Analysis of beta-cell gene expression reveals inflammatory signaling and evidence of dedifferentiation following human islet isolation and culture

Abstract

The stresses encountered during islet isolation and culture may have deleterious effects on beta-cell physiology. However, the biological response of human islet cells to isolation remains poorly characterized. A better understanding of the network of signaling pathways induced by islet isolation and culturing may lead to strategies aimed at improving islet graft survival and function. Laser capture microdissection (LCM) was used to extract beta-cell RNA from 1) intact pancreatic islets, 2) freshly isolated islets, 3) islets cultured for 3 days, and changes in gene expression were examined by microarray analysis. We identified a strong inflammatory response induced by islet isolation that continues during in-vitro culture manifested by upregulation of several cytokines and cytokine-receptors. The most highly upregulated gene, interleukin-8 (IL-8), was induced by 3.6-fold following islet isolation and 56-fold after 3 days in culture. Immunofluorescence studies showed that the majority of IL-8 was produced by beta-cells themselves. We also observed that several pancreas-specific transcription factors were down-regulated in cultured islets. Concordantly, several pancreatic progenitor cell-specific transcription factors like SOX4, SOX9, and ID2 were upregulated in cultured islets, suggesting progressive transformation of mature beta-cell phenotype toward an immature endocrine cell phenotype. Our findings suggest islet isolation and culture induces an inflammatory response and loss of the mature endocrine cell phenotype. A better understanding of the signals required to maintain a mature beta-cell phenotype may help improve the efficacy of islet transplantation.

Figure 1. Schematic diagram of the workflow.

Frozen sections of the intact human pancreas, freshly isolated islets (d0) and cultured islets (d3) were processed for LCM. The beta-cells were identified by their intrinsic autofluorescence and captured by LCM, followed by RNA extraction, amplification and labeling. Labeled RNA was hybridized to Human WG-6 Expression Arrays (Illumina) and scanned by BeadArray Reader. Expression data were analyzed by Flexarray, DAVID and IPA and validated by immunostaining.

Figure 2. Principal Component analysis (PCA) plot of variability in whole genome expression from intact or isolated islets.

Samples were distributed by their similarity in expression data using dimensionality reduction. Pancreatic intact islets (green dots) cluster at right whereas samples from d3-islets (red dots) cluster at left, apart from other groups. The freshly isolated islet samples (yellow dots) cluster at middle and partially intermix with other groups. Principal component 1 and Principal component 2 together accounted for 27% variation.

Figure 3. Hierarchical clustering of differentially expressed genes in fresh (d0) and cultured islets (d3).

The expression data for intact islets and isolated islets (d0-islets & d3-islets) were analyzed and top 359 genes with p<0.05 and fold change >2.5 were used to create cluster map.

Figure 4. IL-8 expression in isolated islets.

Immunofluorescence staining was performed for IL8 in frozen sections from pancreas (a–h), d0-islets (i–p) and d3-islets (q–x) and costaining was performed with insulin (c, k, s) or glucagon (g, o, w). IL-8 expression increases after islet isolation and culture. Most of IL-8 positive cells also costain with insulin but not glucagon. Scale bar (10 micron) on merged image represents scale for all images on its left side.

Figure 5. Validation of microarray data by quantitative real-time PCR.

Quantitative real-time PCR was performed for IL-8, ID2, SOX4 and SOX9. Fold-changes represent expression level of genes in isolated islets relative to intact islets. Mean fold changes derived from quantitative PCR (white bars) were plotted along with the corresponding mean fold changes from microarray analysis (solid bars). Error bars represent standard deviation. * represents t-test p-values<0.05. For microarray standard deviation and t-test was performed on the expression values obtained after normalization with Lumi. The mean fold changes determined by quantitative PCR were in concordance with the mean fold changes from microarray data.