Generation of functional insulin-producing cells from neonatal porcine liver-derived cells by PDX1/VP16, BETA2/NeuroD and MafA

Abstract

Surrogate β-cells derived from stem cells are needed to cure type 1 diabetes, and neonatal liver cells may be an attractive alternative to stem cells for the generation of β-cells. In this study, we attempted to generate insulin-producing cells from neonatal porcine liver-derived cells using adenoviruses carrying three genes: pancreatic and duodenal homeobox factor1 (PDX1)/VP16, BETA2/NeuroD and v-maf musculo aponeurotic fibrosarcoma oncogene homolog A (MafA), which are all known to play critical roles in pancreatic development. Isolated neonatal porcine liver-derived cells were sequentially transduced with triple adenoviruses and grown in induction medium containing a high concentration of glucose, epidermal growth factors, nicotinamide and a low concentration of serum following the induction of aggregation for further maturation. We noted that the cells displayed a number of molecular characteristics of pancreatic β-cells, including expressing several transcription factors necessary for β-cell development and function. In addition, these cells synthesized and physiologically secreted insulin. Transplanting these differentiated cells into streptozotocin-induced immunodeficient diabetic mice led to the reversal of hyperglycemia, and more than 18% of the cells in the grafts expressed insulin at 6 weeks after transplantation. These data suggested that neonatal porcine liver-derived cells can be differentiated into functional insulin-producing cells under the culture conditions presented in this report and indicated that neonatal porcine liver-derived cells (NPLCs) might be useful as a potential source of cells for β-cell replacement therapy in efforts to cure type I diabetes.

Figure 1. Histological and cellular morphologies in the neonatal and adult porcine liver.

(A) Hematoxylin and eosin staining revealed differences in the lobular structures of these 2 tissues (upper panel). Albumin, a marker of mature hepatic cells, was weakly expressed in the neonatal liver compared to the adult liver (middle panel). Cytokeratin 19 (CK19), an epithelial or ductal cell marker, was detected in the liver sections. (B) Immunohistochemical staining to detect CD34 (upper panel), alpha-fetoprotein (AFP; a stem cell marker; middle panel), and sox9 (a marker of early cells in the bile duct; bottom panel) in liver sections. Note that CD34, AFP, and sox9 were only expressed in neonatal tissue and not in adult tissue. (C) Proliferating cells were only detected in neonatal livers via Ki67 staining. (D) CCK-8 activity was measured in primary isolated and cultured liver cells from neonatal and adult livers. The measurements were conducted in three independent experiments. ***P<0.005. Error bar represents the SE.

Figure 2. Viral transduction efficiency determined via FACS and relative gene expression determined via quantitative real-time PCR.

(A) Various viral titers were tested using green fluorescent protein (GFP) as a reporter gene. Sampling was performed at 72 h after the transduction of Ad-GFP. The obtained transduction efficiency was 78% using a multiplicity of infection (MOI) of 50 after 16 h (upper) and 87% at an MOI of 100 (bottom). (B) Monolayer cells spontaneously formed clusters on low-attachment culture dishes at 24 h after transduction. Most of the clusters expressed GFP well at 72 h after transduction (C–E) qRT-PCR analysis in transduced cells. Sampling was performed during final step of differentiation. Ct values were normalized to GAPDH gene expression within the same cDNA sample. The results are presented as the fold increase based on GAPDH gene expression. All primers were designed based on the pig-aligned sequence from NCBI. The presenteddata are the means of six independent experiments. The meansand standard error values are presented. *P<0.05; **P<0.01; ***P<0.005;Error bars, S.E.

Figure 3. Immunofluorescence staining of NPLCs following the transduction of Ad-GFP.

NPLCs were transduced with Ad-GFP 1 day after isolation for 24 h. (A) Sox+ expression in DAPI+ nuclei. Numerous Sox9+ cells were observed in aggregates following (B) FACS sorting of NPLCs. Sampling for FACS was performed at 48 h after transduction. A total of 78% of the GFP+ cells were sorted. (C) Immunofluorescence staining to detect albumin (ALB), alpha fetoprotein (AFP) and CD34.

Figure 4. In vitro functional analyses and quantification of cellular differentiation.

(A) Measurement of glucose-stimulated insulin secretion from NPLCs transduced with Ad-GFP or Ad-PDX1/VP16, BETA2/NeuroD and MafA (n = 3). Islets from adult SNU pigs were used as a positive control. (B) Cellular insulin content. The insulin content was normalized to the amount of total cellular protein (mg). (C) Immunofluorescence visualization of treated NPLCs. Sampling was performed at 6 days after reaggregation, which is the final step of the induction of differentiation. Nuclei are stained with DAPI (a–c) in high-magnification images. (D) Quantification data indicate that the treated NPLCs were trans-differentiated into insulin-producing cells at an efficiency of 11%. Scale bars, 100 µm. *P<0.05;**P<0.01; ***P<0.005; Error bars, S.E.

Figure 5. Effect of the subcapsular transplantation of treated NPLCs on hyperglycemia in STZ-induced nude mice.

(A) Non-fasting blood glucose levels were monitored for 6 weeks after transplantation. (B) The average blood glucose levels over the recording period, from 7–42 days after transplantation, are reported (Ad-GFP, animals receiving cells transduced with Ad-GFP, n = 6; TPL, animals receiving cells transduced with triple adenoviruses, n = 9). (C) At 6 weeks after transplantation, the intra-peritoneal glucose tolerance test (IP-GTT) was performed in 3 groups: normal control mice (n = 3), STZ diabetic immune-deficient mice (diabetic control, n = 3), and mice receiving cells containing Ad-PDX1/VP16, BETA2/NeuroD, and MafA (transplanted group; TPL, n = 3). (D) The values for the area under the curve for glucose (AUCg) indicate an improvement of diabetes upon transplantation of treated NPLCs (NC, normal control; DC, diabetic control; TPL, transplanted group). The Bonferroni method was used for statistical analysis. (E) At 6 weeks after transplantation, the grafts were harvested. Note the insulin-positive cells in the graft. Nuclei were stained with DAPI. (a–c) High-magnification images.(F) Quantification of insulin-positive cells in the graft. Scale bars, 50 µm. *P<0.05; ***P<0.005; Error bars, S.E.