Establishing a multicellular model by three-dimensional cell-assembly technique for metabolic syndrome

One of the major obstacles in developing multifunctional drugs for the metabolic syndrome (MS) is lack of in vitro models that capture more complex features of the disease. Here we give the first report that establishes an energy metabolic system model using cell-assembly technique which can assemble cells into designated places to form complex three-dimensional structures. Adipose-derived stromal cells were assembled and induced differentiation into adipocytes and endothelial cells; pancreatic islets were then deposited at designated locations and constituted adipoinsular axis with adipocytes. Analysis of the factors involved in energy metabolism showed our system could capture more physiological and pathophysiological features of the in vivo energy metabolism. Drugs known to have effects on MS showed accordant effects in the systems. Construction and study of such multicellular systems could help us better understand pathogenesis of MS, develop new technologies for drug discovery, and foster applications in tissue engineering and metabolomics profiling. Researchers have recently developed techniques to fabricate tissues and organs in which both cells and biomaterials have carefully defined architectures. Three-dimensional (3-D) bioassembly tool is capable of extruding cells and biomaterials into spatially organized, 3-D constructs . Cell printing equipments can print cells as a stream of drops in 3-D positions that mimic their respective positions in organs . We have recently developed a 3-D cell-assembly technique N at ur e P re ce di ng s : h dl :1 01 01 /n pr e. 20 08 .1 49 6. 1 : P os te d 8 Ja n 20 08 based on Rapid Prototyping (RP) . This technique can put different cells and materials into designated places to form complex 3-D structures. The designed architecture facilitates cell growth, organization and differentiation. Hepatocytes have been assembled with gelatin hydrogel to build 3-D structures, in which the viabilities and functions of the cells could remain more than 60 days . Tissue engineering endeavored to manufacture tissues and organs, but pay little attention to establish physiological system models. However, in study of complex physiological processes, there is an increasing demand for in vitro 3-D models that can capture more of the relevant complexity than what traditional two-dimensional (2-D) cultures achieve . In the field of drug discovery, high-throughput screening is ineffective due to its failure in capturing complex pathological features in vivo environment . More recently, some in vitro 3-D models have been applied in research pattern formation, tumor growth and apoptosis, but limited to study of simple physiologic processes by slow spontaneous aggregation of spheroids with poorly controlled structure. Metabolic syndrome (MS) is a cluster of growing epidemic diseases including obesity, diabetes, hypertension, and atherosclerosis. Recently, the pharmaceutical industry has shown great interest in developing drugs that can target MS as a whole . Researchers including ourselves have devoted to find out drug targets and establish models for MS. However, there are still no approved drugs that can reliably reduce all of the metabolic risk factors over the long term. One of the major obstacles is the lack of in vitro models that can capture more pathological features of MS. MS is largely presented as energy metabolic turbulence and cardiovascular disease (CVD); adipocytes and β-cells usually constitute adipoinsular axis to regulate energy metabolism; and endothelial cell dysfunction can connect the pathogenesis of energy metabolic turbulence with CVD. With these mechanisms in mind we attempt to organize the related cells to establish MS models. Here we report the feasibility of constructing an in vitro multicellular 3-D model of energy metabolic system for MS, using our cell-assembly technique. Adipose-derived stromal (ADS) cells were assembled in a well designated 3-D structure and were controlled differentiation into adipocytes and endothelial cells based on their respective positions within the structure; pancreatic islets were then deposited at designated micro-holes. We also tested the factors N at ur e P re ce di ng s : h dl :1 01 01 /n pr e. 20 08 .1 49 6. 