Regulation of Neurod1 contributes to the lineage potential of Neurogenin3+ endocrine precursor cells in the pancreas

Abstract

During pancreatic development, transcription factor cascades gradually commit precursor populations to the different endocrine cell fate pathways. Although mutational analyses have defined the functions of many individual pancreatic transcription factors, the integrative transcription factor networks required to regulate lineage specification, as well as their sites of action, are poorly understood. In this study, we investigated where and how the transcription factors Nkx2.2 and Neurod1 genetically interact to differentially regulate endocrine cell specification. In an Nkx2.2 null background, we conditionally deleted Neurod1 in the Pdx1+ pancreatic progenitor cells, the Neurog3+ endocrine progenitor cells, or the glucagon+ alpha cells. These studies determined that, in the absence of Nkx2.2 activity, removal of Neurod1 from the Pdx1+ or Neurog3+ progenitor populations is sufficient to reestablish the specification of the PP and epsilon cell lineages. Alternatively, in the absence of Nkx2.2, removal of Neurod1 from the Pdx1+ pancreatic progenitor population, but not the Neurog3+ endocrine progenitor cells, restores alpha cell specification. Subsequent in vitro reporter assays demonstrated that Nkx2.2 represses Neurod1 in alpha cells. Based on these findings, we conclude that, although Nkx2.2 and Neurod1 are both necessary to promote beta cell differentiation, Nkx2.2 must repress Neurod1 in a Pdx1+ pancreatic progenitor population to appropriately commit a subset of Neurog3+ endocrine progenitor cells to the alpha cell lineage. These results are consistent with the proposed idea that Neurog3+ endocrine progenitor cells represent a heterogeneous population of unipotent cells, each restricted to a particular endocrine lineage.

Figure 1. Neurod1 deletion in the pancreas progenitors, in an Nkx2.2 null background, phenocopies the Nkx2.2^null;Neurod1^null double knockout.

Pancreatic tissue from wildtype, Neurod1^Δpanc, Nkx2.2^null, and Nkx2.2^null;Neurod1^Δpanc was analyzed by immunofluorescence for expression of the islet hormones glucagon (gluc), ghrelin (ghr) and insulin (ins) at P0 (A–L; white bar indicates 50 microns; DAPI marks all nuclei). Amylase (amyl) expression marks exocrine tissue in all genotypes (A–D). The quantitative expression of glucagon (Gcg) (M), ghrelin (Ghr) (N), and pancreatic polypeptide (Ppy) (O), as well as deletion of Neurod1 (P) was determined by real time PCR using RNA extracted from wildtype, Neurod1^Δpanc, Nkx2.2^null, Nkx2.2^null;Neurod1^Δpanc, and Nkx2.2^null;Neurod1^null pancreas (P0; N = 3–8). Relative mRNA expression was normalized to the housekeeping gene, cyclophilinB. Data are represented as mean+/−SEM. * p<0.05; ** p<0.01; *** p<0.001.

Figure 2. The genetic interaction of Nkx2.2 and Neurod1 is required in the Neurog3+ endocrine cells to specify PP and epsilon cells.

Pancreatic polypeptide-expressing PP cells (A) and ghrelin-expressing epsilon cells (G) were quantified by morphometric analysis, comparing wildtype, Neurod1^Δendo, Nkx2.2^null, Nkx2.2^null;Neurod1^Δendo, Nkx2.2^null;Neurod1^Δpanc, and Nkx2.2^null;Neurod1^null at P0. Cell numbers were quantified relative to total pancreas area and displayed normalized to wildtype. Representative sections stained for ghrelin and insulin illustrate the change in ghrelin-expressing cells between genotypes, and the absence of insulin-expressing cells the Nkx2.2^null and Nkx2.2^null;Neurod1^Δendo (C–F; white bar indicates 50 microns; DAPI marks all nuclei). The expression of pancreatic polypeptide (Ppy) (B) and ghrelin (Ghr) (H) was determined for all genotypes by real time PCR (P0; N = 3–7). Relative mRNA expression was normalized to the housekeeping gene, cyclophilinB. Data are represented as mean+/−SEM. * p<0.05; ** p<0.01; *** p<0.001.

Figure 3. Alpha cells are only minimally restored in the Nkx2.2^null;Neurod1^Δendo.

