Pax6 is crucial for β-cell function, insulin biosynthesis, and glucose-induced insulin secretion

Abstract

The Pax6 transcription factor is crucial for endocrine cell differentiation and function. Indeed, mutations of Pax6 are associated with a diabetic phenotype and a drastic decrease of insulin-positive cell number. Our aim was to better define the β-cell Pax6 transcriptional network and thus provide further information concerning the role of Pax6 in β-cell function. We developed a Pax6-deficient model in rat primary β-cells with specific small interfering RNA leading to a 75% knockdown of Pax6 expression. Through candidate gene approach, we confirmed that Pax6 controls the mRNA levels of the insulin 1 and 2, Pdx1, MafA, GLUT2, and PC1/3 genes in β-cells. Importantly, we identified new Pax6 target genes coding for GK, Nkx6.1, cMaf, PC2, GLP-1R and GIPR which are all involved in β-cell function. Furthermore, we demonstrated that Pax6 directly binds and activates specific elements on the promoter region of these genes. We also demonstrated that Pax6 knockdown led to decreases in insulin cell content, in insulin processing, and a specific defect of glucose-induced insulin secretion as well as a significant reduction of GLP-1 action in primary β-cells. Our results strongly suggest that Pax6 is crucial for β-cells through transcriptional control of key genes coding for proteins that are involved in insulin biosynthesis and secretion as well as glucose and incretin actions on β-cells. We provide further evidence that Pax6 represents a key element of mature β-cell function.

Fig. 1.

Specific inhibition of Pax6 gene expression by siRNA identifies target genes in primary rat β-cells. Primary β-cells (∼40,000 cells/condition) were transfected with 100 nm Pax6 siRNA cocktail (siPax6-614 and siPax6-1007) or corresponding scramble (+) during 96 h. A, Effects of Pax6 siRNA on Pax6 gene expression by real-time RT-PCR. Data are corrected by β-actin mRNA values. B, Western blot analyses of Pax6 protein content from transfected primary β-cells with scramble (Scr) or specific Pax6 siRNA (siPax6) after 96 h. C, Quantitative analyses of the expression of key pancreatic endocrine genes coding for insulin and proteins involved in β-cell differentiation and function or insulin gene transcription 96 h after scramble or Pax6 siRNA transfection. Data are expressed relative to β-actin mRNA values; corrected for transfection efficiency as the means ± sem for at least five different experiments. *, Statistical significance with P < 0.05 value using a Student's t test; NS, no significant effect.

Fig. 2.

Specific inhibition of Pax6 gene expression decreases protein content of corresponding putative Pax6 target genes in primary β-cells. A, Western blot analyses of putative Pax6 targets from total cell extracts of transfected primary β-cells with scramble (Src) or specific Pax6 siRNA (siPax6) after 96 h. B, Histograms represent protein amounts after 96 h transfection with specific Pax6 or scramble siRNA. Each Western blotting was performed on three different transfection experiments. Data are presented as the means ± sem. *, Statistical significance with P < 0.05 using Student's t test; NS, no significant effect.

Fig. 3.

Organization of Pax6 target gene promoters. Schematic representation of the GLUT2, PC1/3, GK, Nkx6.1, GLP-1R, and GIPR gene promoters. Putative Pax6-binding sites are identified using in silico analyses of binding site search with the Genomatix software (Matinspector, available at http://www.genomatix.de) and represented by a black circle. Sequences of the Pax6 putative binding site is mentioned in capitals.

Fig. 4.

Analyses of Pax6 binding on the PC1/3, GK, Nkx6.1, GLP-1R, and GIPR gene promoters by EMSA experiments. Representative gels for Pax6 binding on PC1/3, GK, Nkx6.1, GLP-1R, and GIPR gene promoters. EMSA were performed with 5′ end-labeled PC1/3 (+243/−261 bp), GK (−34/−16 bp), Nkx6.1 (+377/+395 bp), GLP-1R (−233/−216 bp), and GIPR (−368/−348 bp) oligonucleotides (listed in Supplemental Tables 2 and 3) in the presence of nuclear extracts from BHK-21 cells (pSG5, negative control) or BHK-21 cells overexpressing mouse Pax6 (p46 isoform). EMSA experiments were performed at least twice for each probe.

Fig. 5.

In vivo interactions of Pax6 on the Insulin, PC1/3, PC2, GK, Nkx6.1, GLP-1R, and GIPR gene promoters. Quantitative ChIP analyses in rat islet cells (A) and β-TC3 insulin-producing cell line (B). Histograms represent relative binding of Pax6 on the insulin (pInsulin), PC1/3 (pPC1/3), PC2 (pPC2), GK (pGK), Nkx6.1 (pNkx6.1), GIPR (pGIPR), GLP-1R (pGLP-1R) gene promoters, and the negative control CNAP1 gene. Binding was analyzed by real-time PCR, and amount of amplified DNA for different target gene promoters were corrected by total precipitated DNA in each condition. Binding intensity data are expressed relative to IgG immunoprecipitation (nonspecific binding) and are presented as the mean ± sem for at least three independent experiments for β-TC3 and two independent experiments for rat islets cells. An antihistone H4 immunoprecipitation was also performed as a positive control for each promoter (data not shown). *, Statistical significance with P < 0.05 value using a Student's t test, NS, no significant effect.

