Loss of the transcription factor MAFB limits β-cell derivation from human PSCs

Abstract

Next generation sequencing studies have highlighted discrepancies in β-cells which exist between mice and men. Numerous reports have identified MAF BZIP Transcription Factor B (MAFB) to be present in human β-cells postnatally, while its expression is restricted to embryonic and neo-natal β-cells in mice. Using CRISPR/Cas9-mediated gene editing, coupled with endocrine cell differentiation strategies, we dissect the contribution of MAFB to β-cell development and function specifically in humans. Here we report that MAFB knockout hPSCs have normal pancreatic differentiation capacity up to the progenitor stage, but favor somatostatin- and pancreatic polypeptide-positive cells at the expense of insulin- and glucagon-producing cells during endocrine cell development. Our results describe a requirement for MAFB late in the human pancreatic developmental program and identify it as a distinguishing transcription factor within islet cell subtype specification. We propose that hPSCs represent a powerful tool to model human pancreatic endocrine development and associated disease pathophysiology.

Fig. 1. MAFB is expressed in human β-cell counterparts and generation of MAFB KO hESCs.

a Schematic outlining the differentiation protocol to generate β-like cells from hESCs. The key lineage markers are outlined at each stage. Chemicals and durations for each differentiation stage are indicated in the Methods section. d, day(s); hESC, undifferentiated hESC stage; DE, definitive endoderm stage SOX17 + FOXA2+; PP, pancreatic progenitor stage PDX1 + NKX6.1+; β, β-like cell stage. b Representative FC plots outlining a typical differentiation of cells through the DE, PP, and β-like cell stages using the indicated markers. c Representative IF images from four independent experiments at indicated stages of differentiation depicting SOX17 and FOXA2, PDX1 and NKX6.1, C-PEP co-staining with PDX1 and NKX6.1, and MAFB co-staining with C-PEP and NKX6.1. DAPI indicates nuclear staining. Scale bars, 50 μm. d MAFB mRNA expression during β-cell differentiation as shown via qPCR. Each data point represents an independent biological sample. e Representative western blot from three independent experiments showing MAFB protein levels during β-cell differentiation at the indicated stages. f Targeted disruption of MAFB in hESCs. Targeting strategy for CRISPR-mediated gene editing of MAFB. g Representative western blot analysis of MAFB in targeted clones at the β-cell of differentiation from three independent experiments. Sizes of molecular weight markers are shown on the right. Similar results were obtained using two independent antibodies. h Representative IF images from three independent experiments at the β-cell stage of differentiation depicting MAFB expression in MAFB +/+, +/−, and −/− cells. DAPI indicates nuclear staining. Scale bars, 50 μm.

Fig. 2. Loss of MAFB limits β-cell derivation.

a Representative FC plots depicting the percentages of SOX17 + FOXA2+ cells and quantification at the DE stage (n = 6, biologically independent samples). b Representative FC plots depicting the percentages of PDX1 + NKX6.1+ cells and quantification at the PP stage (n = 6, biologically independent samples). c Representative FC plots depicting the percentages of GFP+ cells and quantification at the β-cell stage (n = 6, biologically independent samples). d Representative FC plots depicting the percentages of C-PEP + NKX6.1+ cells and quantification at the β-cell stage (n = 6, biologically independent samples). e Representative IF images from six independent experiments at the indicated stages of differentiation depicting SOX17 and FOXA2, PDX1 and NKX6.1, and C-PEP co-staining with PDX1 and NKX6.1. Scale bars, 50 μm. GFP and Brightfield images from live in vitro cell cultures. Scale bars, 100 μm. P values by one-way ANOVA followed by Dunnett’s multiple comparisons test were *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as individual biological replicates and represent the mean ± SD.

Fig. 3. MAFB is required to generate functional β-cells.

a Schematic created with Biorender.com outlining the transplantation strategy to test the functionality of MAFB+/+ and −/− β-like cells under the kidney capsule of NSG mice. b Serum glucose measurements from glucose-stimulated insulin secretion (GSIS) experiments at 5 weeks post-transplantation (n = 3 independent mice). c GSIS experiment at 5 weeks post-transplantation. Experiments were performed one time with three independent NSG mice (n = 3) per group. P values by paired two-tailed t-test. d Representative IF images from six independent animals (3 MAFB+/+ and 3 MAFB−/−) for C-PEP co-staining with PDX1 and NKX6.1 of grafts removed from kidney capsules 8 weeks post-transplantation. Scale bars, 50 μm. Data are presented as individual biological replicates and represent the mean ± SEM.