1 : P os te d 8 Ja n 20 08 involved in energy metabolism and endothelial dysfunction of the multicellular model, tested whether chronic exposure to high glucose, a major inducement of MS, could lead to similar pathological changes of the multicellular model to MS, and tested whether drugs known to have effects on MS manifest accordant effects in the multicellular system. RESULTS Assembling of 3-D multicellular system Considering the important roles of adipocytes, β-cells and endothelial cells in energy metabolism and MS pathogenesis, we assumed that these types of cell would be required for construction of model for MS. We selected ADS cells isolated from rat, which can differentiate into adipocytes and endothelial cells. After several comparative experiments about biocompatible materials, we mixed ADS cells with gelatin, alginate and fibrinogen solutions sequentially. A software package was employed to fabricate the complex structure model with orderly channels (Fig. 1 a). Following the designed model, a nozzle controlled by computer was used to deposit the mixture on a glass chip at a temperature of 10 °C (gelatin at gel state) and generate 3-D structures (Fig. 1 b). Then, alginate was crosslinked with CaCL2 and the fibrinogen with thrombin. Once the mixture was crosslinked, the whole construct could be handled easily without losing its integrity; cells were homogeneous embedded in the matrices (Fig. 1c, d), and the hydrogel state could remain at least 8 weeks. The complete process required only ~20 min. Scanning electron micrographs showed the development of extensive extracellular matrix (ECM) and cell networks in the structure after 6 days of culture (Fig. 1d). The pancreatic islets (contain β-cells) from rat were deposited to designated micro holes in the 3-D structure at the 7 d. Anti-insulin immunostaining showed that 90% of the pancreatic islets could keep their integrity of globoid shape more than three weeks in the structure (Fig. 1e ). Some envelopes of islet were broken, but the β-cells still congregated and kept secreting insulin. (Fig. 1f). To test whether ADS cells in the 3-D structure can be controlled differentiate into endothelial cells and adipocytes, we examined composition and distribution of the cells in the 3-D structure by immunostaining and Oil Red O staining. After 3 days of culture with endothelial growth factor (EGF), CD34 (ADS cells and endothelial cells indicator) and CD31 (mature endothelial cells indicator) staining revealed that over 90% of ADS cells on the walls of the channels were differentiated into mature endothelial cells (CD31+/CD34+) and connected with each other N at ur e P re ce di ng s : h dl :1 01 01 /n pr e. 20 08 .1 49 6. 1 : P os te d 8 Ja n 20 08 forming vessel-like structure. (Fig. 2a, c, e, f). This differentiation of ADS cells was based on EGF inducing (Fig2 b, d) and cell position in the 3-D structure (Fig. 2e, f). From the 4 day, the structures were treated with insulin, IBMX and dexamethasone for 3 d. At the 12 day, Oil red O staining revealed that ADS cells in the structure differentiated into adipocytes with a spherical shape (Fig. 2g). The cells under the walls of the channels were more sensitive to differentiation into adipocytes than those on the surface of the walls. However, if not induced with EGF, more than 90% of cells on the walls of channels would differentiate into adipocytes (Fig. 2h). Dynamic insulin secretion To determine whether β-cells assembled in the system can mimic in vivo function, we used a perfusion system for a time-dependent estimation of insulin response to glucose stimulation. Figure 3 shows the dynamic insulin release patterns of free or assembled β-cells stimulated by glucose. In response to 15 mM glucose stimulation, assembled β-cells showed a significant increase in insulin secretion compared to free β-cells. To determine whether chronic exposure to high glucose could cause pathological changes of the β-cells, we culture free or assembled β-cells with high-glucose (15 mM). After 5 days of culture, both free β-cells and assembled β-cells showed dramatic decreases in glucose-induced insulin release, and the decrease rate of the assembled β-cells was higher than that of the free β-cells. Nateglinide could revise the insulin secretion decrease of the assembled and free β-cells; whereas rosiglitazone only had the same effects on the assembled β-cells. Notably, after pre-cultured with high glucose, the glucose-induced insulin secretion peak of the assembled β-cells was delayed compared to normal control, this delay could be revised by nateglinide, however, this phenomenon was not observed in the free β-cells. Derangements in the kinetics of insulin release, including the decrease and delay of the secretion peak value, were one of main characteristics of diabetes. These results are consistent with the other in vivo experiments, suggesting that the adipocytes in the 3-D structure could help the β-cells to capture more pathologic feature of MS. Further investigation of adipocytokine secretion would subsequently reveal the mechanism of this phenomenon. Glucose consumption, FFA release, and Adipogenesis Glucose consumption and free fatty acids (FFA) mobilization of adipocytes are two basic metabolic processes in energy metabolism. To determine the glucose consumption of cells in the N at ur e P re ce di ng s : h dl :1 01 01 /n pr e. 20 08 .1 49 6. 1 : P os te d 8 Ja n 20 08 2-D, 3-D (without β-cells) and multicellular 3-D (with β-cells) culture systems, we culture the cells with 25 mM glucose, and thereafter the media were collected for glucose consumption assays (Fig. 4a). Cells in the 3-D structure showed higher glucose consumption than in the 2-D