Direct Differentiation of Human Adult Adipose Tissue into Multilineage Functional Organoids

doi:10.21203/rs.3.rs-5054966/v1

Download PDF

Article

Direct Differentiation of Human Adult Adipose Tissue into Multilineage Functional Organoids

https://doi.org/10.21203/rs.3.rs-5054966/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Current strategies for generating organoids typically relay on intricate in vitro manipulation of dissociated stem cells including the isolation, expansion, and genetic processing of cells, which pose safety and scalability challenges for therapeutic translation. The direct generation of organoids from readily accessible tissues may offer a solution to these issues. Herein, we demonstrate the direct generation of functional organoids representing all three germ layers from human adult adipose tissue without any single-cell manipulation steps. Specifically, we generated reaggregated microfat (RMF) tissues from human adipose tissue using a specialized suspension culture system. These RMF tissues can be converted into mesodermal bone marrow organoids capable of reconstituting human normal hematopoiesis in immunodeficient mice; endodermal insulin-producing organoids able to reverse hyperglycemia in STZ-induced diabetic mice; and ectodermal nervous-like tissues resembling neurons and neuroglial cells. These findings highlight the potential of human adult adipose tissue as a safe and scalable platform for generating diverse clinically-relevant organoids for regenerative therapy.

Health sciences/Medical research/Translational research

Health sciences/Medical research/Stem-cell research

Direct differentiation

Regenerative Medicine

Human adipose tissue

Islet organoids

Bone marrow organoids

Neural differentiation

Adult stem cells

The replacement of dysfunctional tissues or organs with healthy ones is considered the ultimate solution in the field of medicine. The emergence of organoid technology in recent years has opened up new avenues to create biological substitutes for tissue repair or organ replacement¹. Currently, functional organoids are primarily derived from human pluripotent stem cells (PSCs)^2–4 and adult tissue-derived stem cells (ASCs)^5,6 through a series of intricate in vitro cell manipulation steps, which typically encompass cell isolation, expansion, and genetical processing to achieve the necessary quality and quantity criteria. Subsequently, these dissociated stem cells are cultured within three-dimensional (3D) culture systems and sequentially exposed to a series of signaling cues that provide appropriate biochemical and biophysical stimuli targeting key regulatory pathways. These processes promote the differentiation of stem cells into diverse cell types and facilitate their assembly into multicellular organoids. However, these complex procedures pose a scalability challenge for translating organoid technologies into clinically-relevant therapeutics due to labor-intensive, high cost, and time-consuming issues. Additionally, the cellular-level approaches currently employed present several challenges for clinical implementation, including limited availability of transplantable stem cells, difficulties in controlling cell fate, ethical considerations, immune rejection, and tumorigenic risks. Therefore, it is imperative to develop an efficient and safe strategies that can provide an unlimited regenerative source for generating therapeutically-relevant organoids.

The direct generation of functional organoids from easily accessible adult tissues may potentially solve these limitations and greatly extend the clinical applications of organoids. Most adult tissues harbor resident ASCs within specialized microenvironments known as niches⁷. The intricate interplay between stem cells and their niche components, including extracellular matrix (ECM) molecules, growth factors, cytokines, and physical cues, plays a crucial role in maintaining tissue homeostasis under normal physiological conditions and enables notable regenerative capacities in response to injuries⁸. Recent studies have demonstrated the remarkable plasticity of ASCs to differentiate across tissue lineage boundaries under both in vivo and in vitro conditions, irrespective of classical germ layer designations^9,10. These findings underscore the potential of adult tissues as a promising platform for generating functional organoids.

In this study, we present a simple, safe and scalable platform for the direct generation of multilineage functional organoids, demonstrating promising potential for disease modelling and therapeutic applications. By culturing human adult adipose tissues in a specialized suspension culture system, we generated reaggregated microfat (RMF) tissues capable of direct differentiation into functional organoids representing all three germ layers, including mesodermal bone marrow organoids that reconstituted normal human hematopoiesis in immunodeficient mice, endodermal islet organoids that reversed hyperglycemia in STZ-induced diabetic mice, and ectodermal neural-like tissues containing neuronal or neuroglial cells.

The in vitro self-reaggregation process enhances the multipotent differentiation capacity of microfat tissue

We first established a protocol for generating RMF tissues from human adult adipose tissue, as outlined in Fig. 1a and Extended Data Fig. 1a. Briefly, human adult subcutaneous adipose tissues were obtained from health individuals who underwent liposuction surgery. The harvested adipose tissues (lipoaspirate) were washed, minced, and mechanically emulsified into microfat tissues. These resultant microfat tissues were subsequently cultured in growth medium on ultra-low attachment plates for a duration of 3 weeks to facilitate their self-reaggregation. Following the suspension culture period, the fractionated microfat tissues gradually condensed into multiple distinct clusters of microtissues (Fig. 1a). Scanning electron microscopy (SEM) and hematoxylin and eosin (HE) staining results revealed a substantial presence of fibroblast-like cells within the generated RMF tissues, particularly on the surface (Fig. 1b, c). These observations were further supported by positive Ki67 staining (Fig. 1d), indicating active cell proliferation during suspension culture. Additionally, there was a gradual transition of the extracellular matrix (ECM) from an adipose phenotype to a stromal phenotype in the RMF tissues. This transition was evidenced by an increase in laminin and fibronectin expression in the RMF tissues, especially in regions with dense proliferating cells (Fig. 1e). Laminin and fibronectin are major ECM components of stem cell niches that significantly influence stem cell regenerative capacity^11,12. Quantitative analysis revealed a significant increase in DNA content throughout the 3-week period of self-reaggregation culture (Fig. 1f), suggesting notable augmentation in cell numbers within the microfat tissues. To further compact the RMF tissues, they were punched using a 4-mm biopsy punch after reaggregation. The harvested RMF tissue pieces were subsequently cultured for an additional week in growth medium, resulting in the formation of smooth and rounded RMF pellets, each containing approximately 5.92 × 10⁴ cells (Fig. 1g). Another week of self-reaggregation culture led to an increased presence of fibroblast-like cells on both surface and within the RMF pellets (Fig. 1b, c).

To characterize the cellular population within RMF pellets, we enzymatically dissociated RMF pellets into dissociated cells (referred to as RMF cells) and performed flow cytometry analysis to determine their cellular composition. Additionally, stromal vascular fraction cells (SVFs) obtained from native fat tissue, primary adipose tissue-derived stem cells (ADSCs) cultured for 5 to 7 days without cell passaging (referred to as P0 ADSCs), and expanded ADSCs that were regularly cultured and passaged for 4 weeks (referred to as expanded ADSCs) were included as control groups for comparative analysis of cellular composition (detailed information on cell harvesting and group design is provided in Extended Data Fig. 1b). Flow cytometry analysis revealed that the RMF pellets contained approximately 66.1% of mesenchymal stem cells (MSCs; CD45⁻CD73⁺CD90⁺), 11.15% of pericytes (CD45⁻CD31⁻CD146⁺), and 5.95% of endothelial progenitor cells (EPCs; CD45⁻CD31⁺CD34⁺). Compared with native fat tissue, the RMF pellets exhibited a significantly higher percentage of MSCs and comparable percentages of pericytes and EPCs. Compared with expanded ADSCs, the RMF pellets showed a lower percentage of MSCs but a significantly higher number of pericytes and EPCs (Fig. 1h). We further characterized the self-renewal and differentiation potential of the cell population within RMF pellets. As shown in Extended Data Fig. 1c, RMF cells exhibited comparable proliferative capacity to SVFs, but significantly higher than both P0 ADSCs and expanded ADSCs. Similarly, the colony-forming ability of RMF cells was comparable to SVFs but superior to expanded ADSCs (Extended Data Fig. 1d). Furthermore, RMF cells demonstrated a mesoderm trilineage differentiation potential similar to freshly harvested SVFs but greater than both P0 ADSCs and expanded ADSCs (Extended Data Fig. 1e).

To further elucidate the changes in cell population following a 4-week in vitro self-reaggregation culture, we conducted single-cell RNA sequencing (scRNA-Seq) on cells derived from native fat tissues, RMF pellets (4 mm), and expanded ADSCs using 10× Genomics platform. The resulting atlas consisted of 65501 cells that were clustered based on differential expression of cell-type-specific marker genes and visualized using a uniform manifold approximation and projection (UMAP) plot. Clustering analysis identified 8 distinct cell clusters across all 3 sample types, which were categorized into 5 main cell types based on their spatial distribution on the UMAP plots: adipose-derived stem and progenitor cell (ASPC) clusters (ASPC-1, ASPC-2, ASPC-3, and ASPC-4), preadipocyte cluster, endothelial cell cluster, smooth muscle cell cluster (SMC), and immune cell cluster, (Fig. 1i and Supplementary Fig. 1a). Cell type assignment was determined by previously reported markers and signatures for identifying cellular subpopulations in adipose tissue^13,14 (Supplementary Fig. 1b, c). The analysis of the proportion of the 5 main cell types in all samples revealed that, following a 4-week in vitro culture, the percentage of ASPCs increased from 24.57% in native fat to 97.17% in RMF pellets and further to 99.9% in expanded ADSCs (Fig. 1i). These findings align with flow cytometry analysis results obtained for the cell populations present within RMF cells, SVFs (native fat), and expanded ADSCs (Fig. 1h). To visually represent the relationship among these 3 sample types, we employed a ternary plot to position each cell according to its expression of known signature genes associated with these samples. The ternary plot shows that the cells in RMF pellet samples exhibit distinct differences compared to those in expanded ADSC samples but are more similar to those in native fat samples (Fig. 1j). We next performed cell-to-cell correlations to evaluate similarity between cultured samples and native fat samples. Notably, we observed that the cell clusters within RMF pellets displayed stronger correlation than those within expanded ADSCs with most corresponding cell clusters in native fat tissue (Fig. 1n and Supplementary Fig. 1d). Collectively, these data indicate that the self-reaggregation process during in vitro culture better preserves cellular phenotype similarity to native fat tissue compared to regular two-dimensional (2D) expansion methods.

In this study, we identified 4 distinct subpopulations of ASPCs across all 3 sample types. We employed pseudotime analysis to evaluate the stemness levels of these ASPC subpopulations in each sample type. Our findings revealed that the ASPC-1 cluster, predominantly derived from native fat samples (96.04%), occupied the initial position on the developmental trajectory, followed by ASPC-2 and ASPC-3 clusters primarily originating from RMF pellet samples (94.91% and 96.55%, respectively). Conversely, the ASPC-4 cluster mainly derived from expanded ADSC samples were positioned at the terminal end of the trajectory (Fig. 1k, m). Furthermore, pseudotime values indicated that both ASPC-2 and ASPC-3 clusters from RMF pellet samples exhibited an intermediate differentiation state between those observed in native fat and expanded ADSC samples (Fig. 1l). Notably, high expression levels of stemness-related genes such as SPON2, HTRA1, TIMP1, STMN2, PRDX2, PRDX4, CCND1, and NUMB, cell cycle-related genes such as MMP2, MMP14, and TWIST1, along with mesodermal lineage differentiation-related genes such as ADH1B, CTSK, LRP1, TIMP2, and FGF7 were observed in both ASPC-2 and ASPC-3 clusters (Supplementary Fig. 1e). These observations suggest that the ASPC clusters from RMF pellet samples possess an intermediary differentiated phenotype compared to their counterparts derived from native fat and expanded ADSC samples.

Generation of humanized bone marrow organoids from RMF pellets

We further evaluated the multipotency of the RMF tissues towards mesodermal lineages. Our objective is to transform RMF pellets into functional humanized bone marrow organoids, which should faithfully recapitulate the complexity of biological architecture and functionality of living human bone marrow niche in an animal model¹⁵. Figure 2a illustrates our 2-stage protocol for generating humanized bone marrow organoids from RMF pellets. The 4-mm RMF pellets were initially chondrogenically primed on ultra-low attachment plates for 4 weeks and subsequently induced to undergo hypertrophy for an additional period of 2 weeks. The chondrogenically primed RMF pellets displayed typical cartilaginous features, including cuboidal cells within large lacunae and abundant deposition of cartilaginous matrix that strong positive staining for Safranin O and collagen type II. Following the 2-week hypertrophic induction stage, the RMF pellets displayed a mature hypertrophic cartilaginous phenotype as evidenced by significantly increased expression of collagen type X, a marker specific to hypertrophic cartilage (Fig. 2b). Moreover, after the hypertrophic induction stage, active remodeling of typical cartilaginous ECM towards hypertrophic ECM was observed through the expression of matrix metalloprotein 9 (MMP9) and MMP13. Additionally, augmented osteogenic and angiogenic potential were indicated by expression of collagen type I and bone sialoprotein (BSP) within the RMF pellets (Extended Data Fig. 2a), indicating the potential of endochondral ossification upon in vivo transplantation. To evaluate reproducibility in generating hypertrophic chondroid pellets from RMF pellets using our protocol, we analyzed a total of 240 chondroid pellets derived from 6 independent adipose tissue donors (40 pellets per donor) using Safranin O staining. Microscopy analysis revealed that out of these samples, 44 were graded as worst quality, while 98 were graded as moderate quality and another 98 were considered best quality based on their characteristics (Extended Data Fig. 2b).

