Modeling Genetic Risk of β-Cell Dysfunction in Human Induced Pluripotent Stem Cells From Patients Carrying the MTNR1B Risk Variant

Abstract

Disruptions in circadian rhythm, partly controlled by the hormone melatonin, increase the risk of type 2 diabetes (T2D). Accordingly, a variant of the gene encoding the melatonin receptor 1B (MTNR1B) is robustly associated with increased risk of T2D. This single-nucleotide polymorphism (SNP; rs10830963; G-allele) is an expression quantitative trait locus (eQTL) in human pancreatic islets, conferring increased expression of MTNR1B, which is thought to perturb pancreatic β-cell function. To understand this pathogenic mechanism in detail, we utilized human induced pluripotent stem cells (hiPSC), derived from individuals with T2D carrying the MTNR1B G-allele. Patient-derived fibroblasts were reprogrammed to hiPSC and single-base genome editing by CRISPR/Cas9 was employed to create isogenic lines of either the C/C or G/G genotypes (nonrisk and risk, respectively). In addition, the human embryonic stem cell (hESC) line (HUES4) was subjected to genome editing to create isogenic lines of either the C/C or G/G genotypes. hiPSC and hESC were differentiated into β-like cells, using a 50-day 2D protocol. Single-base genome editing generated cells with the desired genotype at a success rate of > 90%. Expression of stage-specific markers confirmed differentiation of both hiPSC and hESC into β-cells. MTNR1B mRNA levels were consistently low in differentiated β-cells, precluding quantitative analysis of gene expression. Western blot analyses indicated slightly higher levels of the MTNR1B protein in differentiated β-cells carrying the risk allele, which is in accord with the notion that rs10830963 (G-allele) functions as an eQTL in β-cells. Insulin secretion in response to the combination of high glucose and IBMX was comparable between genotypes, whereas the addition of melatonin appeared to reduce insulin secretion more efficiently in cells carrying the G-allele. While our data suggest elevated MTNR1B protein levels in stem cell-derived β-like cells carrying the risk allele, these cells do not appear to be sufficiently mature to establish rs10830963 as an eQTL at the mRNA level. The observed nominal increase in melatonin sensitivity in G-allele-carrying cells is suggestive of a functional contribution of rs10830963 to β-cell dysfunction; however, this interpretation remains tentative and will require further validation in more mature β-cell models.

Keywords: melatonin receptor 1B; single‐nucleotide polymorphism; stem cell derived β‐cells; β‐cell function.

Figure 1

Graphical representation of derivation of hiPSCs and reprogramming into β‐ cells in vitro. (A, B) The image illustrates the workflow of procuring skin biopsies from patients, hiPSC reprogramming, single base genome editing, and differentiation into β‐cells. (B) The procedure creates two isogenic cell lines, which differ only with respect to one single base, that is, the single‐nucleotide polymorphism (SNP) that constitutes the risk allele. Image created in Biorender.

Figure 2

Genome editing of the MTNR1B (rs10830963) risk allele and edited clone selection strategy. (A) The graphical representation shows the sequence surrounding rs10830963 in the intron of the MTNR1B gene. G, indicated by red arrowhead, is the minor allele conferring the increased risk of impaired insulin secretion and future risk of T2D. A guide RNA (in purple)‐targeting Cas9 to this sequence was prepared and introduced to hiPSCs along with a single strand (SS) DNA template which contains a G instead of a C, thus G will be transcribed to a C (green) instead of a G, resulting in a correction of the G‐risk allele. (B–C) Sanger sequencing of DNA in hiPSCs carrying the homozygous (G/G) risk allele of MTNR1B and upon editing; sequencing after single base gene editing in a pool of hiPSC where the vast majority of cells now carries the nonrisk C‐allele. Image created in Biorender (A) A small peak below the major peak indicates that a minor fraction of cells still remains unedited. (B) MF002B2 edited cells were expanded and clonal cell lines were selected as pure edited and control lines. (C) As the editing efficiency was more than 90%, clonal selection for edited MF007C1 cells was not performed. (D) hESCs (HUES4) are normally heterozygous G/C genotype at the rs10830963 locus in MTNR1B gene. Therefore, these cells were edited both ways from G to C and C to G in parallel to create cell lines carrying homozygous nonrisk C/C and risk G/G risk alleles. Editing efficiency was more than 90%, hence no clonal cell lines were prepared. (E) Editing was also assessed by BsaXI restriction digestion where G/G risk genotype resulted in a prominent DNA band at a roughly 510 bp size and C/C nonrisk genotype at a roughly 210 bp size.

Figure 3

hiPSC and hESC lines and β‐cell differentiation characterization. Immunocytochemistry images of differentiated hiPSCs and hESCs that display C‐peptide⁺ cells (β‐cells in green) co‐expressing PDX1 (red), and NKX6.1 (blue) in fully differentiated cells at Day 50 of differentiation (A–H). The scale bar represents 100 μm, and the images were captured at 20× magnification with a confocal microscope. Gene expression profile of INS and PDX1 in hiPSC and hESC derived β‐cells differentiated for 50 days (I–L) (N = 3 biological replicates, mean ± SD).

Figure 4

MF002B2 hiPSC line, clonal selection and MTNR1B expression. Edited nonrisk (C/C) and risk (G/G) clones were selected based on BsaXI restriction digestion genotyping (A). All clones with clear bands were selected (lower panel in A). At the end of differentiation, cells (bulk) were assessed for gene expression of INS (B) and MTNR1B (C) (N = 6, mean ± SD; each data point represents one individual differentiation experiment from each selected clone). (D, E) MTNR1B protein levels were assessed at the end of differentiation in bulk cells and presented as individual bars for each clone (D: white bars edited C/C clones and black bars, unedited G/G clones) as well as cumulatively (E) (N = 4–6, mean ± SD; each data point represents one individual differentiation experiment from each selected clone, (p > 0.05 was considered significant).

Figure 5

MTNR1B expression in fully differentiated β‐cells. Images of GFP expression (green) in hiPSC and hESC derived β‐cells post lentivirus infection (A–C). Graphs show INS (D, F, H) and MTNR1B (E, G, I) expression in sorted GFP + β‐cells (N = 3, mean ± SD; each data point represents one individual differentiation experiment). Immunocytochemistry images show robust C‐Peptide+ cells (β‐cells, green) (J), MTNR1B (red) expression (K), with merge of the two stainings (L) and enhanced image displaying colocalization (L*). Scale bar represents 100 μm, and the images were captured at 10× magnification with a confocal microscope.

Figure 6

Insulin secretion in differentiated cells. Fold secreted insulin (normalized to low glucose) in hiPSC derived β‐cells (A–C). Differences in fold (1 mM glucose (LG), 20 mM glucose (HG)) and HG + IBMX (50 μM) (A), stimulation with HG and melatonin (100 nM) (B). Stimulation with the combination of HG + IBMX (50 μM) + melatonin (100 nM) (C). Fold secreted insulin (normalized to low glucose) in hESC derived β‐cells (D–F). Differences in fold (1 mM glucose (LG), 20 mM glucose [HG]) and HG + IBMX (50 μM) (D), stimulation with HG and melatonin (100 nM) (E). Stimulation with the combination of HG + IBMX (50 μM) + melatonin (100 nM) (F). (N = 3 biological replicates, mean ± SD).