of preferential differentiation of murine MSCs derived from different tissues and the possible manipulation of the epigenetic memory in MSCs, we propose that our findings can be applied to human MSCs for the selection of optimal tissue source and the strategy to enhance the multipotency of human MSCs for therapeutic applications.

The immunophenotype of MSCs is dynamic over the culture period. MSCs isolated from various tissues were reported to express a common set of cell surface markers, such as CD105, CD90, CD73, CD29, and CD44, with a lack of CD34, CD45, CD11b, and major histocompatibility complex (MHC) class II expression [25]. The MSC isolation protocols retained the plastic adherent cells from the tissue, which invariably consist of a heterogeneous cell population including hematopoietic cells and other tissue cells. Nevertheless, the non-MSC populations presumably undergo depletion gradually in the MSC culture condition, resulting in a more homogeneous MSC population. In our study, we observed that nearly all the early passage MSCs from both compact bone and adipose tissue expressed CD29 and Sca-1, but only 34–50% are CD105<sup>+</sup> and 38–60% are CD106+, suggesting the heterogeneity of MSC immunophenotypes. Prolonged culture of MSCs leads to loss of Sca-1 and CD105 expression, which is particularly obvious in the AT-MSCs and CB-MSCs, respectively. Similar observation for the loss of CD15, CD90, and CD309 was also reported in the neoplastic transformation of bone marrow-derived MSCs after numerous passages [34]. The loss of MSC immunophenotype is correlated with the findings of fibroblastoid morphology change, reduced cell proliferation and differentiation potentials of the late passage MSCs, suggesting that the "stemness" of MSCs cannot be maintained by prolonged culture. This is in agreement with the previous study of the prolonged culture of human bone marrow-derived MSCs with a loss of osteogenic potential [35]. However, the loss of stemness could be less prominent in other tissue-derived MSCs, for example, umbilical cord MSCs, which retained comparable growth rate and osteogenic capacity after 16 passages when compared to the freshly isolated one [36]. Although the MSC culture condition used in this study follows a common protocol in the field, the alterations of MSC phenotypes and cell functions owing to the prolonged culture might reflect a suboptimal condition that needs to be further optimized for better maintenance of the stemness of MSCs. In addition, it has been reported that a subpopulation of MSC is CD105 negative [37,38], which varies in the differentiation potentials and modulation of CD4<sup>+</sup> T cell proliferation when comparing to the CD105<sup>+</sup> counterpart. Interestingly, the CD105 expression in the CD105<sup>+</sup> MSCs can be altered by the culture condition, such as passage number, cell density, and medium composition [37,39]. Being a component of the TGF-β receptor, CD105 also serves as a proliferation marker of endothelial cells [40]. As we observed a reduction of MSC proliferation upon prolonged culturing, we speculate that this could be associated with the loss of CD105 expression in the P7 MSCs.

The biased differentiation capacity of MSCs derived from different tissues remains controversial. While several studies have reported that bone marrow-derived MSCs were more prone to osteogenic differentiation [17,41,42] and adipose tissue-derived MSCs showed decreased chondrogenic differentiation capacity [43,44], others demonstrated no significant differences in differentiation potentials of MSCs derived from various tissues [16–18]. In our study, we observed that both types of MSCs showed comparable chondrogenic differentiation, but preferential differentiation of AT-MSCs to adipo-lineage and CB-MSCs to osteo-lineage. It is proposed that the isolation procedures, culture condition, such as with serum or serum-free, and the heterogeneity of MSC populations may account for

