## Introduction

Fabrication of structured CM composed of multiple tissue types involves complex co-culturing under co-differentiation conditions5. Finding suitable conditions for co-culture, especially co-differentiation, of adipose and muscle cells is challenging, since each of these cell types requires markedly different matrix properties and media composition12. Alternatively, following separate cultivation and differentiation, multiple tissues can be integrated, as recently shown by Kang et al. in their assembly of differentiated muscle, adipose, and blood capillary cell fibers to construct steak-like meat using 3D bioprinting13.

In this study, we engineered a 3D fat tissue from isolated bovine adipose-derived MSCs loaded within alginate hydrogel. Differentiation and maturation periods of the isolated MSCs were optimized to yield the most fat-rich tissue in the shortest time, under the tested conditions. To create a well-integrated structured CMS that replicates the texture of conventional meat, the engineered adipose tissue was integrated within engineered bovine muscle tissue using a gentle stitching process that allowed the co-culture of the integrated construct while preserving the delicate mature adipocytes.

## Results

### Bovine MSC isolation and characterization

Bovine MSCs (BMSCs) isolated from renal adipose tissue of an adult cattle exhibited a fibroblast‐like appearance with a spindle‐shaped morphology (Supplementary Fig. S1a). Gene expression profiling by RT-PCR analysis (Supplementary Fig. S1b) found the cells to be positive for CD29, CD73, CD105, CD90, and CD44 and negative for CD45 (Fig. S1b), a typical profile for adipose-derived MSCs14.

### Adipogenic differentiation of BMSCs within 3D constructs

Towards the goal of bioengineering a 3D adipose tissue-like construct composed of isolated BMSCs, the isolated cells were first induced to differentiate into adipocytes in 2D culture. At passages 2–5, the BMSCs adipogenic differentiation was triggered by application of differentiation medium for 12 days. Twelve days post-differentiation, BMSC had accumulated lipids in their cytosol and acquired a round shape, typical of mature adipocytes (Fig. 1 top panel; b, c, e, f). The adipogenic differentiation was confirmed by Oil Red O (ORO) staining (Fig. 1b, e).

After establishing their adipogenic differentiation capacities in 2D culture, BMSCs were loaded into 3D alginate hydrogel beads and were subjected to the same differentiation conditions. Lipid droplet accumulation, as visualized by ORO staining and peroxisome proliferator-activated receptor gamma (PPARγ) staining, indicated robust differentiation of BMSCs upon adipogenic stimulation (Fig. 1, bottom panel). In parallel, lipoprotein lipase (LPL) and PPARγ expression levels were 3–6 times higher than in non-induced BMSCs cultured in standard medium only (Fig. 1m).

### Optimizing adipogenic differentiation and maturation periods

After selecting the matrix material for adipogenic differentiation in 3D, the differentiation profile of BMSCs in 3D alginate constructs was evaluated over time (Fig. S3a). Whole-mount immunofluorescence staining for CCAAT/enhancer-binding protein alpha (C/EBPα) (an adipogenic marker), and LipidTox revealed larger adipocytes with higher lipid droplet content in their cytosol after 21 days of differentiation as compared to 14 days (Fig. S3b). To obtain more mature adipocytes and thereby enrich the construct with lipid droplets, a maturation phase was added to the construct preparation protocol, using two methods: (i) short differentiation—2 weeks of differentiation and then 1–4 weeks of maturation, or (ii) long differentiation—3 weeks of differentiation and then 1–3 weeks of maturation, totaling a maximum 6-week differentiation and maturation period in both cases (Fig. 3a–e). To compare the outcomes of each protocol, the percentage of lipid droplet coverage, differentiation percentage and lipid droplet content per differentiated cell were quantified (Fig. 3b–d). At the 4-week time point, the lipid droplet coverage was significantly higher using the first method as compared to the second method (Fig. 3b). Within each method, when comparing between 4-, 5-, and 6-week time points, the increase in the lipid droplet coverage was not statistically significant (Fig. 3b). In addition, at the 4-week time point, larger fat-laden adipocytes accumulated in constructs with the short differentiation (2 weeks), as indicated by LipidTox staining and by the lipid droplet size distribution analysis, which showed a significantly higher average diameter (Fig. 3e). In conclusion, 2 weeks of differentiation followed by 2 weeks of maturation provided for the richest 3D engineered bovine adipose tissue in the shortest time, under the tested conditions.