Fig. 4. MAFB controls pancreatic endocrine cell lineage specification.

a UMAP projections of integrated MAFB+/+ and −/− transcriptomes, color-coded by genotype (left) and cell populations (right). b Scatter plot showing the average gene expression (log scale) for MAFB+/+ and −/− cells at the β-like cell stage. Differentially expressed genes (log fold change >0.5, adjusted P value <0.1) between MAFB+/+ and −/− cells are shown in red. Significance was calculated using the MAST test and P values were adjusted for multiple testing using the Benjamini–Hochberg method. The top FIVE up- and downregulated genes are indicated. c qPCR analysis showing mRNA levels of islet hormones (INS, GCG, SST, PPY, and GHRL) at the β-cell stage by. P values by unpaired two-tailed t-test were *P < 0.05 and ***P < 0.001. d–f Representative FC plots depicting d C-PEP+ and CHGA+ cells (n = 5, biologically independent samples), e C-PEP+ and GCG+ cells (n = 5, biologically independent samples), and f C-PEP+ and SST+ cells (n = 4, biologically independent samples) and respective quantification at the β-like cell stage. g Representative IF images from four independent experiments at the β-cell stage of differentiation depicting GCG, SST, NKX6.1, and DAPI. Scale bars, 50 μm. h Western blotting for SST protein (14 kDa) expression at the β-cell stage. GAPDH (38 kDa) was used as a loading control (n = 2 biologically independent MAFB+/+ samples and n = 3, biologically independent samples for MAFB + /− and −/−). i Representative FC plots depicting the percentages of C-PEP + PPY+ cells and quantification at the β-like cell stage (n = 4 for MAFB+/+ and n = 3 for MAFB−/−, biologically independent samples). P values by unpaired two-tailed t-test were ****P < 0.0001. j Representative IF images from four independent experiments depicting SST and HHEX (upper panel) and PPY and GCG (lower panel) and DAPI. Scale bars, 50 μm. P values by one-way ANOVA followed by Dunnett’s multiple comparisons test were *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as individual biological replicates, representing the mean ± SD. N numbers indicated in Supplementary Fig. 2a, b.

Fig. 5. MAFB is critical to β-cell identity.

a Representative Western blotting from three independent experiments for MAFB protein (37 kDa) expression in MAFB−/− and iHygroMAFB rescue cells with or without DOX as indicated. Vinculin (114 kDa) was used as a loading control. b Representative FC plots depicting the percentages of GFP+ cells and quantification at the β-cell stage (n = 3). P values by paired two-tailed t-test. c Representative GFP and Brightfield images from three independent experiments with or without DOX from live in vitro cell cultures at the β-cell stage. Scale bars, 100 μm. d The mRNA expression patterns of designated genes for islet hormones and selected transcription factors (INS, GCG, SST, PPY, PDX1, and MAFA) as measured by quantitative PCR analysis in the presence or absence of DOX (n = 3). P values by paired two-tailed t-test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. e Representative FC plots depicting the percentages of indicated markers at the indicated cell stages (n = 3 or 4, biologically independent samples for MAFB+/+, +/−, and −/− iPSC differentiation). f Quantification of the indicated cell populations from FC plots in (e), normalized to the unedited control, set to 1. g Representative IF images from three independent experiments at the DE, PP, and β-like cell stages of differentiation for the respective markers as indicated. Scale bars, 50 μm. P values by one-way ANOVA followed by Dunnett’s multiple comparisons test were *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 6. Loss of MAFB limits α-cell differentiation.

a Schematic outlining the differentiation protocol to generate α-like cells from hESCs. The key lineage markers are outlined at each stage. Chemicals and durations for each differentiation stage are indicated in the Methods section. d, day(s); hESC, undifferentiated hESC stage; DE, definitive endoderm stage (SOX17 + FOXA2+); EP, endocrine progenitor (CHGA + PDX1+); immature α-like cells (GCG + INS + ARX+) and enriched α-like cells (GCG + ARX+). b Representative FC plots at the indicated time point depicting the percentages of GCG + C-PEP+ cells and quantification at the DE stage. c, d Quantification of FC data outlined in (b) (n = 3). P values by one-way ANOVA followed by Dunnett’s multiple comparisons test were *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as individual biological replicates and represent the mean ± SD. e Representative FC plots at the α-like cell stage depicting the percentages of ARX+, PAX6+ NKX2.2+, and PDX1+ NKX6.1+ cells. f Quantification of FC data in (e) (n = 3, biologically independent samples). g The mRNA expression patterns of designated genes for islet hormones and selected transcription factors (CHGA, INS, GCG, SST, PPY, ARX, and PDX1) as measured by quantitative PCR analysis (n = 3, biologically independent samples). P values by unpaired two-tailed t-test were *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. h Representative IF images from three independent experiments at the α-cell stage depicting GCG, INS and ARX and PDX1, SST and NKX6.1 in d26 α-cells from MAFB+/+ and −/− differentiations. Scale bars, 50 μm. i Quantification of total GCG content from MAFB+/+ and −/− cells at d26–34 in vitro (n = 3, biologically independent samples). P values by paired two-tailed t-test.