the preferential differentiation to certain lineages through enrichment of distinct MSC subpopulations. In addition, MSCs resided in different tissues are subjected to the distinct cellular microenvironments, e.g., signaling molecules, extracellular matrix components, metabolites, etc. These extrinsic factors can induce alterations of the MSC epigenome, leading to potential variations in transcriptomes and biological responses related to stemness [45]. This is indeed supported by the global transcriptomic and proteomic studies which demonstrated substantial differences in the expression of genes or proteins between MSCs derived from bone marrow and adipose tissue [17,20,46,47]. The biased differentiation of CB-MSC and AT-MSC also suggests that the altered epigenome is retained as the "epigenetic memory" of tissue origin. It is reported that the reprogrammed stem cells showed preferential differentiation towards their somatic lineage origins. The epigenetic memory of donor cell origin was found in *Xenopus* nuclear transplanted embryos [48] and early passage of induced pluripotent stem cells (iPSCs) [49,50]. It has been shown that blood-derived iPSCs were preferentially differentiated towards blood lineages and were defective to osteogenic differentiation, whereas bone marrow-derived iPSC demonstrated the opposite [50]. Although the derivation of MSCs does not involve cellular reprogramming, MSCs located at different tissues might be epigenetically reprogrammed by the tissue niche environment. Previous studies reported that the memory status in iPSC can be erased by extensive cell culture passages [49] or by epigenetic modifying agents, such as DNA methylation inhibitor [50]. It remains a challenge to test if the memory status of MSCs can be erased by prolonged culture because late passage MSCs undergo senescence and reduced overall differentiation capacity. Interestingly, our data showed that inhibition of DNA methylation can partially restore the adipogenic differentiation capacity of CB-MSCs, suggesting a possible way to erase the epigenetic memory in MSCs. However, we noticed that AT-MSCs failed to restore osteogenic differentiation after 5-aza treatment, suggesting the involvement of other types of epigenetic modifications, e.g., histone protein methylation or acetylation. Therefore, it is worth evaluating the effects of other epigenetic inhibitors, such as histone deacetylase inhibitors, on the multipotency of MSCs derived from other tissues.

#### **5. Conclusions**

We have demonstrated that early passage MSCs derived from compact bone or adipose tissue are highly proliferative and retain multipotent nature. The tissue origin of the MSCs results in epigenetic memory which implicates a preference for lineage differentiation. A better understanding of the molecular nature of such tissue origin memory can facilitate the choice of optimal sources of MSCs for tissue engineering and regenerative medicine.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4409/9/3/756/s1, Figure S1: Immunophenotypes of hematopoietic markers on MSCs, Figure S2: Expression of lineage marker genes in the differentiated MSCs with reference to β-Actin gene, Figure S3: Expression of lineage marker genes in the differentiated 5-aza-treated MSCs with reference to β-Actin gene, Table S1: Primer sequences for qRT-PCR.

**Author Contributions:** Conceptualization, T.T.N. and R.K.N.; methodology, T.T.N.; validation, T.T.N., K.H.-M.M. and C.P..; formal analysis, T.T.N., R.K.N.; investigation, T.T.N.; data curation, K.H.-M.M., C.P.; writing—original draft preparation, T.T.N.; writing—review and editing, R.K.N.; supervision, R.K.N.; project administration, R.K.N.; funding acquisition, R.K.N. All authors have read and agreed to the published version of the manuscript

**Funding:** This research was funded by Research Grants Council General Research Fund, grant number HKU774712M, and Theme-based Research Scheme, grant number T12-708/12-N. The APC was funded by National Natural Science Foundation of China (NSFC) - Science Fund for Young Scholars, grant number 026-NG KIT\_81200341, and NSFC-Shenzhen Matching Fund 2013, grant number 121-NG KIT\_SIRI/04/04/2014/22.

