## Introduction

Injectable polymers (IPs) [19,20,21] are materials that exist in a sol state outside the body but form gels in situ when the aqueous polymer solution is injected into the body. In particular, temperature-responsive IPs that form a hydrogel upon injection into the body have attracted attention [22,23,24,25]. Such IPs may potentially be applied as biomedical materials in regenerative medicine (as cellular delivery systems and tissue regeneration scaffolds) [26, 27], drug delivery systems [28, 29], and other applications [30,31,32,33,34]. However, because most temperature-responsive IP hydrogels are physically cross-linked systems, they are limited by their relatively low elastic modulus and short in vivo duration. Therefore, IP systems with sufficient gel duration and physical strength for the desired application must be developed.

In recent years, several reports have been published on IP systems that can form covalent cross-links after temperature-responsive gelation via chemical reactions such as the Michael addition-type thiol-ene reaction [35, 36], condensation reactions using activated esters [37], and click reactions [38, 39]. The Michael addition-type reaction can proceed under physiological conditions (37 °C, pH 7.4) and can involve biomolecules (i.e., it is bioorthogonal). We previously developed a temperature-responsive IP system with covalent cross-linking ability by using an ABA-type triblock copolymer of poly(ε-caprolactone-co-glycolide) and poly(ethylene glycol) (PEG), hereafter referred to as tri-PCG [40, 41]. We then introduced acryloyl groups at both termini of tri-PCG to produce tri-PCG-acryl and mixed tri-PCG-acryl and tri-PCG micellar solutions containing dipentaerythritolhexakis(3-mercaptopropionate) (DPMP) as a hydrophobic hexafunctional thiol [42]. Thus, we succeeded in constructing IP systems exhibiting covalent cross-linking ability in response to temperature increases via the thiol-ene reaction. The maximum duration of the IP hydrogel exceeded 90 d, and the duration and mechanical strength (storage modulus) of the IP hydrogel could be controlled by changing the mixing ratio of tri-PCG-acryl in the IP formulation. We also reported that AdSCs embedded in this covalent bond-forming IP system could effectively treat ischemic heart disease [43].

Stem cells differentiate into different cell types depending on the elastic modulus of the culture matrix [44, 45]. Suppressing unintended differentiation and maintaining multipotency are desirable when performing regenerative medicine using the multipotency or inherent properties of stem cells. These requirements highlight the necessity of constructing a system that can maintain undifferentiated stem cells by default and differentiate them into target cells on demand using differentiation-inducing factors.

In this study, we investigated whether the undifferentiated state and multipotency of AdSCs could be maintained in IP hydrogels. We also examined the effect of the mechanical strength of the IP hydrogel on these properties by using a covalent bond-forming IP hydrogel system based on tri-PCG-acryl, which features controllable mechanical strength. Initially, we attempted to maintain the undifferentiated state of the AdSCs in the IP hydrogel and then differentiate them into the intended direction after culture by applying differentiating factors in the IP hydrogel or recovery from the IP hydrogel (Fig. 1). However, we found that the differentiation ability of the cells decreased after culture, which defeats our original purpose. These results reveal that the chemical and physical properties of the scaffold hydrogel should be carefully controlled. These phenomena are extremely important information for future attempts at tissue engineering using stem cells in such gels.

## Materials and methods

### Materials

Tri-PCG and tri-PCG-acryl were synthesized according to previously reported methods [42]. The materials for the synthesis and synthetic details are described in the Supplementary Information (SI). A LabAssay ALP kit was purchased from FUJIFILM Wako Pure Chemical Co., Ltd. (Osaka, Japan). Water was purified using a Millipore Elix UV3 Direct-Q UV system (Merck, Darmstadt, Germany). Dulbecco’s phosphate-buffered saline (PBS) (−) was purchased from Nissui Pharmaceutical Co., Ltd. (Tokyo, Japan). Dulbecco’s modified Eagle’s medium nutrient mixture F-12 (DMEM/F12), collagenase type VIII, 1% BSA solution, and Coomassie protein assay reagent were obtained from Sigma‒Aldrich (St. Louis, MO, USA). Ethylenediamine-N,N,N’,N’-tetraacetic acid and disodium salt (EDTA·2Na) were purchased from Dojindo Laboratories (Kumamoto, Japan). AdSC culture medium (KBM ADSC-1) and a lipid assay kit were purchased from Cosmo Bio Co., Ltd. (Tokyo, Japan). Fetal bovine serum (FBS) was purchased from Biowest (Nuaillé, France). FITC anti-mouse CD31, CD45, and CD90.2, FITC anti-mouse/human CD44, FITC anti-mouse/rat CD29, and PE anti-mouse CD34 antibodies were obtained from BioLegend (San Diego, CA, USA). NucleoSpin RNA and a PrimeScriptTM RT Reagent Kit (Perfect Real Time) were purchased from Takara Bio Inc. (Shiga, Japan). SsoAdvancedTM Universal SYBER Green Mix was purchased from Bio-Rad Laboratories Inc. (Hercules, CA, USA). Hank’s balanced salt solution (HBSS) was obtained from Thermo Fisher Scientific (Waltham, MA, USA). StemXvivo® Osteogenic Supplement, Osteogenic/Adipogenic Base Media, and Adipogenic Supplement were purchased from R&D Systems, Inc. (Minneapolis, MN, USA).