Fig. 6.

Mutations of Pax6-binding sites affect the promoter activities of Pax6 target genes. The native and Pax6 mutant constructs of PC1 (−1500/+354 bp), GK (−1118/+14 bp), Nkx6.1 (−90/+1000 bp), GLP-1R (−490/+10 bp), and GIPR (−155/+166 bp) gene promoters were transfected in BHK-21 and HIT-T15 insulin-producing cell lines for 48 h. Histograms represent normalized luciferase activities of native and mutated promoter constructs vs. human placental alkaline phosphatase activities to monitor transfection efficiency. A, The native PC1/3, GK, Nkx6.1, GLP-1R, and GIPR (black bars) and corresponding mutated-Pax6 +243/+261 PC1/3, −34/−16 GK, +377/+375 Nkx6.1, −233/−216 GLP-1R, and −368/−348 GIPR promoter constructs (gray bars) were cotransfected in BHK-21 cells with pSG5 or mouse Pax6 cDNA (mPax6). Overexpression of mouse Pax6 was systematically verified in each cotransfection experiment by Western blot analyses. Histograms represent the Pax6 transactivation effects on native and mutated promoter constructs with pSG5 condition used as references (black bars). B, The native and mutated promoter constructs were transfected in HIT-T15 cells. The relative activities of each promoter constructs were expressed compared with pGL3basic activities. A minimum of three independent transfections were performed, each of them carried out in duplicate. Data are presented as the means ± sem. *, Statistical significance with P < 0.05 using Student's t test; NS, no significant effect. WT, Wild-type (native promoter); Mut, mutated construct (mutated promoter).

Fig. 7.

Pax6 gene silencing in rat primary β-cells affects insulin content and maturation. A, Insulin content after specific Pax6 gene silencing. Histograms represent insulin content in primary β-cells 96 h after scramble or Pax6 siRNA transfection (40,000 cells/condition). Data are presented as the means ± sem for at least three different experiments. *, Statistical significance with P < 0.05 value using a Student's t test; NS, no significant effect. B, KCl-stimulated insulin secretion. Insulin values from primary β-cells supernatants after 1 h of incubation in KRB-2.8 mm G and an additional hour in KRB-2.8 mm G + KCl (50 nm) in the presence of scramble or Pax6 siRNA. C, Effects of Pax6 silencing on proinsulin processing in rat primary β-cells. Data are represented by relative proinsulin/insulin ratio for scramble or Pax6 siRNA conditions after 1 h incubation of in 2.8 or 16.7 mm glucose medium. D, Data are also represented by the fold induction between 16.7 and 2.8 mm, reflecting the difference between the proinsulin/insulin ratio in scramble and Pax6 siRNA conditions. *, Statistical significance compared with scramble conditions with P < 0.05 value using a Student's t test. Data are presented as the means ± sem for at least three independent experiments.

Fig. 8.

Pax6 affects glucose- and incretin-stimulated insulin secretion in rat primary β-cells. Glucose, GLP-1, and GIP effects on insulin secretion after Pax6 gene silencing in rat primary β-cells. Insulin values were determined from primary β-cells supernatants after 1 h of incubation in KRB-2.8 mm G and an additional 1 h in KRB-16.7 mm G or KRB-16.7 + GLP-1 (10 nm) or GIP (10 nm) in the presence of scramble or Pax6 siRNA. *, Statistical significance compared with 2.8 mm G condition with P < 0.05 value using a Student's t test; NS, no significant effect; δ, statistical significance compared with 16.7 mm G condition with P < 0.05 value using a Student's t test; $, statistical significance compared with scramble condition with P < 0.05 value using a Student's t test.

Fig. 9.

Pax6 affects directly and indirectly GSIS in rat primary β-cells through Pdx1 and MafA silencing. Rat primary β-cells were transfected with scramble or Pax6 siRNA during 16, 24, and 48 h. A, Effects of Pax6 siRNA on Pax6, Pdx1, and MafA gene expression by real-time RT-PCR. Data are corrected by β-actin mRNA values. Histograms represent the relative amount of Pax6, Pdx1, and MafA mRNA levels after 16, 24, and 48 h of Pax6 siRNA transfections compared with scramble conditions. B, Western blot analyses of Pax6, Pdx1, and MafA protein cell contents from transfected primary β-cells with scramble (C.) or specific Pax6 siRNA (Si.) after 16, 24, and 48 h. Quantitative analysis of Western blottings is illustrated in the inset. Histograms represent the effects of Pax6 knockdown on Pax6, MafA, and Pdx1 protein amounts after 16, 24, and 48 h of gene silencing. C, Insulin secretion measurements after 16, 24, and 48 h of Pax6 or scramble siRNA. Cells were incubated for 1 h in KRB-2.8 mm G (basal) and for an additional 1 h in KRB-16.7 mm G (glucose stimulation). The GSIS values are represented by fold inductions compared with basal values. *, Statistical significance compared with scramble condition with P < 0.05 using Student's t test; NS, no significant effects; $, statistical significance compared with 16-h conditions with P < 0.05 value using a Student's t test. Data are presented as the means ± sem for at least three different experiments.