Multiplexed CRISPR gene editing in primary human islet cells with Cas9 ribonucleoprotein

Abstract

Successful genome editing in primary human islets could reveal features of the genetic regulatory landscape underlying β cell function and diabetes risk. Here, we describe a CRISPR-based strategy to interrogate functions of predicted regulatory DNA elements using electroporation of a complex of Cas9 ribonucleoprotein (Cas9 RNP) and guide RNAs into primary human islet cells. We successfully targeted coding regions including the PDX1 exon 1, and non-coding DNA linked to diabetes susceptibility. CRISPR-Cas9 RNP approaches revealed genetic targets of regulation by DNA elements containing candidate diabetes risk SNPs, including an in vivo enhancer of the MPHOSPH9 gene. CRISPR-Cas9 RNP multiplexed targeting of two cis-regulatory elements linked to diabetes risk in PCSK1, which encodes an endoprotease crucial for Insulin processing, also demonstrated efficient simultaneous editing of PCSK1 regulatory elements, resulting in impaired β cell PCSK1 regulation and Insulin secretion. Multiplex CRISPR-Cas9 RNP provides powerful approaches to investigate and elucidate human islet cell gene regulation in health and diabetes.

Keywords: Techniques in genetics; biology experimental methods; cell biology; human genetics.

Graphical abstract

Figure 1

Efficient CRISPR/Cas9 RNP-mediated targeting of PDX1 in primary human islet cells (A) Schematic of the human pseudoislet electroporation system with CRISPR/Cas9 RNP complexes. (B) Fragment of human PDX1 exon 1 sequence, showing the sgRNA sequence in gray. (C) Human pseudoislets 1 day post-electroporation with Cas9-EGFP RNP and PDX1 sgRNAs complexes (top panel: bright field; bottom panel: blue light, 488 nm, scale bar: 500 μm). (D) Quantification of editing frequency mapped to the reference amplicon using CRISPResso analysis. As expected, the mutations cluster around the predicted cleavage position based on the sgRNA sequence (expected cleavage site indicated by a vertical dotted line; sgRNA sequence: violet line). (E) Quantification of indels after CRISPR/Cas9 RNP targeting of PDX1 (PDX1, red) or using a control sgRNA sequence (Cas9, gray) (n = 6 independent human islet donors). (F) Total PDX1⁺ nuclei scored relative to total cells (DAPI+) counted from immunostaining. See also Figure S2 (n = 3 independent donor samples). (G) qRT-PCR of pseudoislets, CRISPR-PDX1 (red), normalized to the CRISPR-Control (n = 3 independent donors). (H) Total Insulin content of CRISPR/Cas9 electroporated cells normalized to genomic DNA (gDNA) content (n = 4 independent donor samples). Data are presented as mean values ±SE. Two-tailed t tests were used to generate p values. See also Figures S1–S3.

Figure 2

Linking candidate cis-regulatory genomic regions to their target genes in human islet cells (A) Genome browser tracks of the MPHOSPH9 loci, highlighting in yellow a candidate cis-regulatory genomic region within the PITPNM2 gene, and linked by pC-HiC to PITPNM2, MPHOSPH9, and C12orf16. See also Figure S4. The predicted enhancer (red, yellow highlight) was targeted with two sgRNAs (sg1 and sg2, turquoise arrows) designed to delete this element. (B) Human islet cells expressing GFP one day following electroporation with the CRISPR/Cas9 RNP complexes. (C) Quantification of editing frequency by each sg RNA mapped to the reference amplicon using CRISPResso analysis. See also Figure S5. (D) Histogram showing deletion of the region in between the two sgRNAs, generated with CRISPResso. (E) RT-qPCR showing reduced expression of MSPHOSPH9 and INS, but not PITPNM2, c12orf65, and ABCC9 after deletion of the putative enhancer site with CR_PEnh compared to the Cas9 control. (F) Glucose-Stimulated Insulin secretion for the control (Cas9) versus CR_PEnh groups. (G) Total Insulin content for the Cas9 control versus CR_PEnh groups. Data are presented as mean values ±SE. Two-tailed t tests were used to generate p values. ∗∗p < 0.05, ∗∗∗p < 0.01.