To further optimize culture conditions toward the goal of cost-effective production of CM, the impact of shortening the adipogenesis period and decreasing cell density were evaluated. To this end, cells were seeded at a lower density and subjected to a 2-week adipogenesis protocol, comprised of differentiation and maturation phases of different lengths (Fig. 3f–i). The highest lipid content was achieved after 5 days of differentiation followed by 9 days of maturation (Fig. 3g, h). The differentiation percentage (i.e., the percentage of cells that had undergone adipogenic differentiation) and lipid droplet content per cell was significantly higher compared to the 4–6-week protocol (differentiation: ~45% vs. 10–30%, respectively; mean lipid droplet content per cell: ~2700 µm2 vs. 500–1000 µm2, respectively (Fig. 3c, d, i)).

### Engineered bovine adipose-muscle marbled-like construct

To mimic the structure of a natural steak, a combined fat and muscle 3D construct was created mimicking inter- and intra-muscular fat.

### Mold-cast marbled-like construct

To mimic intermuscular fat structure, mature engineered adipose tissue, formed of BMSCs and mature engineered muscle tissue, formed of bovine satellite cells (BSCs), were cultured separately for up to 1 month. Then, the engineered adipose and muscle constructs were integrated in a ring shape (Fig. 4 a(i)) and in a semicircular shape (Fig. 4 a(ii)). This was achieved by localized chelation of calcium ions at the connection area and re-cross-linking of the alginate using calcium solution. The combined tissues were co-cultured in adipogenic maturation medium for 1 week to achieve a fully integrated marble-like construct (Fig. 4a). The mold-cast marble-like constructs appeared to be integrated, without detachment points after 1 week of co-culture (Fig. 4a and Fig. S4a). Desmin-stained myotubes alongside mature adipocytes took on the ring and semicircular configurations of the mold (Fig. 4a). To further prove the differentiation of the satellite cells into muscle cells, confocal laser scanning microscope (LSM) images of samples stained for myogenin, a muscle-specific transcription factor involved in myogenesis, were taken from the areas of the muscle in the integrated constructs (Supplementary, Fig. S4b).

### Marble-like construct on 3D-printed scaffold

To mimic the structure of intramuscular fat, we developed a technique for extracting the mature adipocytes from the 3D alginate construct and incorporated them within an engineered muscle tissue which was made via 3D-printing (Fig. 4b). Staining of lipid and desmin demonstrated that the mature adipocytes remained intact, retained their lipid droplets during the extraction process, and were surrounded by muscle fibers, possessing a marble pattern, like marbled meat (Fig. 4b).

To further evaluate the potential of co-culturing the integrated adipose-muscle construct, a quantitative comparison of constructs before the integration procedure versus after integration and co-culture, for 2 types of adipose constructs (i.e., that were generated by different adipogenesis protocol) was conducted (Fig. 5). For this experiment, BMSCs-loaded alginate plugs were subjected to ‘2w diff. + 3w mat.’ or the ‘3w diff. + 2w mat.’ adipogenesis protocol. Then, the constructs were integrated with muscle cells-loaded alginate plugs or cell-free alginate plugs (‘control’), in a ring-shape configuration, and were co-cultured for 1 week. The integrity of the mature adipocytes, which are known to be highly sensitive, as a result of the integration procedure and throughout 1 week of co-culture was assessed by measuring the lipid content (Fig. 5). Lipid content was preserved following the integration process and the 1-week co-culture period in both adipose construct types (i.e., ‘2w diff. + 3w mat.’ and ‘3w diff. + 2w mat.’) (Fig. 5a). Furthermore, the lipid content was not affected by muscle cell presence, as determined by comparing the lipid droplet content in the constructs containing both adipocytes and muscle cells (‘co-culture’; Fig. 5b) to those containing adipocytes only (‘control’; Fig. 5b).