After hypertrophic induction, the generated chondroid pellets were subcutaneously implanted into the dorsal region of NOD/LtJ-Prkdc^scidIl2rg^em1/Shjh (NPSG) mice for further maturation and development (Fig. 2a). The implanted chondroid pellets were explanted at 2, 4, 6, 8, and 12 weeks post-transplantation (Extended Data Fig. 2c) and subsequently subjected to Movat’s pentachrome staining to demonstrate the presence of five distinct tissue types within the retrieved samples: bone, cartilage, bone marrow, adipose tissue, and fibrotic tissue. At the 2-week time point, the retrieved samples displayed a white appearance. Histologically, these pellets mainly exhibited a hypertrophic cartilaginous phenotype that predominantly stained blue in Movat’s staining. Meanwhile, few newly formed bone tissue and hematopoietic tissue were observed on the peripheral region of the samples. More bone and hematopoietic tissues were observed 4 week after transplantation. The retrieved samples acquired a dark-purple color at 6, 8, and 12 weeks. Additionally, Movat’s staining revealed a gradual replacement of cartilage-rich areas by bone tissue along with the formation of a thin cortical bone shell in the outer zone and concurrent expansion of the bone marrow cavity at 8 and 12 weeks (Extended Data Fig. 2c-f). These findings suggest progressive remodeling towards bone and hematopoietic tissues occurred in the implanted chondroid pellets. microCT analyses demonstrated that the implanted chondroid pellets gradually matured into ossicles with an average volume of 30 mm³ while exhibiting a progressive decrease in bone volume/total volume ratio (Extended Data Fig. 2g, h). To assess reproducibility of matured ossicles, we analyzed a total of 160 implanted chondroid pellets derived from 4 independent adipose tissue donors (40 pellets per donor) by HE staining. By week 12, bone and hematopoietic tissues were observed in 155 out of the 160 retrieved samples. Among these 155 retrieved ossicle samples, 48 ossicles were graded as worst quality, 76 were graded as moderate quality, and 31 were graded as best quality (Extended Data Fig. 2i), indicating a high level of reproducibility of mature ossicles.

We further characterized the ossicles retrieved at 12 weeks post-transplantation to elucidate the structural components of bone marrow organoid, including the endosteal niche adjacent to the trabecula and comprising skeletal cells, namely osteoblasts, osteoclasts, and osteocytes, as well as the perivascular niche adjacent to the sinusoid and containing multiple cells from mesenchymal and hematopoietic cell lineages¹⁶. The retrieved ossicles developed an architectural resemblance to normal bone marrow organ, with a predominant outer shell of cortical bone encapsuling spaces filled with trabecular bone, vascular structures, and various lineages of bone marrow cells, such as erythroid, myeloid, and megakaryocytic lineages (Fig. 2c). On average, 4.75 × 10⁶ bone marrow cells per ossicle could be extracted (Fig. 2d). The newly formed bone matrix was positive for osteocalcin (OCN). A portion of cells located within the bone matrix (osteocytes) or lined on the surface of bone matrix (osteoblasts) were positive for osteorix. Additionally, tartrate-resistant acidic phosphatase (TRAP) positive osteoclasts were observed on both trabecular and cortical bone surfaces, indicating an active remodeling process during ossicle development (Fig. 2e). Immunofluorescence staining of laminin, a marker of vascular wall basement membrane, demonstrated that blood vessels occupied the bone marrow cavity of ossicles, including newly formed vessels with sinusoid-like structures and matured vessels with arteriole-like structures. These vessels were further confirmed by NG2⁺ pericytes surrounding the newly formed vascular structures and sinusoid-like structures stabilized by α-smooth muscle actin (α-SMA) (Fig. 2g), respectively. These findings suggest a robust vascular network within the bone marrow cavity. The most noteworthy finding was a substantial proportion of human-specific NuMA positive cells were scattered in critical bone, trabecula bone, vessel structures, stromal tissues, and adipose tissues within the ossicles (Fig. 2h), indicating successful establishment of a humanized bone marrow microenvironment within the ossicles (hereafter the ossicles were referred to as hOssicles).

Immunofluorescence staining murine CD45⁺ revealed that the majority of these bone marrow cells were recruited from the host mice (Fig. 2f). Additionally, immunostaining showed a significant presence of CD146 positive stromal cells and leptin receptor positive (LepR⁺) MSCs (Fig. 2f), along with rare hematopoietic stem cells (HSCs; characterized by CD48⁻CD150⁺) (Fig. 2j), among the bone marrow cells. Flow cytometry analysis confirmed that the composition and proportion of mouse hematopoietic stem and progenitor cells (HSPCs) in hOssicles closely resembled those detected in mouse femurs (referred to as mFemurs), including total HSCs (lineage⁻ sca-1⁺ c-kit⁺, L⁻S⁺K⁺), short-term HSCs (ST-HSCs; L⁻S⁺K⁺CD34⁺CD135⁻), long-term HSCs (LT-HSCs; L⁻S⁺K⁺CD34⁻CD135⁻), multipotent progenitors (MPPs; L⁻S⁺K⁺CD34⁺CD135⁺), hematopoietic progenitor cells (HPCs; lineage⁻ sca-1⁺ c-kit⁺ progenitors, L⁻S⁻K⁺), megakaryocyte-erythroid progenitors (MEPs; L⁻S⁻K⁺CD34⁻CD16/32⁻), common myeloid progenitors (CMPs; L⁻S⁻K⁺CD34⁺CD16/32⁻), granulocyte-monocyte progenitors (GMPs; L⁻S⁻K⁺CD34⁺CD16/32⁺), and common lymphoid progenitors (CLPs; Lin⁻IL-7R⁺CD135⁺) (Extended Data Fig. 3a-d).

Collectively, our data demonstrate that the RMF pellets can be converted into hOssicles via recapitulating the endochondral bone formation process, which reconstitute a bone marrow niche of host mouse origin.

RMF-derived bone marrow organoids support human normal hematopoiesis in immunodeficient mice

To investigate whether the hOssicles can support human hematopoiesis and maintain functional hematopoietic cells in vivo, NPSG mice bearing hOssicles for 8 weeks were sublethally irradiated and subsequently intravenously transplanted with 5 × 10⁵ human umbilical cord blood-derived CD34⁺ (hUCB-CD34⁺) cells after a 24-hour interval, following a standardized protocol¹⁷. The engraftment level, self-renewal, and differentiation of transplanted hUCB-CD34⁺ cells within both hOssicles and mFemurs were analyzed using flow cytometry, colony formation assays, and immunohistochemical staining analysis (1st xenotransplantation; Fig. 3a).

Human cell chimerism was detected in all hOssicles at both 8 and 16 weeks after the 1st xenotransplantation, as evidenced by the presence of hNuMA⁺ cells within the bone marrow cavity of hOssicles through immunohistochemical staining (Fig. 3b). Over time, there was a significant increase in engraftment levels of human hematopoietic cells (percentage of human CD45⁺ cells out of all CD45⁺ cells, including human and mouse CD45⁺ cells) in hOssicles, with a rate of 30.72% at 8 weeks post xenotransplantation and 67.12% at 16 weeks post xenotransplantation, surpassing those observed in their corresponding mFemurs (12.67% at 8 weeks and 24.55% at 16 weeks; Fig. 3c), which is comparable to previously reported studies^18,19. These findings suggest that hOssicles exhibit a high frequency and level of engraftment for human HSPCs compared to mouse bone marrow. Additionally, immunohistochemical analysis further confirmed the presence of various types of human cells within the hOssicles explanted at 12 weeks, these include hCD45⁺ hematopoietic cells, hCD34⁺ HSCs, hCD19⁺ B lymphoid cells, hCD33⁺ myeloid cells, hCD14⁺ monocytes, and hCD15⁺ granulocytes; the hCD34⁺ cells were mainly located adjacent to sinusoids, whereas the hCD45⁺ cells were scattered throughout the bone marrow cavity (Extended Data Fig. 4). Quantitative analysis revealed that hOssicles maintained significantly higher levels of hHSCs (CD45⁺Lin⁻CD34⁺CD38⁻) and hHPCs (CD45⁺Lin⁻CD34⁺CD38⁺) than mFemurs at both 8 and 16 weeks post the 1st xenotransplantation (Fig. 3d and Supplementary Fig. 2a, b). The proportions of mature cells of human origin, including human T lymphoid cells (CD45⁺CD3⁺), human B lymphoid cells (CD45⁺CD3⁻CD19⁺), and human myeloid cells (CD45⁺CD3⁻CD33⁺), in hOssicles were comparable to those in mFemurs at both 8 and 16 weeks post 1st xenotransplantation (Fig. 3e and Supplementary Fig. 2a, b). To assess whether hOssicle can maintain functional hHSPCs within the microenvironment of the hOssicles, human cells extracted from explanted hOssicles at 16 weeks post transplantation, along with their corresponding mFemurs, were subjected for in vitro CFU assays. Both cell populations derived from the hOssicles and mFemurs generated CFUs, including burst forming unit-erythroid (BFU-E), colony forming unit-granulocyte and macrophage (CFU-GM), and colony forming unit-granulocyte, erythrocyte, monocyte, and megakaryocyte (CFU-GEMM). However, the numbers of BFU-E and CFU-GM formed by cells from the hOssicles were significantly higher than those observed in the cells from mFemurs (Fig. 3f). These findings indicate that RMF-derived hOssicles facilitate homing and repopulating human HSPCs in mice.

To further assess the presence of functional and self-renewing human HSPCs within the hOssicles, bone marrow cells were extracted from hOssicles or mFemurs and then transplanted into secondary NPSG mice, respectively (2nd xenotransplantation). hCD34⁺ cells were isolated from hOssicles and mFemur bone morrow cells from NPSG mice using magnetic-activated cell sorting (MACS), followed by intravenous transplantation into sublethally irradiated secondary recipients without hOssicle bearing (Fig. 3a). Eight weeks after the 2nd transplantation, we observed that the engraftment levels of hCD45⁺ cells and the proportions of immature hHSPCs (gated as hCD34⁺ cells) in mice transplanted with hOssicles-derived hCD34⁺ cells were higher than those transplanted with mFemurs-derived hCD34⁺ cells at both 8 and 16 weeks post the 2nd transplantation (Fig. 3g, h and Supplementary Fig. 2c, d), suggesting a higher self-renewal potential for hOssicle-derived hCD34⁺ cells. Similar to the 1st transplantation, the proportions of human T lymphoid cells, human B lymphoid cells, and human myeloid cells in mouse bone marrow were comparable in the mice transplanted with mFemurs-derived hCD34⁺ cells and hOssicles-derived hCD34⁺ cells (Fig. 3i and Supplementary Fig. 2c, d). Taken together, the results of this study indicate that RMF-derived hOssicles provided an optimal humanized microenvironment for long-term maintenance of self-renewal and differentiation capacity of human HSPCs and reconstitute human normal hematopoiesis in mice.

Generation of insulin-producing organoids from RMF pellets

To investigate whether the RMF tissues are capable of generating functional organoids of endodermal lineage, we converted RMF pellets into insulin-producing islet organoids. Initially, our objective was to establish a differentiation protocol with high efficiency and good reproducibility that would meet the cell quantity and quality thresholds required for preclinical research. We observed that the formation of dense RMF pellets with high cell density is crucial in generating functional islet organoids, probably due to native islets being mini-organs characterized by extensive metabolic requirements and high cell density. Consequently, we developed a modified protocol that facilitated the generation of smaller-sized RMF pellets with increased cellular density (Extended Data Fig. 5a). Implementing this modified protocol, we successfully produced round and dense 1-mm diameter RMF pellets from human adipose tissue after a total suspension culture period of 4 weeks (Extended Data Fig. 5b, c). Notably, as the pellet diameter decreased, there was a significant elevation in the DNA content per cubic millimeter of RMF pellets as well as an increase in cell density (Extended Data Fig. 5d, e). The 1-mm RMF pellets contained approximately 2.77 × 10⁴ cells (Extended Data Fig. 5e), predominantly comprising proliferating cells located on the pellet surface (Extended Data Fig. 5f). Furthermore, reducing the pellet diameter resulted in marked improvements in cell viability (Extended Data Fig. 5g, h). Therefore, due to its superior cell density and viability compared to 2-mm and 4-mm RMF pellets, we utilized 1-mm RMF pellets for subsequent experiments involving islet organoid generation.