### Animals

Nine-week-old C57BL/6 N male mice were purchased from SHIMIZU Laboratory Supplies Co., Ltd. (Kyoto, Japan). AdSCs were collected from the mice as reported previously [13, 43]. Prior to the experiments, the mice were anesthetized with isoflurane using a small animal anesthesia station (DS Pharma Biomedical Co. Ltd., Osaka, Japan). All animal experiments were performed following the guidelines listed at Kansai University and approved by the Ethical Committee for Animal Experiments of Kansai University (7 May 2019; Identification No. 1903). This study was carried out in compliance with the ARRIVE guidelines.

### Measurements

Quantitative real-time reverse-transcription polymerase chain reaction (RT‒qPCR) was performed on a CFX96 deep-well real-time system (Bio-Rad Laboratories, Inc.). Flow cytometry (FCM) was performed using a Gallios flow cytometer (Beckman Coulter, Inc., CA, USA) with an Ar-ion coherent beam laser (488 nm). The UV absorbance of the microplated solutions was measured using an iMark microplate reader (Bio-Rad Laboratories, Inc.).

### Evaluation of surface expression markers for AdSCs

We collected AdSCs from 9-week-old male C57BL/6 N mice (SHIMIZU Laboratory Supplies Co., Ltd., Kyoto, Japan), as reported previously [13, 43], and confirmed the expression of AdSC-specific surface markers in these cells. The mice were euthanized by blood removal from the epigastric cavity using a syringe with a 26–29 G needle under isoflurane anesthesia. Subcutaneous inguinal adipose tissue was collected and placed on a culture dish (ϕ = 3.5 cm). The adipose tissue was cut into small pieces with scissors and placed in a 15 mL centrifuge tube. HBSS (3 mL) containing BSA (1%) and collagenase VIII solution (2 mg/mL) was added to the tube at 37 °C. After incubation at 37 °C for 30 min with inversion and mixing, 3 mL of DMEM/F12 medium containing 10% FBS was added to stop the enzymatic reaction. The top fat layer was removed by suction, and the supernatant was passed through a cell strainer (40 μm, BD Bioscience, Franklin Lake, NJ, USA) and collected into a 50 mL tube. Centrifugation (1200 rpm × 5 min) was performed to remove the supernatant. The cell pellet was suspended in PBS(−) containing 1 mM EDTA (1 mL) up to a volume of 10 mL. Centrifugation was performed once more, and the supernatant was removed. The cell pellet was suspended in DMEM/F12 or KBM ADSC-1 containing 10% FBS, seeded in culture dishes, and cultured in a humidified atmosphere containing 5% CO2 for 3 d at 37 °C to obtain AdSCs as adherent cells.

The cell suspension (2 × 105 cells/500 μL) was centrifuged (400 g) at 4 °C for 5 min, and the supernatant was removed. Antibody solution (100 μL) was added to the cells. Incubation in an ice bath was then performed for 30 min. FITC anti-mouse/rat CD29 (0.5 mg/mL), FITC anti-mouse/human CD44 (0.5 mg/mL), and FITC anti-mouse CD90.2 (0.5 mg/mL) antibodies were used as positive markers, whereas FITC anti-mouse CD31 (0.5 mg/mL), PE anti-mouse CD34 (0.2 mg/mL), and FITC anti-mouse CD45 (0.5 mg/mL) antibodies were used as negative markers. After incubation, the cells were washed twice with PBS(−), added to PBS (500 µL), and then subjected to FCM measurements.