## Discussion

In this study, we engineered edible fat tissue for cultured steak production. The fat-rich 3D construct was created by loading an alginate hydrogel with isolated adipose-derived BMSCs, and allowing them to differentiate into mature adipocytes. In addition, we optimized differentiation and maturation periods to achieve lipid-rich tissue within minimal culture time. To create an edible marble-like construct, we then integrated engineered bovine fat tissue and muscle tissue using two techniques: (1) a gentle attachment technique that allows for continued co-culture; (2) extraction of mature adipocytes, that were cultured in alginate hydrogel, and placing them within the muscle tissue. The pioneering CM product first produced in 2013, was composed of muscle cells only15, as opposed to natural steak which consists of varying proportions of both muscle and fat. The fat melts during cooking and lubricates the muscle fibers, thus providing juiciness and mouthfeel.

CM production requires a reliable and reproducible stem cell source. Bovine adipose-derived MSCs are promising adult stem cells due to their multiple potential extraction sites, easy isolation, and high proliferation capacity in vitro16. In the present work, the cells from three independent extractions were characterized for their MSC morphology and gene expression and demonstrated a profile typical of adipose-derived MSCs14. The independent isolations produced consistent results in terms of morphology and gene expression, indicating successful and reproducible BMSCs isolation protocol.

CM production requires extensive cell expansion prior to the differentiation phase, to yield high volumes of cell material from a limited donor tissue. Repetitive passaging, however, may come at the cost of loss of differentiation potential. Indeed, the adipogenic potential of human MSCs has been shown to decline with increasing passages17,18, alongside increased adipogenic potential with older donor age18. In contrast, bovine adipose-derived MSCs showed significantly lower adipogenic differentiation in passage 5 as compared to passage 219. However, here we present preserved BMSC adipogenic differentiation capacity throughout passages 2–5. Further research will be needed to explore the limits of BMSC expansion and adipogenic differentiation capacity.

Further improvements of culture conditions were investigated, resulting in significant enhancement of differentiation percentage and lipid droplet content while shortening culture time. These improvements will directly reduce production costs35. The end-purpose of generating edible fat tissue is to integrate it within muscle tissue to create CMS. One of the challenges in developing a multi-tissue-type construct is to find the optimal culture conditions that will support selective differentiation of multiple stem cell types. This challenge can be overcome using a two-step process, as reported by Strobel et al.31. This approach first involves cell commitment to a specific lineage, usually the sensitive cell type that does not differentiate spontaneously, before combining it with another cell type and co-culturing them in an optimized medium. Alternatively, Shahin-Shamsabadi et al. combined two types of sheets in parallel, made of partially differentiated cells, to create a multi-tissue type construct36. However, co-culture of adipocytes and muscle cell precursors have been shown to suppress muscle cell differentiation37,38,39 and inhibit adipogenic differentiation40. A third approach grows and fully differentiates each cell type separately and then combines them to form an integrated construct. This approach can be applied for CM production and may be the most feasible and cost-effective among the mentioned approaches. Several recent publications have reported on application of this approach for structured CM development13,36. Although Shahin-Shamsabadi et al. attempted to create a multi-tissue-type construct for CM, they used model cell lines (murine C2C12 and 3T3-L1), which may not adequately represent primary cells36. The third approach involves attachment technique of the separated-grown tissues, while Kang et al. used transglutaminase over two days at 4 °C13, which terminates cell growth, the present work used a gentle attachment technique that preserve the cell viability and enable further co-culture of the integrated multi-tissue construct. In the current study, separately grown, differentiated and fully mature tissues formed of BMSCs and BSCs were integrated to form a marbled-like construct. The mold-casted construct was prepared in two geometrical orientations, mimicking intermuscular fat structure with mature adipocytes located between individual muscles. The marble-like construct fabricated on 3D-printed scaffold, mimicked intramuscular fat structures, which contribute to meat juiciness, flavor, and tenderness41. The integrated mold-casted construct was co-cultured in adipogenic maturation medium for 1 week following the attachment procedure. It is important to note that the purpose of this experiment was to assess the potential of attaching two mature tissues while preserving the cells in each of the tissues during a gentle stitching process and throughout the co-culturing. The 1-week co-culture step was implemented in attempt to generate a smooth and well-integrated co-construct, which will likely contribute to the feeling of whole meat-like cut that will not disintegrate during processing/heating.