Previously reported protocols for transdifferentiation of nonpancreatic ASCs across the germ layer border into insulin-producing cells or islet-like organoids involve two main steps: initial differentiation of ASCs into pancreatic progenitors followed by maturation into β-cells without stage-specific endodermal differentiation²⁰. Here, we present an optimized protocol for generating islet organoids from 1-mm RMF pellets that combines previous advances in differentiating insulin-producing cells from both ASCs^5,21,22 and PSCs²³. This protocol mimics the four critical stages of in vivo islet development, including definitive endoderm, pancreatic progenitors, endocrine progenitors, and β-cells (Fig. 4a). Following the initial induction stage, a majority of cells within RMF pellets co-expressed definitive endodermal transcription factors FOXA2 and SOX17 as confirmed by immunofluorescence staining (Fig. 4b). Flow cytometry analysis results revealed that the percentage of FOXA2 and SOX17 double-positive (FOXA2⁺SOX17⁺) cells was 56.68% (Fig. 4f and Extended Data Fig. 6b). Subsequent to the second-stage induction, 46.82% of the RMF cells co-expressed pancreatic progenitor transcription factors PDX1 and NXK6.1 (PDX1⁺NXK6.1⁺; Fig. 4c, f and Extended Data Fig. 6b). At the end of the third-stage induction, a small subset of RMF cells expressed C-peptide (CP), which is produced by β-cells during insulin formation. Approximately 11.93% of RMF cells were CP⁺PDX1⁺ and 15.12% were CP⁺NKX6.1⁺. This subpopulation also exhibited co-expression with key pancreatic endocrine progenitor markers, including NEUROG3 (NGN3) and MAFB (Fig. 4d, f and Extended Data Fig. 6b), indicating the presence of pancreatic endocrine progenitor cells within RMF pellets. At the end of the fourth-stage induction, a significant cell dead rate of approximately 55.94% was observed, while retaining 1.94 × 10⁴ viable cells within the pellets (Fig. 4g and Extended Data Fig. 6a). The expression CP was obviously increased in RMF pellets, the majority of these CP⁺ cells also co-expressed PDX1 and NXK6.1, with 15.78% of CP⁺PDX1⁺ cells and 20.15% of CP⁺NXK6.1⁺ cells in differentiated RMF pellets (Fig. 4e-g and Extended Data Fig. 6b). The β-like cell conversion efficiency of our protocol is relatively lower than those studies using ESCs and iPSCs, which varies widely across different studies (~ 10% to ~ 70%)²⁴. According to the percentage of CP⁺NXK6.1⁺ cells and viable cells within RMF pellets, we determinated that the average number of CP⁺NXK6.1⁺ cells per organoid was approximately 3000 (Fig. 4g). Notably, the expression of mature β-cell markers MAFA and NKX2.2 was also detected. Furthermore, we identified a subset of glucagon (GCG) positive cells and polyhormonal cells those stained positive for both CP and GCG (Fig. 4e). Additionally, stage-specific pancreatic genes were confirmed to be expressed at each induction stage using qRT-PCR (Extended Data Fig. 6c).

A key characteristic of functional pancreatic β-cells is their ability to secrete insulin in response to high-glucose stimulation. To evaluate this functionality in RMF-derived islet organoids, we exposed the final stage organoids to 3 sequential high-glucose challenges and subsequent depolarization with KCl under static in vitro conditions. The resulting organoids secreted minimal amount of human insulin (0.74 µIU per 1 × 10³ cells) when exposed to low glucose but showed a significant increase by 3.14 folds (2.31 µIU per 1 × 10³ cells) upon exposure to the first high-glucose challenge (Fig. 4h). Following the first high-glucose challenge, RMF-derived islet organoids displayed responsiveness towards subsequent challenges and maintained stable elevated levels of insulin secretion. Notably, additional membrane depolarization through KCl stimulation following the third high-glucose challenge resulted in maximal insulin secretion (3.14 µIU per 1 × 10³ cells) (Fig. 4h). Ultrastructure analysis conducted using transmission electron microscopy revealed the presence of insulin granules within RMF-derived islet organoid cells, on average 47 ± 17.8 insulin granules per cell, encompassing both developing insulin granules (diffused gray granules) and mature crystallized insulin granules (dense round and rod-shaped granules) (Fig. 4i). Collectively, these findings suggest that the generated islet organoids possess the capacity for producing packaged insulin granules and sensing high-glucose stimulation in vitro.

To further characterize the cellular composition of the generated RMF-islet organoids, we used scRNA-seq to analyze 8715 organoid cells derived from 2 adipose tissue donors (Fig. 4k and Supplementary Fig. 3a-c). We compared these data with published transcriptomes of human islet cells²⁵ (Fig. 4k and Supplementary Fig. 3d-f). Within the two RMF-islet organoid samples, we identified four distinct populations of pancreatic endocrine cells, including β-, α-, δ-, or γ-like cells. These identified endocrine cell types expressed canonical markers associated with their native counterparts in primary human islets, such as INS in β-cells, GCG in α-cells, somatostatin (SST) in δ-cells, and pancreatic polypeptide in γ-cells (Supplementary Fig. 3b, e). These pancreatic endocrine cells expressed classical human β-cell, α-cell, and δ-cell markers including PCSK2, UCN3, PAX6, SLC30A8, CHGA, and IAPP, (Supplementary Fig. 3c, f). Furthermore, through clustering analysis integrating the RMF-islet organoids and primary human islets database together, we observed that the β-cell identity score was closely resembled that of native islet β-cells (Fig. 4l and Supplementary Fig. 3g), indicating comparable levels of maturation between β-like cells and their native counterparts. Additionally, both groups shared similar gene expression profiles related to β-cell maturation, metabolism, as well as insulin secretion functions (Fig. 4l and Supplementary Fig. 3g). Signature gene analysis also revealed the presence, to some extent, of pancreatic exocrine cell types within RMF-islet organoids, including ductal- and acinar-like cells along with a small population of mesenchymal cells, as reported in previously published literatures using scRNA-seq^26–30 and immunohistochemical stainning^31,32 for stem cell-derived islet organoids. Collectively, these data suggest the generated RMF-islet organoids harbor a population of functionally matured β-like cells exhibiting transcriptional similarities to native human islet β-cells.

RMF-derived islet organoids acquire rapid graft vascularization and reverse diabetes in mice

We next further validated the longevity and functionality of RMF-derived islet organoids in surrogate animals. Previous studies have emphasized the importance of rapid connection with the host circulation for the maturation and subsequent functionality of stem cell-derived islet organoids^33,34. Therefore, in our initial set of experiments, we transplanted RMF-islet organoids derived from 3 independent donors under the kidney capsule of nondiabetic immunodeficient NOD.Cg-Prkdc^scid/Shjh (NOD/SCID) mice (3 mice per adipose tissue donor, 40 to 50 pellets per mouse) to assess graft vascularization and engraftment (Extended Data Fig. 7a). By week 1 post-transplantation, we observed a limited number of blood vessels around the organoid grafts in HE images; however, vessels in-growth within the center of organoid grafts were minimal (Extended Data Fig. 7b, Above). Immunofluorescence images confirmed that insulin positive (INS⁺) β-cells were present within the graft along with abundant single CD31⁺ endothelial cells but few formed tube structures (Extended Data Fig. 7c, Above). By weeks 2 and 4 post-transplantation, both peri-graft and intra-graft perfused vessels containing red blood cells were observed (Extended Data Fig. 7b, c, Center and below). Qualitative analysis revealed progressive increases in the vessel areas in HE images and CD31⁺ areas in immunofluorescence images over time after transplantation, indicating rapid vascularization of the transplanted islet organoid grafts (Extended Data Fig. 7f, g). Histological analysis further demonstrated significant reductions in the number of hypoxia-induced factor 1-α (HIF1α) positive cells and TUNEL positive cells within the grafts over time following transplantation (Extended Data Fig. 7d, e, h, i), suggesting an improved hypoxia microenvironment characterized by decreased hypoxia and apoptotic cell populations possibly attributed to enhanced vascularization within the grafts. Overall, these finding indicate that the RMF-islet organoids facilitate fast graft vascularization and engraftment during early stage of in vivo transplantation.

In the second set of experiments, we transplanted RMF-islet organoids generated derived from 6 independent donors (3 mice per adipose tissue donor) under the kidney capsule of NOD/SCID mice with streptozotocin (STZ)-induced diabetes to evaluate their long-term efficiency in glycemia control (Fig. 5a and Extended Data Fig. 8a). Considering the quantity of CP⁺/NKX6.1⁺ cells within RMF-islet organoids, we transplanted approximately 42 to 48 organoids into a single mouse kidney, which represents the maximum capacity for accommodating RMF-islet organoids in one kidney. Additionally, an equal number of undifferentiated RMF pellets derived from the same donors were employed as controls to evaluate any potential therapeutic effects of undifferentiated cells within the RMF pellets. Following transplantation, all mice receiving RMF-islet organoid grafts (STZ, Organoids) survived until nephrectomy on day 91 after STZ treatment. In contrast, median survival times for mice receiving RMF pellets (STZ, RMF pellets) and those without any transplantations (STZ, No txp) were recorded as 41.5 days and 49 days respectively; however these differences were not statistically significant (Fig. 5b). Regarding non-fasting blood glucose levels, mice receiving RMF-islet organoids exhibited a rapid decline in their levels post-transplantation, reaching stable values close to those of nondiabetic controls within approximately 2 weeks, and maintaining them throughout the analysis period. Notably, removal of the organoid grafts through nephrectomy at 12 weeks post-transplantation resulted in a rapid loss of glycemia control and reoccurrence of diabetes within 1 week (Fig. 5c and Extended Data Fig. 8b). These findings clearly demonstrate that the reversal of diabetes can be attributed to the transplanted RMF-islet organoids. In contrast, mice transplanted with undifferentiated RMF pellets maintained hyperglycemia levels close to 600 mg/dL until death (Fig. 5c and Extended Data Fig. 8b), further confirming the glycemia control effect exerted by pancreatic endocrine differentiated cells within the RMF-islet organoid grafts.

The body weights of diabetic mice with RMF-organoid grafts gradually increased in parallel with the reversal of diabetes, and no instances of overweight were observed (Fig. 5d and Extended Data Fig. 8c). Human insulin was exclusively detected in the serum of RMF-islet organoid transplanted mice as early as 2 weeks post-transplantation. The levels of human insulin in mouse serum were approximately 10 µUI/ml under fasting conditions, which exhibited a 1.6-fold increase at 30 minutes after intraperitoneal glucose injections. All RMF-islet organoid transplanted mice consistently maintained this capacity until nephrectomy (Fig. 5e, f and Extended Data Fig. 8d). Furthermore, intraperitoneal glucose tolerance tests (IPGTT) conducted at both 2 and 12 weeks post-transplantation demonstrated that the RMF-islet organoid transplanted mice displayed significantly improved glucose tolerant and faster glucose clearance compared to those transplanted with RMF pellets or without any transplantations (Fig. 5g and Extended Data Fig. 8e).

Histological analysis confirmed the preservation of cellular complexity in the organoid grafts. This was evident from the presence of a distinct population of cells with abundant cytoplasm within the organoid grafts, as observed in HE images (Extended Data Fig. 8f). These cells were positive for INS and perfused with CD31⁺ vasculatures. A minor populations of GCG⁺ and SST⁺ cells, as well as residual perilipin-1⁺ adipocytes were observed. Furthermore, these INS⁺ cells co-expressed critical β-cell transcriptional factors PDX1 and NKX6.1, as well as mature β-cell markers MAFA and NKX2.2. Approximately 18.63% of cells within the grafts were INS⁺NKX6.1⁺ cells, which is slightly higher than that those observed before implantation. Notably, the human origin of the islet organoid grafts was confirmed by the coexpression of INS and hNuMA (Fig. 5h, i).

We compared the single-cell transcriptomes of RMF-islet organoids before and 12 weeks post-transplantation (Fig. 5j, k and Supplementary Fig. 4a-d). The findings confirm the presence of all four major pancreatic endocrine cell types within the organoid grafts and displayed a similar cellular composition before and after transplantation (Fig. 5j, k). Subtle molecular alterations associated with maturation, insulin secretion, and metabolism functions are observed, including expression of INS, NKX2.2, ENUROD1, UCN3, CPE, G6PC2, CHGA, PCSK1, IAPP, and SLC30A8 (Fig. 5k). These data together indicate the RMF-islet organoids undergo further maturation following transplantation.

Postmortem histological analysis revealed complete destruction of the native pancreatic islets following STZ treatment (Extended Data Fig. 9a), confirming successful establishment of a diabetic model in NOD/SCID mice. After 13 weeks post RMF-islet organoid transplantation (14 weeks post STZ treatment), no evidence of β-cell regeneration was observed (Extended Data Fig. 9b-d). Subsequent postmortem examinations conducted on main organs from the mice transplanted with RMF-islet organoids demonstrated that the absence of macroscopic and microscopic evidence suggestive of tumorigenesis after transplantation (Extended Data Fig. 9e), thereby confirming both efficacy and safety associated with RMF-islet organoid transplantation for diabetes treatment. Collectively, these findings demonstrate the long-term functionality of RMF-islet organoids in a murine model.

Generation of neural-like tissues from RMF pellets in vitro

To investigate whether RMF tissues can be converted into ectodermal tissues, we employed a new strategy to generate neural tissues. Initially, 1-mm RMF pellets were differentiated into neurospheres with a neural stem cell (NSC) phenotype and subsequently induced towards neuronal and neuroglial lineages through sequential culture in specialized differentiation media (Figure S10a). Following 2 weeks of culture in serum-free medium supplemented with FGF-2 and EGF, immunofluorescence analysis revealed that that majority of cells within RMF pellets expressed Nestin, an intermediate filament protein expressed in NSCs. Notably, a subpopulation of these cells also expressed the proliferation marker Ki67 as well as several transcriptional factors involved in neurogenic differentiation including SOX2, PAX6, and FOXG1 (Extended Data Fig. 10b), suggesting their acquisition of a NSC phenotype. These NSC-like cells displayed multipotency as they expressed mature neuron markers such as neuronal nuclei (NeuN), TUJ1, and MAP2 upon culturing in neuronal differentiation medium (Extended Data Fig. 10c). Furthermore, when cultured in neuroglial differentiation medium, they gradually expressed proteins associated with neuroglial lineage commitment including glial fibrillary acidic protein (GFAP) and S100β (Extended Data Fig. 10d). qRT-PCR further confirmed the expression of critical transcriptional factors and markers at the NSC stage including PAX6, SOX2, FOXG1 and Nestin (Extended Data Fig. 10e), neurogenesis stage including NSE (Neuron-specific enolase), MAP2, TUJ1, NEFL (neurofilament light chain), and NEFM (neurofilament medium chain) (Extended Data Fig. 10f), and gliogenesis stage including GFAP and S100B (Extended Data Fig. 10g). Collectively, these findings provided evidence for the successful conversion of human adult adipose tissues into neural tissues.