**Acknowledgments:** The authors thank the Faculty Core Facility of the University of Hong Kong Li Ka Shing Faculty of Medicine for assistance with the flow cytometry analyses.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Spheroid Culture System Methods and Applications for Mesenchymal Stem Cells**

#### **Na-Eun Ryu 1, Soo-Hong Lee 2,\* and Hansoo Park 1,\***


Received: 26 November 2019; Accepted: 9 December 2019; Published: 12 December 2019

**Abstract:** Owing to the importance of stem cell culture systems in clinical applications, researchers have extensively studied them to optimize the culture conditions and increase efficiency of cell culture. A spheroid culture system provides a similar physicochemical environment in vivo by facilitating cell–cell and cell–matrix interaction to overcome the limitations of traditional monolayer cell culture. In suspension culture, aggregates of adjacent cells form a spheroid shape having wide utility in tumor and cancer research, therapeutic transplantation, drug screening, and clinical study, as well as organic culture. There are various spheroid culture methods such as hanging drop, gel embedding, magnetic levitation, and spinner culture. Lately, efforts are being made to apply the spheroid culture system to the study of drug delivery platforms and co-cultures, and to regulate differentiation and pluripotency. To study spheroid cell culture, various kinds of biomaterials are used as building forms of hydrogel, film, particle, and bead, depending upon the requirement. However, spheroid cell culture system has limitations such as hypoxia and necrosis in the spheroid core. In addition, studies should focus on methods to dissociate cells from spheroid into single cells.

**Keywords:** 3D cell culture; spheroid culture; biomaterials

#### **1. Introduction**

Stem cells are valuable resources in regenerative medicine with clinical and research applications (Table 1). Particularly, human mesenchymal stem cells have secretory properties constituted by anti-inflammation, angiogenesis, and immune reaction regulation factors [1,2]. The primary characteristic of stem cells is stemness, represented by their ability of self-renewal, which generate new same cells from the original stem cells, and multipotency, which allows production of new differentiated cells having relatively limited potential [3]. Another characteristic of stem cells is clonality, which is related to lineage of stem cells [4]. Stemness enables the stem cells to have potential for application in numerous biological therapeutic tools such as cell-based therapy, high-throughput pharmacology, drug screening, and tissue engineering. However, stemness is actually maintained in the in vivo microenvironment, which provides growth factors in addition to the cell–cell or cell– extracellular matrix interactions. Therefore, conditions of the cell culture are very essential to facilitate the properties of stemness, maintenance, and proliferation. To fully exploit the properties of stem cells, development of cell culture methods that can increase the proliferation of cells in terms of cell count and superior quality are important. Previously, studies have attempted to develop methods for enhancing stemness and proliferation during in vitro cell culture process [5,6].

Cells constantly require biological signals from the substrates in cellular niches. These signals encourage proliferation and enhance cellular viability. However, the signals can also inhibit proliferation and biological activation of cells, thereby checking the growth [7]. In general, traditional two-dimension cell culture systems, wherein the cells grow as a monolayer, face some limitations in the realization of in vivo multi-cellular conditions [8]. These limitations restrict the cellular studies involving multi-cellular features and cancer cells [9]. When stem cells grow in two-dimensional cultures, maintenance of the differentiation potential and stemness is relatively more difficult than in stem cells growing in actual multi-cellular conditions [6]. However, three-dimensional cell culture systems can reconstitute conditions similar to that in an in vivo microenvironment (Table 2). Three dimensional systems construct cell–cell and cell–extracellular matrix (ECM) interaction networks (Table 2), which play a significant role in various cellular mechanisms, subsequently maintaining the cellular properties [10]. Researchers can attempt to decrease the limitations of conventional monolayer culturing and develop progressive cellular study methods (Table 2).

Spheroid culture system is a promising three-dimensional cell culture method. Morphology of stem cells cultured in spheroid culture system is different from that of cells in the monolayer culture system. In addition, mesenchymal stem cells of spheroids maintain their intrinsic phenotypic properties by cell–extracellular matrix interactions (Table 2) [11].

Spheroid stem cell cultures promote the expression of transcriptional factors of stemness markers such as Oct-4 and Nanog (Table 2). Spheroid stem cells secrete higher levels of cytokines and chemokines that affect the proliferation, viability, and migration of cells and also secrete higher angiogenesis than those secreted in the monolayer stem cells (Table 2). Spheroid stem cells regulated by hypoxia-induced upregulation of gene expression have properties of apoptosis resistance, improved viability, and secretion of angiogenic factors and chemokines (Table 2) [12,13].