Preparation of the marble-like construct on 3D-printed scaffold reported here involved an innovative technique of extracting the mature adipocytes from the alginate hydrogel while maintaining their integrity, loading them in alginate solution and cross-linking it on the engineered muscle tissue. This technique may open doors for a variety of applications. It is possible to control the concentration of the extracted mature adipocytes (it can be concentrated or diluted). They can be further loaded within a desired biomaterial, not necessarily the one required to induce their differentiation, thus the final scaffold material of the product can be replaced in this stage. Under appropriate conditions, the extracted mature adipocytes can be 3D bio-printed to form “highly structured” meat products42.

Both attachment techniques tested here yielded intact and physically stable combined constructs containing bovine fat and muscle cells. Neither the attachment process nor the 1-week co-culturing phase impaired adipocyte integrity. In addition, the presence of the engineered muscle tissue did not impair the mature adipocytes after one week of co-culture. These findings encourage and reveal an additional interesting and instructive information in the complexity of co-culture systems, however, it is certainly worthy of further investigation.

Overall, this study sought to engineer bovine fat tissue for cultured meat by encapsulating adipose-derived bovine MSCs within a 3D alginate scaffold and differentiating them into mature adipocytes. As described herein, efforts were made to economically optimize the production process by reducing culture times. Several strategies may improve CM production such as increasing cell growth and differentiation efficiency or reducing the costs of media for example by substituting low-cost growth factors. Moreover, evaluating the organoleptic properties of the cultured construct, as well as the fatty acid profile of the engineered adipose tissue, is essential for the development of cultured meat and should be further investigated in future studies.

This work reported on engineering of bovine fat tissue for cultured meat which entailed isolation of adipose-derived bovine MSCs, their encapsulation within a 3D alginate hydrogel and their later differentiation into mature adipocytes. Various matrices, fabrication techniques, and differentiation and maturation schedules, were evaluated to optimize the production of fat tissue for cultured meat. BMSCs in a 3D alginate matrix exhibited more robust adipogenic differentiation, while maintaining the construct dimensions, compared to those grown on a collagen matrix. The optimized process yielded lipid-rich constructs within only 2 weeks. Subsequently, we established two techniques for integration of engineered bovine fat tissue within engineered bovine muscle tissue, to imitate inter- and intra-muscular fat structures toward the goal of creating cultured steak.

## Methods

### Adipose-derived stem cell isolation and expansion

BMSC were isolated from the peri-renal adipose tissue of a 1-year-old Holstein Friesian cattle carcass. The resected tissues were soaked immediately in sterile phosphate-buffered saline (PBS) supplemented with 3% penicillin-streptomycin-nystatin solution (PSN; Biological Industries). The tissues were washed 3× with PBS/PSN, cleaned of residual blood vessels and connective tissue and minced with a sterile tweezer/scissors on a plastic dish. Then the tissue was digested with collagenase (type 1; Gibco) diluted in Dulbecco’s modified Eagle medium (DMEM; Gibco). Collagenase activity was terminated by addition of an equal volume of DMEM. The digested tissue was then passed through a 100-μm mesh filter and centrifuged at 1500 × g for 10 min at room temperature. The supernatant was discarded, and the pellet was resuspended in BMSC medium (DMEM - high glucose (DMEM-HG; Gibco) supplemented with 10% fetal bovine serum (FBS; Hyclone). This initial phase of the primary cell culture was identified as passage 0 (P0). Cells were cultured in an incubator at 37 °C under 5% CO2. After 4 days of culture, cells were washed twice with PBS to remove unattached cells and cultured in fresh medium. At 80% confluence, the cells were harvested with 0.25% trypsin-EDTA, reseeded at a density of 5 × 103 cells/cm2 (P1) and maintained in BMSC medium.