### AdSC culture in the IP hydrogels and quantification of gene expression

The preparation of the IP formulations (tri-PCG/DPMP + tri-PCG-acryl mixture containing 16–33 wt% tri-PCG-acryl in the total polymer) and confirmation of their sol–gel transition are described in Figs. S1 and S2 in the SI and our previous study [42]. The IP formulations (tri-PCG only or a mixture of tri-PCG-acryl and tri-PCG/DPMP aqueous solutions) were sterilized by filtration. The AdSC suspension (100 μL) and IP formulation (100 μL) were mixed and dispensed into sterilized Eppendorf tubes. After gelation by incubation at 37 °C for 30 min, KBM ADSC-1 (400 µL) was gently placed on top of an IP gel containing AdSCs. The cells were then further incubated in a humid atmosphere containing 5% CO2 at 37 °C. An aliquot of the supernatant (200 μL) was removed and replaced with fresh KBM ADSC-1 every 2 d. Next, the IP gel was broken, and the homogeneous suspension was collected by repeated pipetting with PBS(−) (3 mL). The cells were collected by centrifugation. This operation was repeated three times to remove the IPs and DPMP.

mRNA extraction and RT‒qPCR were performed using NucleoSpin RNA and PrimeScript RT reagents, respectively, according to the manufacturer’s instructions. The mRNA levels of OCT4 [46], Nanog [47], and Sox2 [48] as undifferentiated markers and AFP [49], Brachyury [50], and Pax6 [51] as embryonic markers were quantified using the Bio-Rad CFX96 Deep Well Real-Time System, SsoAdvanced Universal SYBR Green Supermix, and primers obtained from Eurofin Genomics LLC (Louisville, KY, USA) (see the SI). GAPDH was used as the housekeeping gene in all experiments. Relative RNA expression was determined using Eq. (1):

$${{{{{{{\mathrm{Rel}}}}}}}}\;{{{{{{{\mathrm{ExP}}}}}}}} = 2^{ - \Delta {{{{{{{\mathrm{Ct}}}}}}}}}$$
(1)

where ΔCt = Ct of the gene of interest − Ct. Ct: cycle threshold.

### Differentiation of AdSCs into osteoblasts in the IP hydrogel (Protocol (1))

Tri-PCG-acryl (20 wt%) and tri-PCG/DPMP (20 wt%) were prepared separately via the protocols described above using osteogenic differentiation medium (StemXVivo) instead of ADSC-1 and then mixed and stirred with a vortex mixer to induce osteogenic differentiation (+). The obtained mixture (IP formulation; 75 µL) was sterilized by filtration, mixed with the cell suspension in the same medium (25 µL), and dispensed into a sterilized Eppendorf tube (100 μL). After gelation by incubation at 37 °C for 30 min, the cells were further incubated in a humidified atmosphere containing 5% CO2 at 37 °C for 3 d. The osteogenic differentiation medium (1000 µL) was gently placed on top of the IP gel containing AdSCs, and the cells were further incubated for 14 or 21 d in a humidified atmosphere containing 5% CO2 at 37 °C. The supernatant was replaced with medium every 2 d. Next, the IP gel was broken, and a homogeneous suspension was obtained by repeated pipetting with PBS( − ) (3 mL). The cells were collected by centrifugation. This operation was repeated three times to remove the IPs and DPMP. The collected cells were seeded in 24-well plates at a density of 2 × 104 cells/well and cultured in osteogenic differentiation medium. The medium was replaced every few days. AdSCs treated via the same protocol using KBM ADSC-1 instead of osteogenic differentiation medium were also prepared; these cells were considered osteogenic differentiation induction (−) cells. AdSCs cultured on tissue culture polystyrene (TCPS) dishes were prepared using osteogenic differentiation medium (osteogenic differentiation induction (+)) or KBM ADSC-1 (osteogenic differentiation induction (−)). The AdSC suspensions were used for alkaline phosphatase (ALP), total protein quantification, and RT‒qPCR assays.

### Differentiation of AdSCs into osteoblasts on TCPS after incubation in the IP hydrogel (Protocol (2))

AdSCs were incubated in the IP hydrogel via Protocol (1) using KBM ADSC-1 (osteogenic differentiation induction (−)) for 7 d. The obtained cells were seeded in 24-well TCPS plates at a density of 2 × 104 cells/well and cultured for 14 or 21 d in osteogenic differentiation medium (osteogenic differentiation induction (+)). The obtained cells were subjected to ALP assays, total protein quantification, and RT‒qPCR assays. The ALP assay and total protein quantification procedures are described in the SI.

