A metabolic dysfunction-associated steatotic liver acinus biomimetic induces pancreatic islet dysfunction in a coupled microphysiology system

Abstract

Preclinical and clinical studies suggest that lipid-induced hepatic insulin resistance is a primary defect that predisposes to dysfunction in islets, implicating a perturbed liver-pancreas axis underlying the comorbidity of T2DM and MASLD. To investigate this hypothesis, we developed a human biomimetic microphysiological system (MPS) coupling our vascularized liver acinus MPS (vLAMPS) with pancreatic islet MPS (PANIS) enabling MASLD progression and islet dysfunction to be assessed. The modular design of this system (vLAMPS-PANIS) allows intra-organ and inter-organ dysregulation to be deconvoluted. When compared to normal fasting (NF) conditions, under early metabolic syndrome (EMS) conditions, the standalone vLAMPS exhibited characteristics of early stage MASLD, while no significant differences were observed in the standalone PANIS. In contrast, with EMS, the coupled vLAMPS-PANIS exhibited a perturbed islet-specific secretome and a significantly dysregulated glucose stimulated insulin secretion response implicating direct signaling from the dysregulated liver acinus to the islets. Correlations between several pairs of a vLAMPS-derived and a PANIS-derived factors were significantly altered under EMS, as compared to NF conditions, mechanistically connecting MASLD and T2DM associated hepatic-factors with islet-derived GLP-1 synthesis and regulation. Since vLAMPS-PANIS is compatible with patient-specific iPSCs, this platform represents an important step towards addressing patient heterogeneity, identifying disease mechanisms, and advancing precision medicine.

Fig. 1. A human liver acinus structured biomimetic MPS coupled to a pancreatic islet MPS recapitulates clinically relevant aspects of the dysregulated liver-pancreas axis associated with the progression of MASLD and T2DM.

A Hypothesis: early metabolic syndrome (EMS) medium-induced MASLD secretome directly promotes early islet dysfunction. B Dynamic metabolic dyshomeostasis of the liver and pancreatic-islets associated with disease progression. C Schematic (see Figure S1c for actual photograph) depicting the modular two-organ coupled vLAMPS-PANIS chip. The hepatic and islet modules can also be used in an uncoupled standalone configuration (Figure S1A, B) to enable both intra- and inter-organ functional characterization to be deconvolved under normal (i.e., NF) and disease conditions. The hepatic influx medium composition determines a normal fasting (NF) or EMS state and its continuous flow among the self-assembled multiple cell types (see Methods and Results) can be monitored for biomarkers that complement in situ measurements.

Fig. 2. EMS medium promotes an early MASLD phenotype in the standalone vLAMPS maintained for 8 days.

A Representative immunofluorescence images of lipid accumulation (LipidTOX) in hepatocytes under NF and EMS conditions and B LipidTOX quantification in conjunction with secreted proinflammatory cytokines and chemokines. C Representative immunofluorescence images of Insulin receptor susbtrate-2 (IRS-2) in hepatocytes under NF and EMS conditions and D IRS-2 quantification in conjunction with an increase in hepatokine secretion. E Representative immunofluorescence images of stellate cell activation (α-SMA) under NF and EMS conditions and F α-SMA quantification in conjunction with secreted fibrotic and proinflammatory factors. Scale bar, 100 µm. Scatter plot graphs display the mean and standard error (SEM) of the EMS values normalized to NF for each of the metrics (n = 10 MPS). (see Fig. S4). Points on the graph represent individual chips. Statistical analysis by Wilcoxon signed-rank test of normalized EMS values to their corresponding NF study.

Fig. 3. EMS medium alone is not sufficient to induce islet dysfunction. NF and EMS media maintain islet viability and functionality including glucose stimulated insulin response for 8 days in the standalone PANIS.

A Representative live/dead assay image for islets cultured in NF and EMS media. B Percentage of live cell shows no significant difference between NF and EMS (n = 10 islets and n = 7 islets). C Immunofluorescence labeling for C-peptide expressing β-cells and glucagon expressing α-cells for islets cultured in NF and EMS. D The fraction of quantified C-peptide and glucagon positive cells shows no significant difference (n = 8 islets and n = 10 islets). E Representative images showing mitochondrial associated ROS immunofluorescence labeling of islets. F Percentage of ROS intensity quantification with respect to a positive control (50 µM TBHP) for islets cultured in NF and EMS media (see Methods) demonstrates no significant difference between NF and EMS (n = 8 islets and n = 8 islets). G Standalone PANIS efflux analysis showed the change in islet-derived factors in EMS normalized to NF (see Fig. S6), with no significant difference between them (n = 5 MPS). H Glucose stimulated insulin secretion (GSIS) assay, where islet samples were subjected to low (3 mM), and subsequent high (16 mM) glucose concentration and the resulting secreted insulin was measured and showed no significant difference at either glucose concentration in islets that were maintained in either NF or EMS media (n = 4 islets and n = 3 islets). I Comparison of insulin stimulation index secretion for islets cultured in NF and EMS showing no statistically significant difference. Scatter plot graphs display the mean and standard error (SEM) with p-values shown. For graphs -B, D, F, H, and I- statistical analysis was done by two-tail t test with Welch’s correction. For graph -G- Statistical analysis was done by Wilcoxon signed-rank test of normalized EMS values to their corresponding NF study. All scale bar, 50 µm.

Fig. 4. Coupling vLAMPS to PANIS maintains islet viability under EMS conditions for at least 8 days but induces significant perturbations in the islet secretome and function.

A Representative images demonstrating the effect of EMS medium on vLAMPS indicated by markers of steatosis (LipidTOX), insulin signaling (IRS-2), and fibrosis (α-SMA) in hepatocytes. Scale bar, 100 µm. B Corresponding quantitation of hepatic EMS induced steatosis, decreased insulin signaling, and increasing fibrosis (red). Along with vLAMPS-derived EMS secreted factors (from the coupled efflux) normalized to complementary NF chips (n = 12 MPS). (reproducible, Table 1). C Representative images of the PANIS live/dead assay, immunofluorescence labeling for C-peptide (green) expressing β-cells and glucagon (red) expressing α-cells, and ROS oxidative stress for coupled islets cultured in NF and EMS (n = 10 islets and n = 7 islets). All scale bar, 50 µm. D Percentage of live cell of the islets shows no statical difference between conditions. E No statistically significant difference between the fraction of quantified C-peptide and glucagon positive cells in the islets (n = 7 islets and n = 10 islets). F Percentage of ROS intensity quantification with respect to a positive control for coupled islets cultured in NF and EMS (see Methods) demonstrated no significant differences between islets cultured in either NF or EMS media (n = 12 islets and n = 11 islets). G islet-derived factors (from the coupled efflux) in EMS normalized to NF (reproducible, Table 1) demonstrated significant increases in the secretion of both amylin and pancreatic polypeptide in EMS medium (n = 12 MPS). H The islets from the coupled vLAMPS-PANIS in EMS exhibit significant abnormal insulin response to low glucose in islets that were maintained in either NF or EMS media (n = 10 islets and n = 6 islets), which was not observed in the standalone PANIS islets (see Fig. 3H). I The same significance is observed in the insulin stimulation index response of islets that were coupled with NF medium compared to EMS medium. (see Methods). Scatter plot graphs display the mean and standard error (SEM) with p-values shown. For graphs -D, E, F, H, and I- statistical analysis was done by two-tail t test with Welch’s correction. For graph -B and G- statistical analysis was done by Wilcoxon signed-rank test of normalized EMS values to their corresponding NF study.