**Table 1.** Applications of spheroid of Mesenchymal stem cells.

**Table 2.** Advantages and disadvantages of spheroid culture.


#### **2. Mechanism of Spheroid Formation**

In a cell culture suspension, cells tend to aggregate and go through the process of self-assembly. Self-assembly means single cells constitute multi-cellular spheroids by themselves. Self-assembly is a natural phenomenon that happens during embryogenesis, morphogenesis, and organogenesis. It is affected by various factors, including gradients of nutrients, oxygen, and growth factors in cell culture

medium, as well as cellular paracrine factors. Cell culture medium permeates inside the spheroids by diffusion. The gradient of diffusion is induced by increasing the spheroid size during spheroid culture. The bigger the size attained by the spheroids, the harder it becomes for the medium to reach the core of the spheroids. In addition, the rate of production and consumption of factors can affect self-assembly [19].

Adhesion and differentiation of cells affect the formation of multi-cellular spheroids. In particular, cadherin and integrin are directly related to the mechanism of spheroid formation. The process of spheroid formation is divided into several steps. Firstly, single cells present within the suspension agglomerate to form loosely adhesive cell spheroids. In this step, extracellular matrix fibers including complementary binding of peripheral cell surface to integrin encourages preliminary aggregation. Next, E-cadherin promotes strong adhesion of initial cell aggregate by creating homophilic binding between cadherins of peripheral cells. In addition, β-catenin complex facilitates cellular signal transduction. Actin can also affect agglomeration and stemness by promoting contacts between adjacent cells [20]. As a result, strong adhesive multi-cellular spheroids are formed [21].

E-cadherin (CDH1), a Ca2+-dependent homophilic transmembrane adhesion molecule, could be a central component of spheroid formation [22,23]. The effects of E-cadherin have been demonstrated in cellular experiments involving human breast cancer cell lines and mouse embryonic stem cells [24]. For example, in a mouse embryonic stem cell study, E-cadherin-mediated cell attachment initiates embryonic body agglomeration. In addition, if the reaction of E-cadherin on the embryonic body with α-mouse E-cadherin antibodies is blocked, there is considerable inhibition of embryonic body agglomeration [25]. Integrins are transmembrane adhesion proteins composed of α-subunits and β-subunits of heterodimers that facilitate the cell-ECM connection during cell invasion and migration. Apart from E-cadherin, β1-integrin also plays a role in the attachment of the early spheroid formation. Interaction of integrin-ECM affects multi-cellular spheroid formation rates [21].

#### **3. Spheroid Formation Methods**

#### *3.1. Technical Methods*

#### 3.1.1. Pellet Culture

In this system, cells are concentrated to the bottom of the tube by centrifugal force. Cell–cell adhesions are maximized by proximity of the single cells at the bottom of the tube (Figure 1a), (Table 3). To harvest the cell pellet, supernatants are removed, and cell pellets are resuspended in spheroid formation cell culture medium. After estimating the cell count, cells in medium are dispensed into each well of a 96-well U bottom plate with cell repellent surface [19,26].

**Figure 1.** Schemes of technical methods. (**a**) Pellet Culture, (**b**) Liquid Overlay, (**c**) Hanging Drop, (**d**) Spinner Culture, (**e**) Rotating Wall Vessel, (**f**) Microfluidics, (**g**) Magnetic Levitation.


**Table 3.** Properties of technical methods.

Pellet culture can be used to induce differentiation of mesenchymal stem cells. In particular, a pellet culture system is suitable for stem cell differentiation by chondrogenesis since the interaction between adjacent cells in a pellet culture microenvironment is similar to the interaction in pre-cartilage condensation occurring during embryonic development. In pellet culture, mesenchymal stem cells can change their morphological shape from fibroblastic to polygonal in a manner similar to that in chondrocytes. Therefore, pellet culture system can be used for the study of signal pathways of chondrogenesis and for assessing the chondrogenic potentiality of stem cells [27,28].