The leading strategy for generating organoids currently is differentiated from dissociated stem cells and requires intricate in vitro cell manipulation steps, which limits their clinical translation potential. Herein, we demonstrated a straightforward, safe and scalable approach for the direct differentiation of functional organoids and tissues from human adult adipose tissues in vitro. These include the generation of mesodermal humanized bone marrow organoids that capable of constituting normal human hematopoiesis in immunodeficient upon xenotransplantation of human HSCs, endodermal insulin-producing organoids that effectively ameliorated hyperglycemia in STZ-induced diabetic mice following kidney subcapsular transplantation, and neuroectodermal neural-like tissues exhibiting phenotypic features of neurons and neuroglial cells. Importantly, these direct differentiation processes occurred at the tissue level in vitro without any in vitro cell manipulation steps, representing a significantly departure from conventional cellular-based strategy. Collectively, our findings suggest that, as an abundant and easily accessible tissue source, human adult adipose tissue holds potential as a simple, safe, and scalable platform for generating diverse clinically relevant-organoids for regenerative therapy.

In this study, we hypothesized that by preserving the tissue integrity of adult tissues to maintain ASCs within their native niches and remodelling their niche microenvironment towards a regenerative state in vitro, the lineage-specific differentiation potential of ASCs can be extended beyond germ boundaries, enabling their differentiation into cell types from different germ layers. To validate this hypothesis, we employed human adult adipose tissue as an exemplar of tissue stem cell niches for generating functional organoids due to its high accessibility, easy to handling, less ethical concerns, and most importantly, its remarkable regenerative plasticity in tissue turnover^35,36. Previous studies have shown that after stimulated with growth factors and small molecules in vitro, terminally differentiated adipocytes can undergo dedifferentiation into a multipotent state with the potential to generate tissues of all three germ layers after transplantation, including muscle, bone, cartilage, neural tissues, and blood vessels^37,38. In our approach, we cultured fractionated microfat tissues in a suspension culture system to maintain their histological integrity as adipose tissues, thereby ensuring the maintenance of ADSCs within their native niche microenvironments; simultaneously allowing for increased ADSC proliferation within RMF tissues. As expected, we observed that the cells within RMF pellets exhibited enhanced tri-mesodermal lineage differentiation potential and displayed higher pseudotime values compared to ADSCs expanded under 2D conditions. Similarly, previous studies have demonstrated that preserving tissue integrity enables long-term expansion of fetal brain-derived organoids³⁹ and maintains cellular heterogeneity in tumor-derived glioblastoma organoids⁴⁰, potentially by creating an appropriate tissue-like niche microenvironment in intact tissues.

In this study, we successfully generated humanized bone marrow organoids in mice for modeling human hematopoiesis. Bone marrow organoids should faithfully replicate the 3 components of the human bone marrow niche in animal models: the endosteal niche, the perivascular niche, and the hematopoietic niche^4,16,41. Currently, the leading strategy for establishing bone marrow organoids involves generating hOssicles by co-culture of human MSCs within different matrices¹⁵, which are able to form an ectopic humanized bone marrow niche when transplanted into immunodeficient mice^18,41,42. We employed a developmental bioengineering approach to generate hOssicles from RMF tissues. In this approach, the RMF pellets were initially induced into a hypertrophic cartilaginous phenotype in vitro and subsequently transplanted NPSG mice for further maturation by recapitulating the process of endochondral ossification. By tracing cell origin, we confirmed the endosteal niche was mainly comprised of cells of human origin, indicating the establishment of a humanized microenvironment. This facilitated successful reconstitution of human normal hematogenesis upon xenotransplantation of hUCB-CD34⁺ cells. To our knowledge, this is the first study demonstrating direct generation of humanized bone marrow organoids from adult adipose tissue. These organoids hold great potential for modeling normal and malignant hematopoiesis as well as bone metastasis in solid tumors.

After successfully converting adult adipose tissue into bone marrow organoids of the same mesoderm origin, we further hypothesized that by modifying the niche microenvironment to a specific regenerative state through the provision of a set of biochemical and biophysical cues, adult adipose tissue derived from the mesoderm could also transition beyond germ layer boundaries into tissues originating from endodermal or ectodermal lineages. This parallels the differentiation potential observed in their cellular counterpart, ADSCs, known for their capability to transdifferentiate into cells originating from endodermal lineages such as hepatocytes^43–45 and β-like cells^46,47, as well as cells originating from ectodermal lineages such as neurons^48,49, Schwann cells^50,51, keratinocytes^52,53, and corneal epithelial cells⁵⁴. Thus, we next validated the feasibility of conversing RMF tissue into insulin-producing islet organoids derived from endoderm. Currently, islet organoids are primarily derived from PSCs, encompassing both embryonic stem cells^24,27,34 and induced pluripotent stem cells^28,29, or transdifferentiation from various lineage cells including human fibroblasts^30,55,56, pancreatic exocrine acinar cells⁵⁷, liver cells⁵⁸, and gastric stem cells⁵. As an alternative approach to this paradigm, we successfully converted RMF pellets into islet organoids that resembled similar cellular composition to native islets and displayed biphasic glucose-dependent insulin secretion capacity in vitro. Notably, the conversion of RMF tissues into islet organoids occurred without transitioning through a pluripotent state, distinguishing them from generation of β-like cells from fibroblasts^30,55,56 or ADSCs^59,60 by introducing exogenous lineage-specific transcription factors or exposure to chemical compounds. Upon transplantation into STZ-induced diabetic mice, these organoids showcased their ability to effectively ameliorate hyperglycemia towards normal levels and maintained glycemic improvement throughout the analysis period. These findings hint at the potential application of RMF-islet organoids as a regenerative therapy option for clinical diabetes treatment. Moreover, we further demonstrate the possibility of generating neuroectodermal tissues from RMF tissues. Upon stimulation within specific neurogenic media, RMF tissues exhibited expression of neuron markers or neuroglial cell markers, indicating their potential for application in nervous system generation. However, the maturity and functionality of these future nerve-like tissues needed further investigation.

Collectively, we have developed a simple, safe, and scalable platform for the direct differentiation of functional organoids from human adult adipose tissue. This adult tissue-based strategy bypasses the need for cell isolation, extensive expansion, and genetical processing while also overcoming limitations associated with the use of PSCs, such as ethical and legal concerns as well as tumorigenic risks. Furthermore, our findings unveil the remarkable regenerative plasticity of human adult adipose tissues through the establishment of a RMF culture system. This plasticity highlights the potential of human adult adipose tissue as an abundant source for developing therapeutic agents.

However, as a newly developed organoid generation strategy, there are several aspects that necessitate further optimization. The heterogeneity of human adipose tissue samples from different donors remains an unpredictable factor for successfully generating functional organoids. Additionally, as a pivotal player during the direct differentiation, the development of highly efficient medium and condition combinations to further improve the differentiation rate will be prioritized of future studies. Lastly but significantly, unraveling the detailed mechanisms underlying how the adipose stem cell niche within the RMF tissues supports their multilineage differentiation remains elusive. Dissecting the molecular, cellular, and spatial components of adipose stem cell niches during microfat tissue self-reaggregation and ADSC differentiation could contribute to advancing this technology towards clinical application.

Data and code availability

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information, except for the raw data for scRNA-seq reported in this publication can be accessed under the Gene Expression Omnibus (GSE182159) and the Genome Sequence Archive (https://ngdc.cncb.ac.cn, accession number HRA001730) on request. This paper analyses existing data. These accession numbers for the datasets are listed in the key resources table.

Any additional information is available upon request.

Animal models

All the animal experimental procedures were approved by the Animal Care and Use Committee of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, with project license numbers of SH9H-2021-A79-1 (bone marrow organoid associated animal experiments) and SH9H-2023-A778-1 (islet organoid associated animal experiments). Mice were housed in the specific pathogen-free facility at Shanghai Ninth People’s Hospital under a 12-hour light/12-hour dark cycle (06:00 to 18:00) at room temperature of 22 ± 1 °C, with ad libitum access to food and water.

For subcutaneous implantation of RMF-derived chondroid pellets, NOD/LtJ-Prkdc^scidIl2rg^em1/Shjh female mice aged 4 weeks were used. For secondary transplantation of hCD34⁺ cells, NOD/LtJ-Prkdc^scidIl2rg^em1/Shjh female mice aged 8-10 weeks were used. For kidney subscapular implantation of RMF-derived islet organoids, NOD.Cg-Prkdc^scid/Shjh with aged 8-10 weeks were used. Mice of the same sex were randomly assigned to experimental groups. Both strains of mice were purchased from Shanghai Jihui Laboratory Animal Care Co., Ltd.

Human subjects

Human adult adipose tissue samples were collected from health individuals who underwent liposuction surgery (median age: 31 years, range 20-50). Donor information was shown in Supplementary Table 1. The collection and use of human adult adipose tissue samples were approval by the Institute of Ethics Committee Review Board in Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine (SH9H-2023-TK189-1), in accordance with the principles outlined in the declaration of Helsinki. Informed consent was obtained from the healthy donors before adipose tissue donation.

Human umbilical cord blood samples were collected from healthy full-term births at the Shanghai Ninth People’s Hospital following IRB-approved protocols (IRB no. SH9H-2023-TK189-1), or purchased from the Shangdong Qilu Stemcell Engineering co., ltd. All blood samples were received as de-identified, therefore, the information on the age and/or gender of the donors is not available.

Cell and tissue culture

All cell and tissue culture operations were conducted under sterile conditions in laminar flow hoods. Cells or microfat tissues were maintained in an incubator at 37 °C and a constant 5% CO₂ atmosphere.

For single cell expansion, SVFs or ADSCs were seeded at a density of 1 × 10⁵ cells/cm² in 100-mm dishes and cultured in α-MEM supplemented with 10% FBS, 10 mM HEPES, 1 mM sodium pyruvate, and 1% PSG. The medium was changed twice per week. Prior to reaching confluence, cells were dissociated using 0.25% Trypsin-EDTA for passaging purposes. For microfat tissue culture and subsequent differentiation, RMF tissues or RMF pellets were cultured on ultralow attachment cell culture plates in specific medium, with the medium being changed twice per week.

Generation of RMF pellets from human adult adipose tissue

To generate chondroid pellets, 4-mm RMF pellets were generated through a conventional protocol as illustrated in Extended Data Fig. 1a. Adipose tissues, collected in the form of lipoaspirates obtained from healthy donors during liposuction surgeries, were washed, minced, and centrifuged at 1200 g for 3 minutes. The adipose tissues on the top layer were loaded into a 20-ml syringe connected to another syringe via a three-way connector with a filter hole size of 2.4 mm. Then, the adipose tissues were emulsified into microfat tissues by transferring the content between syringes for 30 cycles. The emulsified microfat tissues were subjected to further centrifugation at 1200 g for 3 minutes. After centrifugation, the microfat tissues were purified with a liquid part in the bottom and an oil part on the top. The purified microfat tissues were subsequently seeded onto ultralow attachment 6-well plates with each well containing 1.5 ml of microfat in growth medium, consisting of α-MEM supplemented with 10% FBS, 10 mM HEPES, 1 mM sodium pyruvate, 1% PSG, 10^-5 M ascorbic acid, 10^-7 M dexamethasone, 5 ng/ml FGF-2, and 10 ng/ml PDGF-BB. Following suspension culture for 3 weeks, the fractionated microfat tissues gradually self-reaggregated into RMF tissue constructs. To obtain uniform-sized RMF pellets measuring approximately 4 mm in diameter, biopsy punches with a diameter of 4 mm were employed to extract them from the RMF tissue constructs, which were then placed into 24-well ultralow attachment plates for an additional week of suspension culture.

To generate islet organoids, 1-mm RMF pellets with high cell density and small diameter were generated from a modified protocol as illustrated in Extended Data Fig. 5a. To increase cellular density, the washed and centrifuged adipose tissues were sequentially passed through a one-way connector with a filter hole size of 2.0 mm and another one-way connector with a filter hole size of 1.0 mm for 30 cycles, respectively. Subsequently, the emulsified microfat tissues underwent centrifugation at 1600 g for 3 minutes. Increasing the number of mechanical emulsification steps and applying higher centrifuge force facilitated the disruption and removal of more mature adipocytes, thereby augmenting cell density. The purified microfat tissues could self-reaggregate into RMF tissues comparable to those generated using the conventional protocol after a 2-week suspension culture period. Subsequently, 1-mm RMF pellets were obtained by punching the resulting RMF tissues with a 1-mm biopsy punch, which were subjected to an additional 2 weeks of suspension culture for further enhancement of cell density.