### Cell-loaded 3D hydrogel constructs

#### BMSC-seeded alginate matrix

BMSCs were cultured until 80–85% confluent, detached with trypsin-EDTA, and suspended in BMSC medium. The cell suspension was mixed at a 1:1 v/v ratio with 1% alginate (FMC Biopolymer) in PBS, to form a solution of 0.5% alginate and 40, 60, or 100 × 106 cells/ml.

The cell-alginate mixture was dripped directly into a calcium chloride bath (CaCl2 solution; 100 mM CaCl2 in DDW with 10 mM HEPES buffer), using a syringe without needle, to form ~2–3-mm-diameter beads. The cell-loaded beads were incubated for 15 min at room temperature and then washed 3 times, for 5 min, in PBS (+Ca, +Mg).

#### Plug constructs

Whatman® paper 1 was placed in the bottom of a sterile petri dish after saturating it with CaCl2 solution. Teflon molds (6 mm inner diameter, 1-mm thick) were then placed on the CaCl2-saturated Whatman paper. The cell-alginate mixture (30 µl) was then pipetted into each mold, and CaCl2 solution was gently sprayed on the upper surface. Next, CaCl2 solution was added to cover the entire plug and samples were incubated for 15 min at room temperature. After 15 min, the hydrogels were separated from the molds and washed in a PBS (+Ca, +Mg) bath 3 times, for 5 min each.

Beads/plugs were gently transferred to a 24-well plate with BMSC medium and incubated (37 °C, 5% CO2, 48 h), after which, the medium was changed.

#### BMSC-seeded collagen matrix

BMSCs were cultured until 80–85% confluent, detached with trypsin-EDTA, and suspended in collagen solution as follows: Type-1 collagen solution (8.44 mg/mL, Corning) was neutralized by mixing 1:1 (v/v) with sterile reconstitution buffer (3 volume of 0.1 N NaOH, 0.07 volume of 5% w/v NaHCO3, 1 volume of 100 mM HEPES buffer, and 1 volume of 10× M199) on ice. The cell pellet was then suspended in the collagen solution, resulting in a final collagen concentration of 5 mg/ml and a cell density of 40 × 106 cells/ml. The BMSC-collagen mixture was poured into culture dishes and polymerized at 37 °C for 30 min, after which BMSC medium was placed over the gel and incubated overnight.

### Cell culture, adipogenic differentiation, and maturation

For 2D culture, BMSCs were seeded at a density of 5 × 103 cells/cm2 on Matrigel (Cultrex)-coated 6-well plates in BMSC medium. After reaching 80% of confluency (within 24–48 h), the medium was changed to adipogenic differentiation medium (ADM; basal medium supplemented with Rock-inhibitor, WNT inhibitor and fibroblast growth factor). Non-induced BMSCs, cultured in standard medium (BMSC medium), were used as a negative control. Media were changed every other day for 2–3 weeks (the exact period mentioned for each experiment). For the differentiation of BMSC seeded within 3D constructs, the constructs were placed in non-TC, 24-well plates (1 construct per well) and cultured in BMSC medium immediately after seeding (1 ml medium per well). The medium was changed to ADM 3 days post-seeding, allowing the cells to adjust the 3D environment within the constructs before differentiation induction. ADM was changed every other day for 2–3 weeks (the exact periods mentioned for each experiment). For the adipocyte maturation phase, adipogenic maturation medium (AMM; BMSC medium supplemented with insulin and a cocktail of non-animal free fatty acids (FFAs)) based on previous work9, was added at different time points following BMSCs differentiation to adipocytes (the exact schedules detailed for each experiment in the “Results” section). The culture handling was performed as described in the differentiation protocol.

### Marble-like construct

Marble-like adipose-muscle tissue was generated using mold casting and 3D scaffold printing methods. In both methods, mature engineered tissues were combined, after culturing each tissue separately.