#### 3.1.2. Liquid Overlay

Liquid overlay culture technique, also called the static suspension culture, forms spheroids by interrupting the adhesion of cells on non-adherent culture plates (Figure 1b), (Table 3). A non-adherent culture layer is typically composed of agar or agarose gel. Agarose is a very efficient material for the inhibition of cell attachment and is superior to agar with respect to its non-adherent properties. Since the cell attachments are inhibited, cells spontaneously form spheroids above the non-adherent surface by promoting cell–cell adhesive molecules [19,29,30].

Despite excellent non-adherent properties of agarose, this biomaterial has drawbacks in terms of culturing cancer cells. Agarose has trouble in interacting with tumor cells and is unable to activate the specific signaling pathways related to reaction of tumor cells to therapy process. Recently, hyaluronic acid can be a suitable alternative biomaterial that can replace agarose. It has the capability to interact with surface receptors of cancer cells during cancer progression. This interaction enhances transduction of cellular signals related to proliferation, angiogenesis, survival, and differentiation, as well as resistance to therapeutics [31,32].

#### 3.1.3. Hanging Drop

Hanging drop culture technique allows single cells to aggregate and fabricate spheroids in the form of droplets (Figure 1c), (Table 3). By controlling the volume of the drop or density of cell suspension, it is possible to control the spheroid size [33]. The novel hanging drop array platform is capable of efficiently forming definite size spheroids [34]. This technique can form circular spheroids having a narrow distribution of size with 10% to 15% variation coefficient, while the spheroid growth in non-adherent surface culture methods has 40% to 60% variation coefficient [35]. A general method involves starting from a monolayer cell culture, after which the cells are prepared as suspension and diluted with culture medium to attain the desired cell density. Subsequently, the cell suspension is dispensed into wells of a mini-tray with the help of a compatible multistep or multichannel pipette. A lid is placed on the mini-tray and the entire mini-tray is reversed upside down. The cell suspension drops attached on the mini-tray would stay on the reversed surface by surface tension. In this method, spheroids are formed as droplets owing to simultaneous action of surface tension and gravitational force [19,36].

Besides the adjustable size of a spheroid, hanging drop system has other advantages. There is no requirement of expensive or professional equipment to form spheroids for small scale experiments. A huge quantity of spheroids can be produced readily by multichannel pipetting and can be harvested

by scraping lids of culture dishes [33]. In addition, mesenchymal stem cells cultured via hanging drop system can secrete considerable quantities of potent anti-inflammatory as well as anti-tumorigenic factors [37].

#### 3.1.4. Spinner Culture

Spinner culture technique refers to the technique wherein the cell suspension in spinner flask bioreactor containers is continuously mixed by stirring (Figure 1d), (Table 3). The resultant spheroid is dependent on size of the bioreactor container [9,38]. Conditions of the fluid and mass in the containers are affected by the convectional force of the stirring bar, which is crucial to form the spheroid. A high stirring rate induces damage to the spheroid cells. However, an extremely slow rate of stirring allows spheroid cells to sink to the bottom of the container, resulting in inhibition of spheroid formation in the container [19].

In addition to adipogenesis, osteogenic differentiation of mesenchymal stem cells is also boosted by improved expression of osteogenic markers such as osteopontin and osteocalcin in the spinner system [39].