SVF cells extraction and culture

The SVFs were isolated from subcutaneous adipose tissues. The adipose tissues were washed, minced, and digested in 0.075% collagenase NB4 for 45 minutes to 1 hour at 37 °C on an orbital shaker. Subsequently, the suspension was centrifuged at 300 g for 10 minutes, and the resulting SVF pellet was rinsed once with PBS, resuspended in α-MEM supplemented with 10% FBS, and filtered through a 100-μm strainer. Next, red blood cell lysis buffer was added to remove erythrocytes followed by straining the cells again through a 70-μm strainer. Nucleated cells were counted and stored at -150 °C for further use. For single-cell expansion, SVFs were seeded at a density of 1 × 10⁵ cells/cm² in 100-mm dishes and cultured in α-MEM supplemented with 10% FBS and 1% PSG in an incubator at 37 °C and 5% CO₂ with medium changed twice per week.

Osteogenic, adipogenic, and chondrogenic differentiation of SVFs and ADSCs

In vitro adipogenic and chondrogenic differentiation of cells and ADSCs were induced.

Adipogenic differentiation: cells were seeded in 6-well plates at a density of 5 × 10³ cells/cm² and expanded until reaching confluence before initiating differentiation.

For adipogenic differentiation, cells were cultured in adipogenic induction medium for 3 days followed by adipogenic maintenance medium for 1 day. This 4-day cycle was repeated 4 times, after which the cells were further cultured in the adipogenic maintenance medium for an additional week. The composition of adipogenic induction medium included high-glucose DMEM, 10% FBS, 10 mM HEPES, 1 mM sodium pyruvate, 1% PSG, 10 μg/ml insulin, 10^-6 M dexamethasone, 500 μM IBMX, and 100 μM indomethacin. The composition of the adipogenic maintenance medium consisted of high-glucose DMEM, 10% FBS, 10 mM HEPES, 1 mM sodium pyruvate, 1% PSG, and 10 μg/ml insulin. Following adipogenic induction, the cells were washed and stained with Oil Red O staining kit for 15 minutes to visualize lipid droplet formation.

Osteogenic differentiation: cells were cultured in osteogenic induction medium for 3 weeks with medium changed twice per week. The osteogenic induction medium comprised high-glucose DMEM supplemented with 10% FBS, 10 mM HEPES, 1 mM sodium pyruvate, 1% PSG, 10^-5 M ascorbic acid, 10^-8 M dexamethasone, and 10^-2 M β-glycerophosphate. After osteogenic induction, the cells were washed and stained with Alizarin Red solution for 10 minutes.

Chondrogenic differentiation: 5 × 10⁵ cells were seeded onto a type I collagen-based cylindrical porous scaffold (4 mm in diameter, 1 mm in thick; Yierkang Bioengineering Co., Ltd., Beijing, China). The constructs were then cultured in chondrogenic medium for 4 weeks with medium changed twice per week. The chondrogenic medium is composed of high-glucose DMEM, 1.25 mg/ml HSA, 10 mM HEPES, 1 mM sodium pyruvate, 1% PSG, 10^-5 M ascorbic acid, 10^-7 M dexamethasone, 1% ITS, 0.47 mg/ml linoleic acid, 10 ng/ml TGF-β₃, and 10 ng/ml BMP-6 for 4 weeks.

Differentiation of RMF pellets toward chondroid tissues using the two-stage endochondral priming protocol

Differentiations of RMF pellets toward chondroid pellets were performed using a four-stage induction protocol (Fig. 2a) using the following an endochondral ossification approach: stage 1 cartilaginous pellets and stage 2 hypertrophic cartilaginous pellets.

Medium preparation for endochondral priming

Stage 1 medium: High glucose DMEM + 1.25 mg/ml HAS + 10 mM HEPES + 1 mM sodium pyruvate + 1% PSG + 10^-5 M ascorbic acid + 10^-7 M dexamethasone + 1% ITS + 0.47 mg/ml linoleic acid + 10 ng/ml TGF-β₃ + 10 ng/ml BMP-6

Stage 2 medium: High glucose DMEM + 1.25 mg/ml HAS + 10 mM HEPES + 1 mM sodium pyruvate + 1% PSG + 10^-5 M ascorbic acid + 10^-8 M dexamethasone + 10^-2 M β-glycerophosphate.

Differentiation of RMF pellets towards chondroid pellets

The differentiation process was initiated 4 weeks later after self-reaggregation culture by changing to differentiation media. 4-mm RMF pellets were in ultralow attachment 24-well plates (1 pellet per well) in a humidified incubator at 37 °C and 5% CO₂. Media changes were as follows.

Stage 1 (chondrogenic priming): RMF pellets were cultured in stage I chondrogenic priming medium for 4 weeks.

Stage 2 (hypertrophic induction): RMF pellets were changed to stage II hypertrophic induction medium and cultured for 2 weeks. The induction medium was changed twice per week.

Differentiation of RMF pellets toward islet organoids using the four-stage induction protocol

Differentiations of RMF pellets toward islet organoids were performed using a four-stage induction protocol (Fig. 4a) using the following formulations: stage 1 definitive endoderm, stage 2 panarctic progenitor, stage 3 endocrine progenitor, and stage 4 islet organoid.

Medium preparation for islet organoid induction

Basal medium 1: MCDB 131 + 8 mM D-(+)-Glucose + 14 mM NaHCO₃ + 0.1% BSA + 2 mM GlutaMAX + 1% PSG.

Basal medium 2: MCDB 131 + 4 mM D-(+)-Glucose + 14 mM NaHCO₃ + 0.1% BSA + 2 mM GlutaMAX + 1% PSG + 0.25 mM Vitamin C.

Basal medium 3: MCDB 131 + 2 mM D-(+)-Glucose + 24 mM NaHCO₃ + 2% BSA + 2 mM GlutaMAX + 1% PSG + 0.25 mM Vitamin C + 1:50 ITS-X.

Basal medium 4: MCDB 131 + 20 mM D-(+)-Glucose + 24 mM NaHCO₃ + 2% BSA + 2 mM GlutaMAX + 1% PSG + 0.25 mM Vitamin C + 1:50 ITS-X + 10 µg/ml Heparin.

Basal medium 5: MCDB 131 + 2 mM D-(+)-Glucose + 2% BSA + 2 mM GlutaMAX + 1% PSG + 10 µg/ml Heparin + 1% MEM nonessential amino acids

Differentiation of RMF pellets towards islet organoids

The differentiation process was initiated 4 weeks later after self-reaggregation culture by changing to differentiation media. 1-mm RMF pellets were cultured in ultralow attachment 48-well plates (1 pellet per well) in a humidified incubator at 37 °C and 5% CO₂. Media changes were as follows.

Stage 1 (definitive endoderm): RMF pellets were exposed to basal medium 1 supplemented with 100 ng/ml Activin A, 3 μM CHIR99021 for 1 day, then changed to basal medium 1 supplemented with 100 ng/ml Activin A for 3 days.

Stage 2 (pancreatic progenitor): On day 5, RMF pellets were cultured in basal medium 2 supplemented with 50 ng/ml KGF for 2 days. For day 7 and day 8, RMF pellets were cultured in basal medium 3 supplemented with 50 ng/ml KGF, 200 nM LDN193189, 2 μM Retinoic Acid, 0.25 μM SANT1, 0.25 uM TPPB, and 0.3 mM Taurine. For day 9 to day 12, RMF pellets were cultured in basal medium 3 supplemented with 50 ng/ml KGF, 0.1 μM Retinoic Acid, 0.25 μM SANT1, 0.25 uM TPPB, 200 nM LDN193189, and 0.3 mM Taurine.

Stage 3 (endocrine progenitor): RMF pellets were then exposed to basal medium 4 supplemented with 10 μM ALK5i II, 20 ng/ml Betacellulin, 0.1 μM Retinoic Acid, 0.25 μM SANT1, 1 μM T3, 1 μM XXI, and 10 mM Nicotinamide for 7 days.

Stage 4 (islet organoid): On day 21, RMF pellets were cultured in basal medium 5 supplemented with 10 µg/ml ZnSO₄·7H₂O, 1:1000 Trace Elements A, and 1:1000 Trace Elements B for 7 to 14 days.

Differentiation of RMF pellets toward neural tissues using the two-stage induction protocol

Medium preparation for neural tissue differentiation

Differentiations of RMF pellets toward neural tissues were performed using a two-stage induction protocol (Extended Data Fig. 10a) using the following approach: stage 1 neural stem cells and stage 2 neuronal cells or neuroglial cells.

Stage 1 neural stem cell differentiation medium: Neurobasal™ Plus medium + 1% B27, 1% N2 + 20 ng/ml EGF + 20 ng/ml FGF-2.

Stage 2 neural differentiation medium: Neurobasal™ Plus medium + 1% N2 + 1% PSG + 5% horse serum + 0.5 μM ATRA + 20 ng/ml EGF + 20 ng/ml FGF-2. For neuronal differentiation, 10 ng/ml BDNF was added; for glial cell differentiation, 10 ng/ml PDGF-BB was added.

Differentiation of RMF pellets towards neural tissues

The differentiation process was initiated 4-weeks later after self-reaggregation culture by changing to differentiation media. 1-mm RMF pellets were cultured in were cultured on ultralow attachment 48-well plates (1 pellet per well) in a humidified incubator at 5% CO₂ at 37 °C. Media changes were as follows.

Stage 1 neural stem cell differentiation: RMF pellets were cultured in stage 1 neural stem cell differentiation medium for 2 weeks.

Stage 2 neuronal or neuroglial differentiation: RMF pellets were exposed to stage 2 neuronal differentiation medium or glial differentiation medium for 2 weeks, respectively. The induction medium was changed twice per week.

Subcutaneous transplantation of RMF-chondroid pellets into NPSG mice

Immunodeficient NOD/LtJ-Prkdc^scidIl2rg^em1/Shjh (NPSG) female mice, aged 4 weeks, were obtained from Shanghai Jihui Laboratory. For subcutaneous transplantation, RMF-chondroid pellets (4 pellets per mouse) were transplanted under the back skin of mice. RMF-chondroid grafts were explanted at 2, 4, 6, 8, and 12 weeks post-transplantation.

Xenotransplantation of hUCB-derived CD34⁺ cells in NPSG mice

Primary xenotransplantation

The hUCB-CD34⁺ cells were isolated using a CD34 microbead kit, following the manufacturer’s instructions. At 8 weeks post implantation, 5 × 10⁵ hUCB-CD34⁺ cells were pooled from a minimum of 6 donors (in total 15 donors, both females and males) and injected intravenously into NPSG mice 24 hours after sublethal irradiation (Redsource RS2000 XPro; 200 cGy). Mice were euthanized at 8 and 16 weeks post hUCB CD34⁺ cell transplantation. The femurs and hOssicles were collected for subsequent analysis of chimerism, self-renewal, and differentiation (Fig. 3a). The transplantation experiments were repeated 5 times with a cohort of 6 recipient mice (each mouse bearing 4 ossicles) per transplant.

Secondary xenotransplantation

After eight weeks of primary xenotransplantation of hUCB-CD34⁺ cells, the recipient mice were euthanized and the femurs and hOssicles were collected to isolate hCD34⁺ cells using MACS. The isolated hCD34⁺ cells were then plated onto cytokine-containing methylcellulose medium. For secondary xenotransplantation, a secondary batch of NPSG mice, aged 8-10 weeks without pre-established hOssicles, received sublethal irradiation (200 cGy) 24 hours prior to the secondary transplantation. Subsequently, 1 × 10⁶hCD34⁺ cells enriched from mouse femurs or hOssicles were injected intravenously into irradiated NPSG mice, respectively. Eight weeks after the secondary transplantation, the femurs were collected for subsequent chimerism, self-renewal, and differentiation analysis. The secondary transplantation experiments were repeated 3 times with a cohort of 6 recipient mice per transplant.

Kidney subcapsular transplantation of RMF-islet organoid in NOD/SCID mice.

Immunodeficient NOD.Cg-Prkdc^scid/Shjh (NOD/SCID) mice, aged 8-10 weeks, were obtained from Shanghai Jihui Laboratory. For nondiabetic cohorts, RMF-islet organoids (40 to 48 pellets per mouse) were loaded into catheter and transplanted under the kidney capsule of mice. RMF-organoid grafts were explanted by nephrectomy for analysis graft survival and vascularization at 1, 2, and 4 weeks post-transplantation (Extended Data Fig. 7a).

For diabetic cohorts, mice were rendered diabetic after a single intraperitoneal injection of streptozotocin (STZ; 150 mg/kg body weight). Four days after STZ treatment, nonfasted blood glucose was measured from tail vein samples using a OneTouch glucometer. Seven days after STZ treatment, hyperglycaemic mice (blood glucose higher than 200 mg/dL) were selected for further transplantation studies. RMF-islet organoids (40 to 48 pellets per mouse) or undifferentiated RMF pellets (42 to 48 pellets per mouse, derived from the same donors of the islet organoid group) were transplanted under the kidney capsule of mice. The mice underwent sham operation were left as controls. Nonfasted blood glucose and body weight were routinely monitored up to 12 weeks after transplantation. At select times, intraperitoneal glucose tolerance tests were performed following standard protocol. Briefly, after fasting the mice for 16 hours overnight, mice were intraperitoneally injected with glucose (2 g/kg body weight), blood glucose measured were measured at 0, 30, 60, 90, and 120 minutes after glucose challenges. Serum samples before and 30 minutes after glucose challenge were also collected. Serum human insulin levels were quantified using a Human Ultrasensitive Insulin ELISA kit. At 12 weeks after transplantation, nephrectomy was performed to harvest the grafted tissues for section and endocrine markers staining or digested into single cells for 10× scRNA-seq. At 13 weeks after transplantation, mice were euthanized and the recipients’ main organs including pancreas, heart, brain, liver, kidney, spleen, and lung were harvested for gross anatomy and histological analysis (Fig. 5a).