Pancreatic tissue was analyzed by immunofluorescence for the presence of glucagon-expressing cells at P0, comparing wildtype (A), Neurod1^Δendo (B), Nkx2.2^null (C), and Nkx2.2^null;Neurod1^Δendo (D). Amylase expression marks exocrine tissue in all genotypes (A–D; white bar indicates 50 microns; DAPI marks all nuclei). Glucagon-expressing alpha cells were quantified by morphometric analysis, relative to total pancreas area and displayed normalized to wildtype (E). The expression of glucagon (Gcg) (F) and Neurod1 (G) was measured by real time PCR using RNA from P0 pancreas for all genotypes (N = 3–7). Relative mRNA expression was normalized to the housekeeping gene, cyclophilinB. Data are represented as mean+/−SEM. * p<0.05; ** p<0.01; *** p<0.001.

Figure 4. Alpha cells are present in the early pancreatic domain of the Nkx2.2^null;Neurod1^null double-knockout mouse.

Sections from E10.5 wildtype (A) and Nkx2.2^null;Neurod1^null (B) embryos were stained for Pdx1 to identify the pancreatic domain. Adjacent sections were stained for FoxA and glucagon (gluc), to identify alpha cells in the early pancreatic domain in both the wildtype (C) and Nkx2.2^null;Neurod1^null (D). White bar indicates 50 microns. DAPI marks all nuclei.

Figure 5. Alpha cells are not rescued with deletion of Neurod1 in glucagon+ cells, in the absence of Nkx2.2.

Pancreatic tissue from wildtype, Neurod1^Δalpha, Nkx2.2^null, and Nkx2.2^null;Neurod1^Δalpha was analyzed by immunofluorescence for expression of the islet hormones glucagon (gluc), ghrelin (ghr) and insulin (ins) at P0 (A–L; white bar indicates 50 microns; DAPI marks all nuclei). Amylase (amyl) expression marks exocrine tissue in all genotypes (A–D). The quantitative expression of glucagon (Gcg) (M), ghrelin (Ghr) (N), and pancreatic polypeptide (Ppy) (O) was determined by real time PCR using RNA extracted from pancreas (P0; N = 3–7). Relative mRNA expression was normalized to the housekeeping gene, cyclophilinB. Data are represented as mean+/−SEM. * p<0.05; ** p<0.01; *** p<0.001.

Figure 6. Neurod1 is expressed in a subset of endocrine progenitor cells.

Utilizing the Neurod1:LacZ knock-in allele (Neurod1^LacZ/+) and immunofluorescence on tissues sections from E9.5 and E13.5 embryos, the expression pattern of Neurod1 (marked by beta-galactosidase; beta-gal) and glucagon (A, B), and Neurod1 and Neurog3 (C, D) was determined. DAPI marks all nuclei. All images are confocal. White bar indicates 50 microns. Boxes denote area magnified for inset, which are +1.75zoom of lower power image. Top and right rectangular panels represent a Z projection of at least 10 stack pictures at the level of intersection of the red/green crosshairs. (E) The percentage of each of the populations of glucagon+ cells, or (F) Neurog3+ cells was quantitated at E9.5 and E13.5. Data are represented as mean+/−SEM.

Figure 7. Nkx2.2 represses the Neurod1 promoter in alphaTC1 cells.

(A) Schematic representation of the Neurod1 minimal promoter, with the areas previously identified to be activated by Nkx2.2 denoted with grey boxes. (B) Luciferase activity was assessed in alphaTC1 cells transfected with Neurod1 promoter constructs (NDfull, NDΔ2, NDΔ3, NDΔ4) in addition to pcDNA3 alone or Nkx2.2. Nkx2.2-dependent activity was determined based on promoter region deletion. Luciferase activity was determined 48 hours post-transfection. Luciferase readings were normalized to Renilla luciferase values. (C) H3K4me3 is enriched in alpha and beta cells, although at significantly lower levels in alpha cells. The Nkx2.2 dephosphorylated mutant (S-11-A) results in a significant increase in H3K4me3 enrichment in alpha cells, comparable to levels observed in beta cell. Conversely, the Nkx2.2 phosphorylation mutant (S-11-D) results in a significant decrease in H3K4me3 in beta cells, comparable to levels in alpha cells. (D) The repressive H3K27me3 mark is not present on the Neurod1 promoter in alpha or beta cells (n = 3). Data was normalized to Gapdh. All data are represented as mean+/−SEM. * p<0.05.

Figure 8. A proposed model for the involvement of Nkx2.2 and Neurod1 in alpha and beta cell specification.

Taking into account both our in vivo and in vitro data, we propose that specific combinations of transcription factors acting in the progenitor cells within the early pancreatic epithelium set up the competency of the unipotent endocrine progenitors to become specific islet cell types. Specifically, we propose a model whereby Nkx2.2 must repress Neurod1 in a Pdx1+ progenitor, and this repression maintained in the Neurog3+ endocrine progenitor, thereby permitting glucagon-expressing alpha cell specification. Conversely, activation of Neurod1 by Nkx2.2 permits beta cell formation.