#### 3.1.5. Rotating Wall Vessel

Rotating wall vessel reconstructs microgravity by constant circular rotation [40]. Due to constant rotation, cells are continuously in a suspended state in the vessel (Figure 1e), (Table 3) [41]. This microgravity can affect gene expression of mesenchymal stem cells. In microgravity conditions, chondrogenic and osteogenic gene expression of stem cells reduces, whereas adipogenic gene expression is elevated [42]. This is because microgravity inhibits expression of *Collagen I* of the osteoblastic marker gene and integrin/Collagen I signaling pathway during the osteoblastic differentiation [43]. In addition, microgravity suppresses stress fiber development and improves intracellular lipid accumulation. However, reduction of osteogenic gene expression by microgravity can be regulated. Expression of RhoA protein switches these microgravitational effects and improves expression of the markers of osteoblastic differentiation of mesenchymal stem cells [44]. Expression of chondrogenic genes is increased by regulation of the p38 MAPK activation pathways [45].

#### 3.1.6. Microfluidics

This microfluidic culture technique, also called lab-on-a-chip technique, is used for applications such as single cell analysis, genetic assays, and drug toxicity studies. This culture method has microscale dimensions corresponding to the scale of in vivo microstructures (Figure 1f), (Table 3). In addition, microfluidic devices easily enable microscale control of the environment, mimicking the in vivo three-dimensional environment. One of the features of the microfluidic method is that it integrates multiple processes including cell capture, mixing, detection, and cell culturing. Another feature is a considerably high cell throughput for cell analysis. Microfluidic devices employ materials permeable to oxygen and growth factors affecting proliferation. This characteristic feature of microfluidics technology can decrease hypoxia, which is an unavoidable disadvantage of spheroid culture [46].

Recently developed fluidic systems overcome the limitations posed by the conventional fluidic system and offer advantages such as diversity of design and cost reduction through smaller requirements for specimens and reagents for cell transport assays [47]. Presently, the fluidic system can produce a distinct concentration of analyte mixtures and facilitates real-time monitoring of living cells. In addition, this system can optimize cell culture conditions for the proliferation and differentiation of stem cells, and be used for tissue engineering processes such as organ replacement and tissue regeneration, and in future clinical trials [48–50]. The currently used microfluidics system can be used to develop a co-culturing system related to the generation of microvascular network using mesenchymal stem cells. The co-culture system can also induce formation of a human microvascular network [51].

#### 3.1.7. Magnetic Levitation

Magnetic levitation-based culturing makes use of magnetic particles and integration with hydrogels according to the given conditions. In the magnetic levitation system, cells are mixed with magnetic particles and subjected to magnetic force during cell culture (Figure 1g), (Table 3). This system utilizes negative magnetophoresis, which can imitate a weightlessness condition, because positive magnetophoresis can hinder the attainment of weightlessness [52]. Due to magnetic force, the cells incorporated with magnetic particles stay levitated against gravity. This condition induces the geometry change of cell mass and promotes contact between cells, leading to cell aggregation. In addition, this system can facilitate multi-cellular co-culturing with agglomeration of different cell types [53,54].

When mesenchymal stem cells and magnetic particles are cultured with collagen gel, particle internalization takes place. Spheroid formation can be reproducible and reduces necrosis in the spheroid core, thus maintaining its stemness as a spheroid [54]. However, some groups have demonstrated that artificially manipulated gravity can lead to changes in cellular structures and can result in apoptosis [55,56].

#### *3.2. Using Biomaterials Methods*

#### 3.2.1. Hydrogels

Hydrogels are widely used for cell culture studies. Hydrogels have been fabricated using biocompatible materials such as alginate [57,58], fibrin [59,60], collagen [54] and hyaluronic acid [61,62]. The primary properties of hydrogels is that mesenchymal stem cells can be entrapped in them (Figure 2a), (Table 4). This method effectively improves the viability of cells while reducing cellular apoptosis. Furthermore, osteogenic differentiation potential is stably maintained and secretion of proangiogenic factors is activated in the hydrogel-entrapped cells compared to that in the non-entrapped cells of the monolayer culture [11,57,59]. Activated secretion of proangiogenic factors implies increased angiogenic potential and highly correlates to improved osteogenesis [63,64].