In vitro glucose-stimulated insulin secretion

Islet organoids were washed twice in Krebs buffer (KRB) and pre-incubated in KRB containing low glucose (2 mM) for 2 hours at 37 °C to remove residual insulin. Islet organoids were then challenged with three sequential treatments of alternating low-high-low KRB containing glucose (high; 20 mM), followed by depolarization with KRB containing 30 mM KCl and low glucose (2 mM). Each treatment lasted 30 minutes, after which 100 μl of supernatant was collected and human insulin quantified using the Human Ultrasensitive Insulin ELISA. The islet organoids were enzymatically dissociated into single cells using TrypLE™ Express Enzyme and viable cells were counted using a Countess II (Thermo). Human insulin measurements were normalized by viable cell counts.

Cell proliferation assay

Duo to the heterogeneity of cell populations in SVFs, RMF cells, P0 ADSCs, and expanded ADSCs. Following several pilot studies, different amounts of cells were seeded in 96-well plates to ensure different samples achieved the same adherent cell numbers at 24 hours post cell seeding: 7000 cells per well for RMF cells, 9000 cells for SVF cells, 3000 cells for P0 ADSCs, and 5000 cells for expanded ADSCs. Six wells were used for each group, and the CCK-8 standard curve was fitted according to the measurement results. Moreover, cells were cultured in α-MEM supplemented with 10% FBS, 10 mM HEPES, 1 mM sodium pyruvate, and 1% PSG for 24, 48, 72, and 96 hours. Then, the cells were incubated with a CCK-8 reagent for 2 hours at 37 °C and 5% CO₂. Thereafter, the medium was harvested to measure the optical density values at 450 nm using a microplate reader. The absolute number of cells was calculated by standard curve and the cell proliferation curve was plotted.

SVF and ADSC colony-forming unit assay

To determine colony-forming efficiency, 1 × 10⁴ cells were seeded on 100-mm cell culture plates and cultured in α-MEM supplemented with 10% FBS, 10 mM HEPES, 1 mM sodium pyruvate, and 5 ng/ml FGF-2. After 2 weeks, cultures were rinsed with PBS, fixed with 4% PFA stained with Crystal Violet for 10 minutes and rinsed with tap water. Colonies were imaged and counted.

Hematopoietic colony-forming unit assay

For the colony-forming unit assays, hCD34⁺ cells were sorted from hOssicle and mouse femur bone marrow using MACS. Then, 1 × 10³ hCD34⁺ cells were seeded into 1 ml MethoCult™ H4434 Classic medium and cultured at 37 °C and 5% CO₂ for 12-14 days. Hematopoietic colonies (BFU-E, CFU-GM, and CFU-GEMM) were quantified according to the manufacturer’s instruction.

Flow cytometry

SVF and ADSC subpopulation analysis

SVFs and ADSCs were trypsinized, washed, and filtered through 40-μm cell strainers. Then cells were incubated with Zombie Aqua™ dye and blocked with Human TruStain FcX™ before staining with the following mix of fluorochrome-conjugated mouse anti-human monoclonal antibodies: CD31-Brilliant Violet 605^TM (WM59), CD34-PE/Cyanine7 (581), CD45-PerCP/Cyanine5.5 (HI30), CD73-Brilliant Violet 421^TM (AD2), CD90-APC/Fire^TM 750 (5E10), and CD146-APC (P1H12) for 30 minutes at 4 °C. All antibodies were used at a 1:100 dilution. Samples were acquired on a CytoFLEX LX (Beckman Coulter) then data was analyzed using FlowJo 10. Flow cytometry analysis data were plotted as a percentage of MSCs (gated as CD45^-CD73⁺CD90⁺), endothelial progenitor cells (EPCs, gated as CD45^-CD31⁺CD34⁺), and pericytes (gated as CD45^-CD31^-CD146⁺) distribution among Zombie Aqua™ dye-negative live cells.

Humanized ossicle and mouse femur bone marrow cell analysis

Mice were euthanized and the femurs and hOssicles were collected. Bone marrow cells from mouse femurs and explanted humanized ossicles were isolated by flushing mouse bones with 29G needles and by crushing ossicles with micro pestles, respectively, into tubes containing FACS buffer (DPBS + 2% FBS). Bone marrow cells were incubated with Red Blood Cell Lysis Buffer, stained with Zombie Aqua™ dye and blocked with Human TruStain FcX™. For mouse HSC analysis, bone marrow cells were then stained with the mix of anti-mouse Lineage cocktail-PE (145-2C11; RB6-8C5; RA3-6B2; Ter-119; M1/70), Sca-1-PE/Cyanine7 (D7), c-Kit-APC (2B8), CD135-Brilliant Violet 421TM (A2F10) and CD34-FITC (RAM34) for 30 minutes at 4 °C. For mouse HPC analysis, bone marrow cells were stained with the mix of anti-mouse Lineage cocktail-PE (145-2C11; RB6-8C5; RA3-6B2; Ter-119; M1/70), Sca-1-PE/Cyanine7 (D7), c-Kit-APC (2B8), CD34-FITC (RAM34), CD135-Brilliant Violet 421TM (A2F10), IL-7Rα-APC/Cyanine7 (A7R34), and CD16/32-PerCP/Cyanine5.5 (93) for 30 minutes at 4 °C. All antibodies were used at a 1:100 dilution.

Samples were acquired on a CytoFLEX LX then data was analyzed using FlowJo 10. Flow cytometry analysis data were plotted as a percentage of LT-HSC (gated as L^-S⁺K⁺CD34^-CD135^-), ST-HSC (gated as L^-S⁺K⁺CD34⁺CD135^-), MPP (gated as L^-S⁺K⁺CD34⁺CD135⁺), CLP (gated as Lin^-CD135⁺IL-7Rα⁺), MEP (gated as L^-S^-K⁺CD34^-CD16/32^-), CMP (gated as L^-S^-K⁺CD34⁺CD16/32^-), and GMP (gated as L^-S^-K⁺CD34⁺CD16/32⁺) distribution among Zombie Aqua™ dye-negative live cells.

Transplanted hUCB-CD34⁺ cell analysis

8 weeks and 16 weeks after xenotransplantation of hUCB-CD34⁺ cells, bone marrow cells were isolated from mouse femurs and explanted humanized ossicles, respectively, into tubes containing FACS buffer. After red blood cells lysis with Red Blood Cell Lysis Buffer, bone marrow cells were incubated with Zombie Aqua™ dye and blocked with Human TruStain FcX™. Monoclonal anti-mouse CD45-FITC (30-F11), anti-human CD45-Brilliant Violet 421^TM (HI30) were used to distinguish human from mouse cells. Monoclonal mouse anti human CD34-APC (581), CD38-PE (HB-7), and a human lineage cocktail of antibodies including PE/Cyanine 7-conjugated mouse anti human CD2 (RPA-2.10), CD3 (HIT3a), CD4 (RPA-T4), CD7 (CD7-6B7), CD8a (RPA-T8), CD14 (HCD14), CD16 (3G8), CD19 (HIB19), CD20 (2H7), CD56 (5.1H11), and CD235ab (HIR2) were used to sort human HSPCs. Monoclonal mouse anti human CD3-PE/Cyanine 7 (SK7), CD19-APC (HIB19), and CD33-PE (WM53) were used to analyze human mature blood cells. All antibodies were used at a 1:100 dilution and incubated at 4° C for 30 minutes. Cells were acquired on a CytoFLEX LX then data was analyzed using FlowJo 10. Flow cytometry analysis data were plotted as a percentage of hHSCs (gated as human lineage^-CD34⁺CD38^-), hHPCs (gated as human lineage^-CD34⁺CD38⁺), human T cells (gated as CD3⁺), human B cells (gated as CD3^-CD19⁺CD33^-), and human myeloid cells (gated as CD3^-CD19^-CD33⁺) distribution among donor-derived hCD45⁺ cells.

8 and 16 weeks after the secondary xenotransplantation of mouse femur or humanized ossicle-derived CD34⁺ cells, bone marrow cells from mouse femurs were isolated by flushing mouse bones with 29G needles. The method for processing and staining BM cells as well as strategies for analyzing human HSPCs and mature blood cells are the same as above.

Islet organoid cell analysis

Islet organoids were dissociated by TrypLE™ Express Enzyme at 37 °C for 15-20 minutes and filtered through 40-μm cell strainers for ensuring single-cell suspensions. Cells were incubated with eBioscience™ Fixable Viability Dye eFluor™ 780 and blocked with Human TruStain FcX™. Then cells were fixed and permeabilized with Foxp3/Transcription Factor Fixation/Permeabilization solution according to the manufacturer’s instruction. For Stage 2 analysis, cells were then stained with anti-human PDX-1-PE (658A5) and NKX6.1-Alexa Fluor 488 (R11-560) for 30 minutes at 4 °C. For Stage 3 and Stage 4 analysis, cells were stained with the mix of anti-human C-Peptide-Alexa Fluor 647 (U8-424) and PDX-1-Alexa Fluor 488 (658A5) or the mix of C-Peptide-Alexa Fluor 647 (U8-424) and NKX6.1-Alexa Fluor 488 (R11-560) for 30 minutes at 4 °C. All antibodies were used at a 1:50 dilution. Samples were acquired on a LSR Fortessa X-20 (BD Biosciences) then data was analyzed using FlowJo 10.

Live and dead cell analysis

RMF pellets with diameters of 1, 2, 4 mm were dissociated by 0.15% Collagenase NB4 at 37 °C for 30-40 minutes and filtered through 40-μm cell strainers for ensuring single-cell suspensions. For live and dead cell analysis, cells were incubated with eBioscience™ Fixable Viability Dye eFluor™780 for 30 minutes 4 °C. Samples were acquired on a CytoFLEX LX and then data was analyzed using FlowJo 10.

Magnetic-activated cell sorting MACS

Enrichment of CD34⁺ cells from hUCB and bone marrow

Mouse femurs and explanted ossicles were collected, crushed, flushed, and resuspended into MACS Buffer (DPBS, 0.5% BSA, 2 mM EDTA) and incubated with anti-human CD34 microbeads at 4 °C. After 30 minutes, the samples were rinsed and resuspended into MACS Buffer. The purification was done with the Mini & MidiMACS Starting Kit according to the manufacturer’s instruction.

Dead cell removal

Islet organoids were dissociated by TrypLE™ Express Enzyme at 37 °C for 15-20 minutes and filtered through 40-μm cell strainers for ensuring single-cell suspensions. Cells were washed and resuspended into binding buffer and incubated with Dead Cell Removal Microbeads for 15 minutes at room temperature. The purification was done with the Mini & MidiMACS Starting Kit according to the manufacturer’s instruction.

Mouse cell removal

Islet organoid grafts were explanted from the kidney and then dissociated by TrypLE™ Express Enzyme at 37 °C for 15-20 minutes and filtered through 40-μm cell strainers for ensuring single-cell suspensions. Cells were washed and resuspended into MACS Buffer and incubated with Mouse Cell Depletion Cocktail for 15 minutes at 4 °C. The purification was done with the Mini & MidiMACS Starting Kit according to the manufacturer’s instruction.

Tissue processing and histological analysis

Tissue samples were fixed in 4% PFA for 24 hours at room temperature. In vivo ossicle samples were further decalcified for 7 days in 0.5 mM EDTA, changing the solution twice per week. Next, samples were embedded in paraffin and cut with a microtome in 5-μm sections. The sections were deparaffinized, rehydrated, and stained with hematoxylin and eosin (HE), 0.1% Safranin O/Fast green solutions to detect glycosaminoglycans, Movat’s pentachrome staining kit, or leukocyte acid phosphatase kit for detecting tartrate-resistant acid phosphatase (TRAP) activity.

Immunofluorescence staining

For immunofluorescence staining, slices were blocked with blocking buffer for 1 hour at room temperature and incubated with primary antibodies with appropriate dilutions overnight at 4 °C, followed by washing in PBST (PBS + 0.1% Triton X-100), and incubation with secondary antibodies for 1 hour at room temperature. Nuclei were stained with DAPI for 10 minutes. Antibody details are provided in the key resources table. Representative images were taken using a Leica Stellaris 8 confocal microscope.

Immunohistochemical staining

For immunohistochemical staining, temperature-based (70 °C/160 W, 40 min) or enzymatic digestion-based antigen retrieval was performed at either high or low pH depending on the antibody. Then, following blocked with blocking buffer for 1 hour at room temperature and incubation with primary antibody or appropriate amount of IgG control overnight at 4 °C, slices were washed and incubated with biotinylated secondary antibodies. Chromagen development was achieved by incubating sections with DAB or Vector red according to manufacturer’s instructions. The slices were mounted and photographed using a Zeiss Axio Imager M2 microscope. Antibody details are provided in the key resources table.