Figure 3

CRISPR/Cas9 RNP electroporation in human islet cells allows multiplex targeting of regulatory regions (A) Genome browser tracks of PCSK1 and its regulatory elements: PCSK1 Promoter (PCSK1-Prom) and PCSK1 Enhancer (PCSK1-Enh). Regulatory regions that show glucose-induced H3K27ac accessibility are highlighted: PCSK1-Prom (light blue, sg1 and sg2 are designed to induce a deletion of the turquoise region) and PCSK1-Enh (light pink, sg3 and sg4 are designed to induce a deletion of the dark pink region). Accessible chromatin regions in the human islets are shown by ATAC-seq, H3K4me3, and H3K27ac ChIP-seq. See also Figure S6. (B) Schematics of the human pseudoislet CRISPR/Cas9 electroporation approach used for targeting of PCSK1-Prom and PCSK1-Enh, or simultaneous targeting of PCSK1-Prom+Enh, followed by culture at either basal (2.8 mM) or high (16.7 mM) glucose concentrations. (C) qRT-PCR of PCSK1 in pseudoislets at 2.8, 5.6, and 16.7 mM glucose (n = 3 independent donors). (D and E) CRISPResso quantification of editing efficiency on (D) PCSK1 Promoter sequence and (E) PCSK1 enhancer sequence, after targeting with CRISPR/Cas9 control (Cas9-sgNT), PCSK1-Prom, PCSK1-Enh, or PCSK1-Enh+Prom (n = 3 independent donors for PCSK1-Enh and PCSK1-Prom and n = 4 for PCSK1-Enh+Prom. See also Figures S7–S10. (F–H) Measurements of PCSK1 expression 5 days after CRISPR/Cas9 RNP targeting of (F) PCSK1-Prom, (G) PCSK1-Enh, and (H) PCSK1-Enh+Prom compared to a control (Cas9-2sg) in human pseudoislets and culture at 2.8 mM versus 16.7 mM glucose (n = 5 independent donors). Data are presented as mean values ±SE. Two-tailed t tests were used to generate p values.

Figure 4

Selective impairment of PCSK1 expression and impaired Insulin processing and secretion following CRISPR/Cas9 RNP targeting of PCSK1 regulatory elements (A and B) RT-qPCR after CRISPR/Cas9 RNP targeting of PCSK1-Prom (turquoise), PCSK1-Enh (pink), or PCSK1-Enh+Prom (green): (A and B) Following culture at 5.6 mM glucose, and measurement of mRNA levels of (A) PCSK1 and (B) PCSK2 and GCG (n = 5 donors). (C and D) Following culture at 2.8 mM versus 16.7 mM glucose and measurement of mRNA levels of glucose regulated (C) INS (n = 3) and (D) IAPP expression (n = 5). (E) Scheme of the PCSK1 locus, showing PCSK1 neighboring genes. (F) RT-qPCR of CAST and ELL2 following CRISPR/Cas9 RNP targeting of PCSK1 regulatory regions (n = 5 independent donors for CAST and n = 4 for ELL2). Data are presented as mean values ±SE. Two-tailed t tests were used to generate p values. ∗p < 0.05. (G–L) (G) Scheme of electroporation of CRISPR/Cas9 RNP complexes with sgRNAs targeting PCSK1-Prom (turquoise bars), PCSK1-Enh (pink bars), PCSK1-Enh+Prom (green bar), or the control Cas9-sgNT followed by measurements of: (H) Glucose-stimulated Insulin Secretion (n = 5), (I) Total INS content (n = 5), (J) Glucose-stimulated Proinsulin Secretion (n = 4), (K) Total Proinsulin Content (n = 6–7), (L) Ratio of Proinsulin/Insulin (n = 6–7). See also Figure S11. Data are presented as mean values ±SE. Two-tailed t tests were used to generate p values. ∗p < 0.05, ∗∗p < 0.005.