**Figure 2.** Schemes of using biomaterials methods. (**a**) Hydrogels, (**b**) biofilms, (**c**) particles.


**Table 4.** Properties of biomaterials.

Physicochemical biomimetic properties of hydrogels similar to those of the extracellular matrix are capable of offering functional niches promoting the self-renewal potential and wound healing. These properties of hydrogels improve angiogenetic capacity and stemness of the cells (Table 4) [65]. By adjusting the physical properties of hydrogel materials, the size of a spheroid can be optimized. To control the size of spheroids, weak adhesive materials can be used and physically embossed patterns on the surface of the hydrogel should be fabricated [66]. In addition, another main property of a hydrogel is the capability to deliver cells directly. Hydrogels can also be prepared in an injectable

form that can directly deliver stem cells to in vivo models and compensate for the necrotic or defective tissues [65,67].

Hydrogels have been developed to study the microenvironment of cancer cells. The stiffness of hydrogels can affect the phenotype and growth of cancer cells. The number and size of cancer cells tend to reduce when cultured in stiff hydrogels. However, tumorigenicity of cancer cells increases after in vivo transfer in softer hydrogels. This stiffness of hydrogels can be optimized by modifying the concentrations of the composing materials [68].

#### 3.2.2. Biofilms

Films made of biomaterials can be constructed by various methods such as photolithography and stamping (Figure 2b). Tumor cell spheroids can be cultured on films for their role in cancer drug discovery [69,70]. Apart from tumor cells, stem cells can be cultured on films. Stemness marker expression and differentiation potential are increased during culturing on film (Table 4) [71,72]. Adhesion and proliferation of stem cells can be enhanced by changing the composition and concentration of the film materials (Table 4) [73]. The component ratios used in the films are critical to the size of spheroids and rate of spheroid fabrication, as well as cellular adhesion and proliferation. In a previous study, hyaluronic acid(HA) modified chitosan film was found to form larger spheroids and induce cell aggregation in lesser time than the unmodified chitosan film [74]. The size of spheroids is also affected by the thickness of the film. Reducing thickness of the film leads to decreased spheroid size [75].

One of the biomaterials used in such films is chitosan. Culturing on chitosan films can improve angiogenesis, chemotaxis, and self-renewal [13,76]. Recently, graphene has been investigated as a cell culturing material [77,78]. Graphene films can provide distinctive environments beneficial to neurogenesis. Moreover, the neurons differentiated on graphene films have a remarkably keen sense of external stimulations. Graphene is believed to be capable of adjusting neural differentiation and growth of mesenchymal stem cells [79].

#### 3.2.3. Particles

Particulate factors have been used in spheroid cultures to control the cell culture microenvironment (Figure 2c). A drawback of spheroid cultures is the inadequate supply of nutrients and oxygen to the core of the spheroid. This is accounted by a rise in the diffusion gradient with increased spheroid size. However, particles within spheroids are capable of controlling conditions inside the spheroids during culturing. Consequently, the viability and proliferation of cells improve (Table 4) [80,81].

Particles are capable of regulating stem cell differentiation by controlling the extracellular environment [82]. Differentiation is also regulated by encapsulating stem cells in these particles. A previous study demonstrated that mesenchymal stem cell encapsulating particles, including a nanofibrous meshwork, could induce osteogenic differentiation [83]. However, particles can inhibit specific stem cell differentiation while inducing differentiation of other stem cells. This is achieved by controlling mechanotransductional mechanisms. Particles act as obstacles of internal adhesion between adjacent cells of spheroids. Alteration of mechanical force, including internal adhesion, surface tension, and interfacial tension in a spheroid, leads to biased differentiation of the stem cells [84].

The desired delivery of growth factors using particles can modulate the spheroid microenvironment. By transferring suitable growth factors into spheroids, differentiation can be spatially controlled [85,86].

#### **4. Applications of Spheroid**

#### *4.1. Study of Tumors*

Tumor cells are affected by cellular structures and extracellular matrix. The conventional 2D culture system has limitations pertaining to tumor cell culture. Spheroid culture system is a promising method for the study of tissue structure, signaling pathways, and immune activation of cancer cells.