Live and dead cell staining

The 1-mm, 2-mm, and 4-mm RMF pellets were stained with the LIVE/DEAD™ Viability/Cytotoxicity Assay Kit according to the manufacturer’s instruction and then washed in PBST. Then, the samples were scanned using a Leica Stellaris 8 confocal microscope.

Tartrate-resistant acid phosphatase (TRAP) staining

TRAP staining was performed in paraffin-embedded tissue sections using a leukocyte acid phosphatase kit. Briefly, deparaffinized slices were immersed in 0.2 M acetate buffer at pH 5.0 containing sodium acetate trihydrate and sodium tartrate for 20 minutes. The slices were then incubated in fresh, pre-warmed staining solution containing 1mg/ml Naphthol AS-MX phosphate powder and 1 mg/ml Fast Red TR Salt 1,5-naphthalenedisulfonate salt at 37 °C for approximately 2 hours. Upon visual confirmation of bright red color development, the slices were rinsed and subsequently counterstained with Mayer’s Hematoxylin solution for 1 minute. Slices were mounted and photographed using a Zeiss Axio Imager M2 microscope.

Confocal imaging, analysis, and quantification

Stained tissue and organoid slices were imaged on a Stellaris 8 confocal microscope or a Zeiss Axio Imager M2 microscope. TIFF files of the merged images were processed and analyzed for measurement of mean fluorescence intensity and manual counting of positive cells for any given markers using ImageJ software. For quantification of cell distribution, regions of interest (ROIs) were defined to divide the organoids in bins, and number of positive cells, mean fluorescence intensity, and area for any given ROI was determined. Similarly, for quantification of cell abundance, positive cells for each marker were manually counted and normalized over DAPI⁺ cells. Quantifications were performed across multiple regions and multiple organoids using at least n = 3 biological replicates. Quantification was performed by two independent investigators.

RNA isolation and reverse transcription (RT)-quantitative PCR (qPCR)

RNA isolation of primary cells, tissues, or organoids were performed following the manufacturer’s instructions. Briefly, total RNA was isolated using the SteadyPure Quick RNA Extraction Kit according to the manufacture’s instruction. Extracted RNA was reverse transcribed into cDNA using the Primerscript RT master kit. cDNA was stored at -20 °C or -70 °C for long time storage.

qPCR was prepared with TB Green Premix Ex Taq kit and detected by Applied Biosystems QuantStudio 3 (Thermo). Raw data were analyzed by ABI Prism 7000 SDS Software (Applied Biosystems). Comparative CT method (2^-^∆∆^CT method) was chosen to evaluate differentially expressed genes. All relative expression levels were normalized to the housekeeping gene GAPDH. A list of the primer sequences for qPCR were listed in Table S2.

DNA content quantification

The DNA content of RMF tissues or pellets were quantified using an Animal Tissues/Cells Genomic DNA Extraction Kit according to the manufacturer’s instruction. The amount of DNA was detected by the NanoDrop Lite Plus Spectrophotometer (Thermo). The DNA content was calculated by dividing the amount of DNA of each RMF sample by the volume of samples.

Scanning electron microscopy (SEM)

Samples were fixed with 0.25% glutaraldehyde overnight at 4 °C and washed with PBS. The samples were gradually dehydrated with 30%, 50%, 70%, 90%, and 100% ethanol, coated with gold and imaged with a microscope (ZEISS GeminiSEM 300).

Transmission electron microscopy (TEM)

To analyze granular ultrastructure, islet organoids were fixed with a mixture containing 1.25% PFA, 2.5% glutaraldehyde, and 0.03% picric acid in 0.1 M sodium cacodylate buffer (pH 7.4) at 4 °C for overnight. Samples were then washed in 0.1 M cacodylate buffer and post fixed at room temperature with a mixture of 1% OsO₄/1.5% KFeCN₆ once for 2 hours then once for 1 hour. After washing with water, samples were stained in 1% aqueous uranyl acetate for 1 hour, washed, and subsequently dehydrated in gradual ethanol (30% to 100%) and propylene oxide. Then, samples were embedded in Epon 812. The samples were sectioned by diamond knife (Leica EM UC7) to 70-90nm, followed by stained with uranyl acetate and lead citrate for contrast. All the specimens were viewed by transmission electron microscope (HITACHI H-7650).

Microcomputed tomography (microCT)

Following fixation in PFA and storage in PBS, microCT data were acquired from explanted ossicles by using a high-resolution scanner (SkyScan1172, Skyscan, Belgium) and 0.5-mm aluminum filtered X-rays (applied voltage 50 kV; current, 200 mA). Transmission images were acquired during a 360° scan rotation with an incremental rotation step size of 0.25°. Reconstruction was performed using a modified Feldkamp algorithm at an isotropic voxel size of 2.5 μm. 3D rendering of the structures was performed using NRecon software (Bruker microCT, Kontich, Belgium) and CTvox (Bruker microCT, Kontich, Belgium) and analyzed by the program CTAn (Bruker microCT, Kontich, Belgium).

Single-cell sequencing and analysis

Single-cell sample preparation

The tissue samples were washed and digested using collagenase and incubated in a shaking water bath set at 37 °C for 60 minutes to facilitate optimal cell dissociation. Subsequently, debris and dead cell removal was performed to eliminate any remaining tissue fragments and acquire a single-cell suspension with high cell viability. Then, cell count and viability was estimated using fluorescence Cell Analyzer (Countess II, Thermo) with AO/PI reagent. Finally, the harvested cells were washed twice with PBS and then resuspended in PBS with 0.04% BSA at a density of 1 × 10⁶ cells per ml.

Single-cell RNA-seq library construction and sequencing

Single-cell RNA-Seq libraries were prepared using a 10× Chromium Next Gel Bead-in-Emulsion (GEM) Single Cell 3’ Reagent Kits v3.1 from 10× Genomics as instructions. Briefly, cells were mixed with reverse transcription reagent and then loaded into the sample well in Chromium Next GEM Chip G. Subsequently Gel Beads and Partitioning Oil were dispensed into corresponding wells separately. After emulsion droplet were formed, reverse transcription was performed. The resulting cDNA was then purified, amplified, cleaned, and ligated to sequencing adaptors. Finally, the indexed PCR was employed to amplify the DNA fragments. The indexed sequencing libraries were cleared with SPRI beads, quantified by qPCR, and ultimately sequenced on a NovaSeq 6000 platform (Illumina) with PE150 read length.

Single-cell RNA sequencing data processing

The Cell Ranger Single-cell Software Suite⁶¹ was used to perform sample demultiplexing, barcode processing, and single-cell 3ʹ gene counting. First, UMI tags and barcode sequences were extracted from Read1, followed by alignment of Read2 (containing the cDNA insert) to a reference genome using STAR⁶². Next, Barcodes and UMIs were filtered based on quality. Last, PCR duplicates were identified by shared barcode sequences, UMI tags, and gene IDs. Confidently mapped non-PCR duplicates with valid barcodes and UMIs were utilized to generate the gene-barcode matrix. Cell barcodes were determined, distinguishing between cells and empty partitions. Meaningful read counts were calculated based on valid barcodes, UMIs, and mapped reads.

Single-cell RNA transcriptome analysis

The clustering and visualization were finished by Python with methods in the package Scanpy (v1.10.1)⁶³. Quality control measures included removing potential doublets with the Scrublet software (v0.2.3)⁶⁴ and filtering out cells with fewer than 2,500 UMIs or fewer than 1,000 genes per cell. Additionally, genes detected in fewer than three cells were excluded. The counts per cell were normalized to reach a total of 10,000 counts per cell and then subjected to a logarithmic transformation. Cells were grouped into clusters by identifying highly variable genes. Subsequently, adjustments were made for total UMI count per cell and expression of mitochondrial genes through regression analysis, followed by standardization of the resulting residuals. Principal component analysis was employed to reduce dimensionality to 50 principal components. A nearest neighbor graph was constructed based on these components, enabling the application of Leiden clustering to segregate cells into distinct clusters.

Ternary plot was used to position cells based on the proportions of signature genes expressed in native fat tissue, RMF pellets, and expanded ADSCs. The MDS plot was created using the vcd package with quantile normalized log2-CPM values. The log2-CPM values for each gene across cells were calculated with a prior count of 1, using the cpkm function from edgeR.

Trajectory analysis was carried out using Slingshot (v1.2.0)⁶⁵. After selecting the subpopulations, the pseudotime values were calculated using slingshot. Then, genes associated with the pseudotime direction were calculated using the ‘associationTest’ in tradeSeq. Finally, the gene list was filtered for plotting based on p-value, meanLogFC, and whether the genes are differential between subpopulations.

The ‘AddModuleScore’ function from the Seurat (v4.4.0)⁶⁶ package was employed to calculate gene expression module scores for each cell. Cells within the same cluster showed comparable module scores, indicating similar gene expression profiles and likely analogous cellular functions or states. The signal distribution within each cluster was analyzed and visualized using violin plots with the ‘VlnPlot’ function in Seurat.

QUANTIFICATION AND STATISTICAL ANALYSIS

All data are presented as means ± SD and were analyzed by GraphPad Prism 9 (GraphPad Software). Paired or unpaired two-tailed Student’s t tests were performed when two groups of samples were compared. One-way or two-way ANOVA with Tukey’s or Dunnett’s tests were performed when multiple groups were compared. Log-rank (Mantel-Cox) test is used for comparison of survival rate between multiple groups. p values are indicated in the figures as: ns; p > 0.05; ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001; ∗∗∗∗ p < 0.0001. Statistical details for each experiment can be found in the figures and the legends.

ACKNOWLEDGMENTS

This research was supported by grants from the National Natural Science Foundation of China (82372535 and 82361138568), Shanghai Clinical Research Center of Plastic and Reconstructive Surgery supported by Science and Technology Commission of Shanghai Municipality (22MC1940300), and Shanghai Plastic Surgery Research Center of Shanghai Priority Research Center (2023ZZ02023). We thank Dr. Haoran Wang for invaluable support. Some illustrations in the graphic abstract and figures were generated using BioRender.

AUTHOR CONTRIBUTIONS

Conceptualization, R.L.H. and Q.L.; methodology and experimental design, R.L.H., J.Y., and Y.Y.; investigation, J.Y., Y.Y., X.L., X.Y., C.L. R.A., and R.L.H.; data analysis, J.Y., Y.Y., X.L., X.Y., C.L. R.A., R.L.H., T.Z., Q.L., and K.Z.; writing – original draft, R.L.H.; writing – review & editing, R.L.H., T.Z., Q.L., and K.Z.; supervision, R.L.H. T.Z., and Q.L..

DECLARATION OF INTERESTS

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Correspondence and requests for materials should be addressed to Q.L.