### Differentiation of AdSCs into adipocytes on TCPS after incubation in the IP hydrogel

AdSCs were cultured in the IP formulation containing KBM ADSC-1 (without differentiation) for 7 d. The obtained cells were then seeded in 24-well TCPS plates at a density of 2 ×  104 cells/well and cultured for 14 or 21 d in StemXvivo® adipogenic differentiation medium. The culture medium was removed, and each well was washed with PBS(−) (500 µL). Neutral buffered formalin solution (10%, 500 µL) was added to the wells for fixation. After incubation at 25 °C for 20 min, the fixed cells were washed with water (500 μL). ORO solution (500 μL) from the lipid assay kit was dispensed into the wells, and the plate was incubated at 25 °C for 15 min. The ORO solution was then removed, and the wells were washed three times with water and dried. The extraction solution from the lipid assay kit (500 μL) was added to the wells, and the solution was incubated at 25 °C for 30 min to extract the dye. Finally, the absorbance of the extracted solution was measured at 570 nm using a spectrophotometer.

### Statistical analysis

Statistical significance was determined using the Tukey–Kramer method.

## Results and discussion

### Characterization of AdSCs collected from mice

We confirmed the expression of MSC-specific surface markers in cells collected from mice by FCM (see Fig. S3 in the SI). Most cells expressed the positive markers CD29, CD44, and CD90. CD29- and CD44-positive cells showed a unimodal distribution, whereas CD90-positive cells exhibited a bimodal distribution. Approximately 84% of the cells analyzed were CD90 positive. Almost no cells expressed the negative markers CD31, CD34, and CD45. These results indicate that nearly all of the cells collected from mouse adipose tissue could be characterized as AdSCs expressing the specific surface markers of MSCs.

### Maintenance of the undifferentiated state of AdSCs in the IP hydrogels

We investigated the gene expression levels of undifferentiated markers (OCT4, Nanog, and Sox2) [46,47,48] in AdSCs cultured in IP hydrogels with various tri-PCG-acryl contents by RT‒qPCR (Figs. 2a–c). The gene expression levels of the undifferentiated markers of cells cultured in the IP hydrogels were equal to or higher than those of AdSCs cultured on TCPS. Moreover, the gene expression level increased as the incubation time increased. Interestingly, the gene expression levels of the markers of undifferentiated AdSCs were higher after culture in the IP hydrogels than before culture.

We previously reported that the mechanical strength (storage modulus, G’) of IP hydrogels could be controlled from approximately 50–5000 Pa by changing the tri-PCG-acryl content [42]. The IP hydrogels used in this study exhibited storage moduli of 50–1000 Pa. However, no obvious trends in the relationship between tri-PCG-acryl content and the gene expression of the undifferentiated markers of AdSCs incubated in the IP hydrogels were observed. These results indicate that mechanical strength in this range (50–1000 Pa) has no significant influence on the differentiation potency of AdSCs. Thus, the AdSCs maintained their undifferentiated state even after culture in the IP hydrogels.

We investigated the gene expression levels of tridermic differentiation markers (AFP, Brachyury, and Pax6) [49,50,51] in AdSCs cultured in the IP hydrogels. The gene expression levels of AdSCs cultured in the IP hydrogels were equal to or higher than those of cells cultured on TCPS (Figs. 2d–f). Gene expression levels increased as the tri-PCG-acryl content increased, although this tendency was not obvious. In fact, most of the data were not significantly different, with small absolute values and large error bars, as determined by the Tukey–Kramer test. These results indicate that the types of genes that show increased expression in the AdSCs cultured in the hydrogels may differ depending on the properties of the hydrogel matrix, such as its elastic modulus. Overall, the increase in the gene expression levels of the undifferentiated cell markers was more apparent than that of the tridermic differentiation markers. The mechanism by which these undifferentiated cell markers increased in the IP hydrogel is currently unclear. Although some reports on the maintenance of the undifferentiated state are available [52,53,54], we did not find any report on the increase in undifferentiated cell markers attributable to the external environment. We presume that this phenomenon is due to the low elastic modulus and cell adhesion ability of the IP hydrogel, but we have not yet obtained concrete data to support this supposition. Based on the results, we concluded that the undifferentiated state of AdSCs was maintained in all IP hydrogels tested despite variations in tri-PCG-acryl content. Therefore, the IP formulation containing 16 wt% tri-PCG-acryl and an equivalent amount of DPMP were used in subsequent experiments.

### Evaluation of multipotency

Next, we investigated whether the AdSCs cultured in the IP hydrogel maintained their ability to differentiate into osteoblasts. The differentiation ability of osteoblasts was investigated using an ALP assay and RT‒qPCR according to the following protocols: Protocol (1): Differentiation was performed during incubation in the IP hydrogel with osteogenic differentiation medium; Protocol (2): The cells were first incubated in the IP hydrogel with normal medium for AdSCs (ADSC-1) and then differentiated on TCPS with osteogenic differentiation medium.