Single tumor cells may form multi-cellular tumor spheroids mixed with other types of cells in a non-adherent 3D culture system, which is more effective in creating cellular heterogeneity [87]. Morphologies of these tumor spheroids are affected by spheroid culture microenvironment. In accordance with the conditions of culture microenvironment, morphologies may form aggregated circles, entangled bundles, elongated ovals, or star-shaped spheroids.

Spheroid form is the most suitable model for cancer study because it has a limited oxygen concentration at its core. This hypoxic nature of spheroids is the primary advantage of spheroid culture. However, in the case of tumors, spheroids larger than 500 μm in diameter undergo necrosis at their core [87] and have a concentration gradient of biological factors similar to tumor cells due to restricted diffusion of nutrients, oxygen, and growth factors [88].

#### *4.2. Drug Screening*

A study using an animal model has a limitation in disease modelling [89]. Presently, parameters of drug screening studies using a mouse model can possibly be overcome by adopting spheroid cell culture [90]. However, lack of uniformity in diameter or morphology of spheroids appears as new parameters for reproducible drug screening. By increasing the uniformity during the spheroid culture period, tumor spheroids can provide precise information on the diseases and suppress undesired side effects of the drugs under development [90,91].

In the context of tumor cell culture for drug screening, co-culture of normal cell and tumor cells can be a potential technique for reconstruction of the heterogenous multi-cellular environment for solid tumors as well as for promoting migration in tumors. The co-culture enables the investigation of interactions between tumor cells and peripheral multi-cellular environments. In addition, because normal host cells proximal to tumor cells can influence drug sensitivity of tumor cells, spheroid co-cultures can be used for drug screening study.

#### *4.3. Regenerative Medicine*

Transplantation is one of the most promising strategies for regenerative therapy. The currently used transplantation therapy has some drawbacks. In case of autografts, the amount of cellular supply is limited and the process of cell collection is cumbersome for the donor. However, allograft transplantation results in problems such as infection, inflammation, and host rejection [16]. Besides, injection in the form of single cells results in the limited immobility of injected cells at the site of the defect [92]. Injectable spheroids of stem cells are considered to improve the engraftment efficiency after transplantation [93]. After the implantation of spheroids, stem cells may be induced to differentiate into suitable cells for reconstructing the defective site [16,94]. Differentiation potential of spheroids has been demonstrated in vitro. Spheroid culture method improves differentiation potential compared to monolayer culturing [95].

Genetically modified spheroids have been developed for cell transplantation therapy [17,96]. These spheroids are prepared in the form of injectable suspensions. After transplanting these spheroids, the altered gene expression is maintained for a longer period of time in host tissues, whereas expression of cells cultured from monolayer plates decreases soon after transplantation. Thus, desired properties of cells transplanted in the host tissues are preserved by the process of spheroid culture [97].

#### **5. Conclusions**

Stem cells have shown applicability in various fields such as regenerative medicine as well as tumor and cancer research. Three-dimensional cultures enhance the applicability of stem cells by increasing the efficiency of culture. Spheroid culture system is an attractive method to overcome limitations of traditional monolayer culture. This system can resolve problems of monolayer culture such as the limited realization of in vivo multi-cellular microenvironments and it can reconstruct biological signal pathways of cell–cell and cell–ECM interactions, which encourage proliferation and viability of cells. Therefore, maintenance of the differentiation potential, stemness and intrinsic

phenotypic properties is improved. To conclude, development of spheroid culture is essential to further optimize formation of spheroids and utilize them as resource in the medical field.

**Author Contributions:** All authors have contributed to the conceptualization, writing and approved the final version of this manuscript.

**Funding:** This work was supported by Research of Korea Centers for Disease Control and Prevention (2018ER610300) and by the Chung-Ang University Research Scholarship Grants in 2018.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Brief Report*