Zhao, Z. et al. Organoids. Nat Rev Methods Primers 2 (2022). https://doi.org/10.1038/s43586-022-00174-y
Dao, L. et al. Modeling blood-brain barrier formation and cerebral cavernous malformations in human PSC-derived organoids. Cell Stem Cell 31, 818-833 e811 (2024). https://doi.org/10.1016/j.stem.2024.04.019
Huang, B. et al. Long-term expandable mouse and human-induced nephron progenitor cells enable kidney organoid maturation and modeling of plasticity and disease. Cell Stem Cell 31, 921-939 e917 (2024). https://doi.org/10.1016/j.stem.2024.04.002
Khan, A. O. et al. Human Bone Marrow Organoids for Disease Modeling, Discovery, and Validation of Therapeutic Targets in Hematologic Malignancies. Cancer Discov 13, 364-385 (2023). https://doi.org/10.1158/2159-8290.CD-22-0199
Huang, X. et al. Stomach-derived human insulin-secreting organoids restore glucose homeostasis. Nat Cell Biol 25, 778-786 (2023). https://doi.org/10.1038/s41556-023-01130-y
Wang, D. et al. Long-Term Expansion of Pancreatic Islet Organoids from Resident Procr(+) Progenitors. Cell 180, 1198-1211 e1119 (2020). https://doi.org/10.1016/j.cell.2020.02.048
Hicks, M. R. & Pyle, A. D. The emergence of the stem cell niche. Trends Cell Biol 33, 112-123 (2023). https://doi.org/10.1016/j.tcb.2022.07.003
Fuchs, E. & Blau, H. M. Tissue Stem Cells: Architects of Their Niches. Cell Stem Cell 27, 532-556 (2020). https://doi.org/10.1016/j.stem.2020.09.011
Beumer, J. & Clevers, H. Hallmarks of stemness in mammalian tissues. Cell Stem Cell 31, 7-24 (2024). https://doi.org/10.1016/j.stem.2023.12.006
Merrell, A. J. & Stanger, B. Z. Adult cell plasticity in vivo: de-differentiation and transdifferentiation are back in style. Nat Rev Mol Cell Biol 17, 413-425 (2016). https://doi.org/10.1038/nrm.2016.24
Yap, L., Tay, H. G., Nguyen, M. T. X., Tjin, M. S. & Tryggvason, K. Laminins in Cellular Differentiation. Trends Cell Biol 29, 987-1000 (2019). https://doi.org/10.1016/j.tcb.2019.10.001
Lukjanenko, L. et al. Loss of fibronectin from the aged stem cell niche affects the regenerative capacity of skeletal muscle in mice. Nat Med 22, 897-905 (2016). https://doi.org/10.1038/nm.4126
Chen, M. et al. Identification of an adipose tissue-resident pro-preadipocyte population. Cell Rep 42, 112440 (2023). https://doi.org/10.1016/j.celrep.2023.112440
Merrick, D. et al. Identification of a mesenchymal progenitor cell hierarchy in adipose tissue. Science 364 (2019). https://doi.org/10.1126/science.aav2501
Dupard, S. J., Grigoryan, A., Farhat, S., Coutu, D. L. & Bourgine, P. E. Development of Humanized Ossicles: Bridging the Hematopoietic Gap. Trends Mol Med 26, 552-569 (2020). https://doi.org/10.1016/j.molmed.2020.01.016
de Janon, A., Mantalaris, A. & Panoskaltsis, N. Three-Dimensional Human Bone Marrow Organoids for the Study and Application of Normal and Abnormal Hematoimmunopoiesis. J Immunol 210, 895-904 (2023). https://doi.org/10.4049/jimmunol.2200836
Reinisch, A., Hernandez, D. C., Schallmoser, K. & Majeti, R. Generation and use of a humanized bone-marrow-ossicle niche for hematopoietic xenotransplantation into mice. Nat Protoc 12, 2169-2188 (2017). https://doi.org/10.1038/nprot.2017.088
Reinisch, A. et al. A humanized bone marrow ossicle xenotransplantation model enables improved engraftment of healthy and leukemic human hematopoietic cells. Nat Med 22, 812-821 (2016). https://doi.org/10.1038/nm.4103
Reinisch, A. et al. Epigenetic and in vivo comparison of diverse MSC sources reveals an endochondral signature for human hematopoietic niche formation. Blood 125, 249-260 (2015). https://doi.org/10.1182/blood-2014-04-572255
Yang, J. et al. Bioengineering and vascularization strategies for islet organoids: advancing toward diabetes therapy. Metabolism 152, 155786 (2024). https://doi.org/10.1016/j.metabol.2024.155786
Lee, S. A. et al. Direct Differentiation of Bone Marrow Mononucleated Cells Into Insulin-Producing Cells Using 4 Specific Soluble Factors. Stem Cells Transl Med 12, 485-495 (2023). https://doi.org/10.1093/stcltm/szad035
Jeong, J. H. et al. Self-organized insulin-producing beta-cells differentiated from human omentum-derived stem cells and their in vivo therapeutic potential. Biomater Res 27, 82 (2023). https://doi.org/10.1186/s40824-023-00419-1
Hogrebe, N. J., Maxwell, K. G., Augsornworawat, P. & Millman, J. R. Generation of insulin-producing pancreatic beta cells from multiple human stem cell lines. Nat Protoc 16, 4109-4143 (2021). https://doi.org/10.1038/s41596-021-00560-y
Liu, H. et al. Chemical combinations potentiate human pluripotent stem cell-derived 3D pancreatic progenitor clusters toward functional beta cells. Nat Commun 12, 3330 (2021). https://doi.org/10.1038/s41467-021-23525-x
Augsornworawat, P. et al. Single-nucleus multi-omics of human stem cell-derived islets identifies deficiencies in lineage specification. Nat Cell Biol 25, 904-916 (2023). https://doi.org/10.1038/s41556-023-01150-8
Wu, J. et al. Treating a type 2 diabetic patient with impaired pancreatic islet function by personalized endoderm stem cell-derived islet tissue. Cell Discov 10, 45 (2024). https://doi.org/10.1038/s41421-024-00662-3
Balboa, D. et al. Functional, metabolic and transcriptional maturation of human pancreatic islets derived from stem cells. Nat Biotechnol 40, 1042-1055 (2022). https://doi.org/10.1038/s41587-022-01219-z
Yoshihara, E. et al. Immune-evasive human islet-like organoids ameliorate diabetes. Nature 586, 606-611 (2020). https://doi.org/10.1038/s41586-020-2631-z
Hogrebe, N. J., Augsornworawat, P., Maxwell, K. G., Velazco-Cruz, L. & Millman, J. R. Targeting the cytoskeleton to direct pancreatic differentiation of human pluripotent stem cells. Nat Biotechnol 38, 460-470 (2020). https://doi.org/10.1038/s41587-020-0430-6
Maxwell, K. G. et al. Gene-edited human stem cell-derived beta cells from a patient with monogenic diabetes reverse preexisting diabetes in mice. Sci Transl Med 12 (2020). https://doi.org/10.1126/scitranslmed.aax9106
Augsornworawat, P., Maxwell, K. G., Velazco-Cruz, L. & Millman, J. R. Single-Cell Transcriptome Profiling Reveals beta Cell Maturation in Stem Cell-Derived Islets after Transplantation. Cell Rep 32, 108067 (2020). https://doi.org/10.1016/j.celrep.2020.108067
Velazco-Cruz, L. et al. Acquisition of Dynamic Function in Human Stem Cell-Derived beta Cells. Stem Cell Reports 12, 351-365 (2019). https://doi.org/10.1016/j.stemcr.2018.12.012
Maxwell, K. G., Kim, M. H., Gale, S. E. & Millman, J. R. Differential Function and Maturation of Human Stem Cell-Derived Islets After Transplantation. Stem Cells Transl Med 11, 322-331 (2022). https://doi.org/10.1093/stcltm/szab013
Aghazadeh, Y. et al. Microvessels support engraftment and functionality of human islets and hESC-derived pancreatic progenitors in diabetes models. Cell Stem Cell 28, 1936-1949 e1938 (2021). https://doi.org/10.1016/j.stem.2021.08.001
Qin, Y. et al. An Update on Adipose-Derived Stem Cells for Regenerative Medicine: Where Challenge Meets Opportunity. Adv Sci (Weinh) 10, e2207334 (2023). https://doi.org/10.1002/advs.202207334
Sakers, A., De Siqueira, M. K., Seale, P. & Villanueva, C. J. Adipose-tissue plasticity in health and disease. Cell 185, 419-446 (2022). https://doi.org/10.1016/j.cell.2021.12.016
Yeola, A. et al. Induction of muscle-regenerative multipotent stem cells from human adipocytes by PDGF-AB and 5-azacytidine. Sci Adv 7 (2021). https://doi.org/10.1126/sciadv.abd1929
Li, Y. et al. Compression-induced dedifferentiation of adipocytes promotes tumor progression. Sci Adv 6, eaax5611 (2020). https://doi.org/10.1126/sciadv.aax5611
Hendriks, D. et al. Human fetal brain self-organizes into long-term expanding organoids. Cell 187, 712-732.e738 (2024). https://doi.org/10.1016/j.cell.2023.12.012
Jacob, F. et al. A Patient-Derived Glioblastoma Organoid Model and Biobank Recapitulates Inter- and Intra-tumoral Heterogeneity. Cell 180, 188-204 e122 (2020). https://doi.org/10.1016/j.cell.2019.11.036
Grigoryan, A. et al. Engineering human mini-bones for the standardized modeling of healthy hematopoiesis, leukemia, and solid tumor metastasis. Sci Transl Med 14, eabm6391 (2022). https://doi.org/10.1126/scitranslmed.abm6391
Altrock, E. et al. Humanized three-dimensional scaffold xenotransplantation models for myelodysplastic syndromes. Exp Hematol 107, 38-50 (2022). https://doi.org/10.1016/j.exphem.2021.12.395
Zhang, J. et al. 3D Printing of a Vascularized Mini‐Liver Based on the Size‐Dependent Functional Enhancements of Cell Spheroids for Rescue of Liver Failure. Advanced Science 11 (2024). https://doi.org/10.1002/advs.202309899
Wang, Z. et al. Molecular mechanism underlying miR-204-5p regulation of adipose-derived stem cells differentiation into cells from three germ layers. Cell Death Discovery 10 (2024). https://doi.org/10.1038/s41420-024-01852-4
Jin, Y. et al. Stem cell-derived hepatocyte therapy using versatile biomimetic nanozyme incorporated nanofiber-reinforced decellularized extracellular matrix hydrogels for the treatment of acute liver failure. Bioact Mater 28, 112-131 (2023). https://doi.org/10.1016/j.bioactmat.2023.05.001
Taguchi, T. et al. Feline Adipose Derived Multipotent Stromal Cell Transdifferentiation Into Functional Insulin Producing Cell Clusters. Front Bioeng Biotechnol 10, 904519 (2022). https://doi.org/10.3389/fbioe.2022.904519
Ghoneim, M. A. et al. Transplantation of insulin-producing cells derived from human mesenchymal stromal/stem cells into diabetic humanized mice. Stem Cell Res Ther 13, 350 (2022). https://doi.org/10.1186/s13287-022-03048-y
Yao, X. et al. Non-electric bioelectrical analog strategy by a biophysical-driven nano-micro spatial anisotropic scaffold for regulating stem cell niche and tissue regeneration in a neuronal therapy. Bioact Mater 20, 319-338 (2023). https://doi.org/10.1016/j.bioactmat.2022.05.034
Wang, J. et al. Rebuilding hippocampus neural circuit with hADSC-derived neuron cells for treating ischemic stroke. Cell Biosci 12, 40 (2022). https://doi.org/10.1186/s13578-022-00774-x
Yang, L., Shen, X. M., Wang, Z. F., Li, K. & Wang, W. The Notch signalling pathway and miRNA regulation play important roles in the differentiation of Schwann cells from adipose-derived stem cells. Lab Invest 102, 320-328 (2022). https://doi.org/10.1038/s41374-021-00687-2
Wang, W. et al. Let-7a-5p regulated by lncRNA-MEG3 promotes functional differentiation to Schwann cells from adipose derived stem cells via directly inhibiting RBPJ-mediating Notch pathway. Apoptosis 26, 548-560 (2021). https://doi.org/10.1007/s10495-021-01685-x
Li, M. et al. Epithelial differentiation of human adipose-derived stem cells (hASCs) undergoing three-dimensional (3D) cultivation with collagen sponge scaffold (CSS) via an indirect co-culture strategy. Stem Cell Res Ther 11, 141 (2020). https://doi.org/10.1186/s13287-020-01645-3
Petry, L. et al. Directing adipose‐derived stem cells into keratinocyte‐like cells: impact of medium composition and culture condition. Journal of the European Academy of Dermatology and Venereology 32, 2010-2019 (2018). https://doi.org/10.1111/jdv.15010
Bandeira, F., Goh, T. W., Setiawan, M., Yam, G. H. & Mehta, J. S. Cellular therapy of corneal epithelial defect by adipose mesenchymal stem cell-derived epithelial progenitors. Stem Cell Res Ther 11, 14 (2020). https://doi.org/10.1186/s13287-019-1533-1
Fontcuberta-PiSunyer, M. et al. Direct reprogramming of human fibroblasts into insulin-producing cells using transcription factors. Commun Biol 6, 256 (2023). https://doi.org/10.1038/s42003-023-04627-2
Leite, N. C. et al. Modeling Type 1 Diabetes In Vitro Using Human Pluripotent Stem Cells. Cell Rep 32, 107894 (2020). https://doi.org/10.1016/j.celrep.2020.107894
Liu, G. et al. Charting a high-resolution roadmap for regeneration of pancreatic beta cells by in vivo transdifferentiation from adult acinar cells. Sci Adv 9, eadg2183 (2023). https://doi.org/10.1126/sciadv.adg2183
Cerda-Esteban, N. et al. Stepwise reprogramming of liver cells to a pancreas progenitor state by the transcriptional regulator Tgif2. Nat Commun 8, 14127 (2017). https://doi.org/10.1038/ncomms14127
Liang, Z. et al. Implantation underneath the abdominal anterior rectus sheath enables effective and functional engraftment of stem-cell-derived islets. Nat Metab 5, 29-40 (2023). https://doi.org/10.1038/s42255-022-00713-7
Du, Y. et al. Human pluripotent stem-cell-derived islets ameliorate diabetes in non-human primates. Nat Med 28, 272-282 (2022). https://doi.org/10.1038/s41591-021-01645-7
Zheng, G. X. et al. Massively parallel digital transcriptional profiling of single cells. Nat Commun 8, 14049 (2017). https://doi.org/10.1038/ncomms14049
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15-21 (2013). https://doi.org/10.1093/bioinformatics/bts635
Wolf, F. A., Angerer, P. & Theis, F. J. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol 19, 15 (2018). https://doi.org/10.1186/s13059-017-1382-0
Wolock, S. L., Lopez, R. & Klein, A. M. Scrublet: Computational Identification of Cell Doublets in Single-Cell Transcriptomic Data. Cell Syst 8, 281-291 e289 (2019). https://doi.org/10.1016/j.cels.2018.11.005
Street, K. et al. Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC Genomics 19, 477 (2018). https://doi.org/10.1186/s12864-018-4772-0
Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol 36, 411-420 (2018). https://doi.org/10.1038/nbt.4096

There is NO Competing Interest.

Download PDF

Version 1

posted

You are reading this latest preprint version

Direct Differentiation of Human Adult Adipose Tissue into Multilineage Functional Organoids

Status:

Version 1

Abstract

Figures

INTRODUCTION

RESULTS

The in vitro self-reaggregation process enhances the multipotent differentiation capacity of microfat tissue

Generation of humanized bone marrow organoids from RMF pellets

RMF-derived bone marrow organoids support human normal hematopoiesis in immunodeficient mice

Generation of insulin-producing organoids from RMF pellets

RMF-derived islet organoids acquire rapid graft vascularization and reverse diabetes in mice

Generation of neural-like tissues from RMF pellets in vitro

DISCUSSION

METHODS

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1