### Mold-cast marble-like construct

The mature bovine muscle tissue was created as we have previously described43,44. Briefly, BSCs were seeded on freeze-dried alginate-RGD scaffolds, incubated in proliferation medium for 1 week and then in myogenic differentiation medium for another 1 week43. The BMSC-loaded alginate plugs were adipogenically differentiated for 3 weeks to engineer the bovine adipose tissue. To combine the mature tissues seeded within an alginate scaffold, a previously published protocol was followed45. First, the tissues were cut, using a scalpel or a puncher, which exposed the ionic bonds. Two forms of cutting were performed: cutting in half and cutting in a ring-shaped manner. Then, a chelation solution (100 mM sodium citrate and 30 mM EDTA) was pipetted into the cutting area for 10 s (to locally chelate the Ca2+ ions from the exposed bonds). The two sections of the different tissues were then attached to each other and re-immersed in CaCl2 solution for 10 min, thereby re-cross-linking the alginate. To support integration of the different tissues, the marble-like constructs were cultured in adipogenic maturation medium for 1 week.

### Marble-like construct on 3D-printed scaffolds

#### Extraction of mature adipocytes from the alginate construct

BMSCs were seeded in alginate hydrogel and were differentiated into mature adipocytes, as described above. After adipocyte maturation within the hydrogel, the alginate hydrogel was dissolved, by adding 50 µl of 55 mM sodium citrate in 10 mM HEPES buffer and gently shaking. Upon dissolution, the cells were centrifuged (100 × g, 3 min) and the supernatant, containing the mature adipocytes, was gently transferred to a new tube, and mixed 1:1 (v/v) with 1% alginate solution.

#### Fabrication of the muscle tissue on a printed scaffold

The 3D-printed construct was fabricated as previously published44. Breifly, ink solution composed of alginate-RGD (1% w/v) and pea protein (1% w/v) was 3D-printed in longitudinal fiber orientation, using a micro-particles-based support bath that was contained 10 mM CaCl2 for cross-linking of the printed ink. Next, the 3D-printed scaffold was freeze-dried and then BSCs were seeded onto the scaffold. The cells were allowed to proliferate for 1 week and myogenically differentiate for 1 week43,44.

#### Creation of the marble-like construct

In the next step, the mature adipocytes in alginate solution were cast on the 3D-printed scaffold that was covered with differentiated BSCs, and re-cross-linked using 100 mM CaCl2 solution, forming an integrated combined construct.

### Oil Red O staining

Prior to staining, cultures were rinsed twice with PBS and fixed with 4% (w/v) paraformaldehyde for 15 min. Oil Red O (ORO) working solution was prepared by diluting 3:2 ORO stock solution (0.5% (w/v) in isopropanol, Sigma) with distilled water and filtering it through a 0.22 µm filter. Cells were incubated with isopropanol (60%) for 5 min and then covered with ORO working solution for 15 min. Cells were then washed with PBS 3–5 times, and then immersed in PBS until analysis. Cells in 3D constructs were fixated in 4% paraformaldehyde (w/v) for 15 min, and then incubated in a 30% (w/v) sucrose solution overnight, embedded in optimal cutting temperature compound (Tissue-Tek), and frozen for subsequent cryo-sectioning (10–20 µm). Samples were imaged under a light microscope (Axio Observer 7 microscope, Zeiss). Images were processed using the Zen Blue software (Zeiss).

### Immunofluorescence (IF) staining of cryo-sections

Cryo-sections were treated with 0.3% Triton X-100 (Bio Lab Ltd.) in PBS (PBS-T) for 5 min, to permeabilize the cell membranes. The sections were then incubated in blocking solution (1% bovine serum albumin (BSA) in PBS-T) for 1 h, at room temperature. Next, sections were incubated with rabbit polyclonal anti-peroxisome proliferator-activated receptor gamma (PPAR-γ) antibody (1:200, Abcam) in PBS-T with 1% BSA, overnight, at 4 °C. After extensive washings with PBS-T 4–5 times, for 5 min each, the sections were incubated with Alexa Fluor 647 goat anti-rabbit IgG (1:400; Invitrogen) and DAPI (1:1000, Sigma) diluted in PBS-T with 1% BSA, for 2 h, at room temperature. Sections were then rinsed 4–5 times with PBS, for 5 min each, and then mounted with Fluromount-G (Southern Biotechnology) and covered with cover slips (#1.5). The slides were then imaged using a fluorescence microscope (Axio Observer 7 microscope, Zeiss).

### Lipid droplet and IF staining of whole-mount scaffolds

Whole scaffolds were fixated with 4% paraformaldehyde (w/v) + 1 mM calcium for 10 min, and then washed several times with PBS containing calcium (PBS-Ca++). Next, permeabilization and blocking steps were carried out: the scaffolds were incubated with 0.05% w/v saponin in PBS-Ca++ containing 0.2% w/v BSA (S-B-P), for 20 min. Then, the constructs were incubated with mouse monoclonal anti-CEBP antibody (1:200, Abcam) diluted in S-B-P, overnight, at 4 °C. Constructs were then washed with PBS-Ca++, 4–5 times, 10 min each, and then were incubated with Alexa Fluor 647 donkey anti-mouse antibody (1:100, Jackson) and DAPI (1:1000, Sigma) in S-B-P, at room temperature, for 1.5 h. Thereafter, HCS LipidTOX™ Green (1:200; Invitrogen) was added to the secondary antibody solution and scaffolds were incubated in the staining solution for another 0.5 h. Then, scaffolds were washed 4–5 times with PBS-Ca++, 10 min each, and imaged using a confocal microscope LSM 700 (Zeiss). The marble-like constructs were stained in the same manner except that different primary and secondary antibodies were used—goat anti-desmin (1:100, Santa Cruz Biotechnology) and Alexa Fluor 546 donkey anti-goat (1:400, Life Technologies) or Cy5 donkey anti-goat (1:100, Jackson) antibody, respectively.

Images were processed in ImageJ software for enumeration, area measurement, and size distribution analysis of cellular lipid droplets. All images were taken under the same acquisition conditions. Therefore, the same parameters were used for calculating the surface areas. Lipid droplet diameters were calculated from the measured areas and imported into Excel software to generate a size distribution plot. Lipid droplet coverage was calculated by the total area of lipid droplets divided by the plug area, multiplied by 100%. In the experiment involving 6-week differentiation/maturation periods (Fig. 3a–e), samples (n = 4) were imaged in bright field mode at each time point and also fluorescently stained for lipid droplets (LipidTox) and for nuclei (DAPI). The lipid droplets were identified from the bright field images and their area was measured using ImageJ software (Fig. 3c). In the experiment involving 2-week differentiation/maturation periods (Fig. 3f–i), each sample was fluorescently stained for lipid droplets (LipidTox) and for nuclei (DAPI or Draq5) and imaged using a LSM microscope. The lipid droplets were identified from the fluorescent images using the analyze particles tool in ImageJ software (Fig. 3h). Differentiation percentage was calculated by dividing the number of lipid-laden cells by the total number of cells in each construct. The mean lipid droplet content per cell was calculated by dividing the total lipid area by the total number of lipid-laden cells.

### Measurement of mechanical properties

#### Stiffness measurements

Compression testing of BMSC-loaded hydrogels (alginate or collagen) after adipogenic differentiation was performed using an AR-G2 rheometer (TA Instruments, New Castle, DE, USA) equipped with parallel-plate geometry. Stress–strain curves were generated. The cross-sectional area of each sample was measured immediately before performing the compression test. Samples were imaged using a light microscope and the cross-sectional areas were measured using imageJ software. The compressive Young’s modulus, given by the slope of the stress–strain curve in the linear region, was used as the stiffness parameter.

#### Shrinkage measurements

Dimensions of BMSC-loaded hydrogels (alginate or collagen) were monitored by imaging the constructs and measuring their area on the day of seeding, and after 7, 14, and 21 days of differentiation. Shrinkage percentage was calculated by Eq. (1).

$$\% \,{{{{{{\rm{shrinkage}}}}}}}=\frac{{{{{{{\rm{area}}}}}}}\,{{{{{{\rm{on}}}}}}}\,{{{{{{\rm{seeding}}}}}}}\,{{{{{{\rm{day}}}}}}}-{{{{{{\rm{area}}}}}}}\,{{{{{{\rm{on}}}}}}}\,{{{{{{\rm{day}}}}}}}\,X}{{{{{{{\rm{area}}}}}}}\,{{{{{{\rm{on}}}}}}}\,{{{{{{\rm{seeding}}}}}}}\,{{{{{{\rm{day}}}}}}}}\;\cdot \;100 \%$$
(1)

### Total RNA extraction, cDNA synthesis, and RT-PCR analysis

Total RNA of BMSCs was isolated using the RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. Briefly, BMSCs were harvested from culture plates by trypsinization and washed twice with PBS. Then, cell pellets were disrupted by adding RLT buffer plus β-mercaptoethanol and the lysates were homogenized. Next, samples were centrifuged, and supernatants were extracted into a new tube containing ethanol. Samples were then transferred into a RNeasy spin column. The RNA concentration of the final elute was measured using a NanoDrop (Thermo Fisher Scientific) and stored at −80 °C. The RNA samples were reverse-transcribed into complementary DNA (cDNA) using the High-Capacity cDNA Reverse Transcription Kit (ABI; 4374966), according to the manufacturer’s instructions. The cDNA obtained was stored at 4 °C until use in PCR amplification reactions. The PCR primers and the length of the amplified products were as follows: CD29 (GACACGCAAGAAAATCCGAT and ACCGGCAATTTAGAGACCA, 89 bp), CD44 (CGGACCTGCCCAATGCCTTTGA and TGCACAGTTGGGAGGTGCGT, 226 bp), CD73 (TTCTCAACAGCAGCATCCCA and CAGTGCCATCCAGATAGACA, 122 bp), CD105 (CCATCAAAAGCCTGACCTTCGG and AGTCTGATGACCACCTCGTT, 138 bp), negative control CD45 (AAGCTGCGCAGGAGGGTAAACG and AAGCTGCGCAGGAGGGTAAACG, 206 bp), and house-keeping 18S (GGAGCGATTTGTCTGGGTTA and GTAGGGTAGGCACACGCTGA, 214 bp). Amplification was performed using the DreamTaq Hot Start Green PCR Master Mix (Thermo Fisher Scientific, USA), according to the manufacturer’s protocol. The PCR program included an initial hot start at 95 °C for 3 min, followed by 30 cycles of denaturation at 95 °C for 30 s, annealing at 51–60 °C, depending on the primer pair, for 45 s, elongation at 72 °C for 1 min, and ended with the final extension at 72 °C for 5 min. After the amplification reaction, the resulting products were separated by electrophoresis in 1.2% agarose gels and photographed with a UV trans-illuminator.

### Real-time quantitative PCR

RT-qPCR was carried out using TaqMan Fast Universal PCR Master Mix (2×) (ABI; 4352042). The following primers were used: 18S (Hs03003631; Thermo Fisher Scientific), GAPDH (Bt03217547_m1; Thermo Fisher Scientific), and PPARG (Bt03217547_m1; Thermo Fisher Scientific). A total of 10 ng cDNA was used for each reaction, according to the manufacturer’s instructions. The reaction was performed using the QuantStudio Real-Time PCR System (ThermoFisher Scientific, Waltham, USA). The amplification reaction conditions: 95 °C for 20 s, 40 cycles at 95 °C for 1 s, 60 °C for 20 s, in a 10 μL reaction volume. Each sample was processed in three replicates, with 18S/GAPDH used as a reference gene. The ΔΔCTt method was used to determine relative expression levels, where the target gene was normalized to 18S/GAPDH expression. Data were analyzed using QuantStudio design and analysis software version 1.5.1.

### Statistics and reproducibility

All statistical analyses in this study were performed using GraphPad PRISM software. For comparison between two groups only (Figs. 1m, 2c, d and 3e) a two-tailed Student’s unpaired t-test was used. For comparison between two groups only at different time points (Fig. 3b) a multiple unpaired t-test followed by Holm-Sidak post hoc test was performed. For comparison between multiple groups during a single time point (Figs. 3h and 5b) a one-way ANOVA followed by Tukey’s post hoc test was used. For comparison between multiple groups during several time points (Fig. 2b) a two-way ANOVA followed by Tukey’s post hoc test was performed. Significance was taken at p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), and p ≤ 0.0001 (****). Graphs were generated using Excel or GraphPad PRISM software and display all data points and/or mean ± standard error of mean (SEM) including the sample sizes as indicated in the legend of each figure.

### Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.