Forced enhancer-promoter rewiring to alter gene expression in animal models

Abstract

Transcriptional enhancers can be in physical proximity of their target genes via chromatin looping. The enhancer at the β-globin locus (locus control region [LCR]) contacts the fetal-type (HBG) and adult-type (HBB) β-globin genes during corresponding developmental stages. We have demonstrated previously that forcing proximity between the LCR and HBG genes in cultured adult-stage erythroid cells can activate HBG transcription. Activation of HBG expression in erythroid cells is of benefit to patients with sickle cell disease. Here, using the β-globin locus as a model, we provide proof of concept at the organismal level that forced enhancer rewiring might present a strategy to alter gene expression for therapeutic purposes. Hematopoietic stem and progenitor cells (HSPCs) from mice bearing human β-globin genes were transduced with lentiviral vectors expressing a synthetic transcription factor (ZF-Ldb1) that fosters LCR-HBG contacts. When engrafted into host animals, HSPCs gave rise to adult-type erythroid cells with elevated HBG expression. Vectors containing ZF-Ldb1 were optimized for activity in cultured human and rhesus macaque erythroid cells. Upon transplantation into rhesus macaques, erythroid cells from HSPCs expressing ZF-Ldb1 displayed elevated HBG production. These findings in two animal models suggest that forced redirection of gene-regulatory elements may be used to alter gene expression to treat disease.

Keywords: MT: Oligonucleotides; Therapies and Applications; enhancer-promoter interactions; fetal hemoglobin; forced chromatin looping; hemoglobin switching; preclinical animal models; sickle cell disease; vector optimization.

Graphical abstract

Figure 1

In vitro testing of ZF-Ldb1 constructs (A) Model of controlling chromatin looping via ZF-Ldb1 constructs to reprogram the β-globin locus (modified from Deng et al.¹⁹). Red⁺ indicates the relative degree of gene expression. (B) Specific target sites within the β-globin locus for each of the ZF-Ldb1 constructs. (C) Anti-HA ChIP-qPCR performed in K562 cells shows the specificity of all zinc-finger (ZF) constructs to the HBG locus. Data are displayed as signal in the immunoprecipitation (IP) fraction compared with total input; n = 1. (D) Expression levels of HbF as a percentage of total globins (HbF + HbA) by cation-exchange HPLC of GFP⁺ cells on day 15 of CD34⁺in vitro culture following lentiviral transduction of CD34⁺ primary human cells with either GFP control or ZF-Ldb1 constructs. n = 2–3 independent donors. Statistical analyses were done using one-way ANOVA. Error bars represent standard deviation. ns, not significant; ∗p < 0.05; ∗∗p < 0.01. (E) RNA-seq analysis of in-vitro-differentiated erythroid cells on day 13 of culture following lentiviral transduction of CD34⁺ primary human cells with either GFP control or ZF-Ldb1 constructs. n = 2 independent donors. Note that multiple genes may be overlapping in this visualization. Red dot, absolute fold change > 4 and false discovery rate (FDR) < 0.05; blue dot, absolute fold change between 1.5 and 4 and FDR < 0.05; black dot, FDR > 0.05 (not significantly changed). (F) Number of DEGs (absolute fold change greater than 1.5 and FDR < 0.05) among ZF constructs. TSS, transcription start site; HS2, DNAse hypersensitivity site 2 of the LCR; HBE, ε-globin; HBG, γ-globin; HBD, δ-globin; HBB, β-globin; LCR, locus control region.

Figure 2

Forced chromatin looping in the humanized BERK mouse model drives in vivo induction of γ-globin expression (A) Experimental design detailing myeloablative transplantation of lentivirally transduced syngeneic mouse bone marrow with either GFP control or GG38/GG63-Ldb1-GFP constructs into heterozygous (mouse β^A human β^S) BERK mice, followed by analysis of HbF and γ-globin expression 10 weeks post transplantation. (B) RP-HPLC analysis and quantification of GFP⁺ sorted bone marrow cells from GFP control-, GG38-, or GG63-transplanted mice. n = 2–3 transplanted mice for each condition for RP-HPLC quantification. Error bars represent standard deviation. (C) Representative western blot analysis and quantification of γ-globin in GFP⁺ sorted cells from GFP control-, GG38-, or GG63-transplanted mice. n = 1–2 transplanted mice for each condition for western blot quantification. Error bars represent standard deviation.

Figure 3

Optimization of ZF-Ldb1 constructs for use in rhesus macaque transplantation (A) Schematic of ZF-Ldb1 constructs. (B) Experimental design for testing HbF induction of ZF constructs. Adult CD34⁺ human cells were transduced at an MOI of 20 or 100 on day 0, and HbF was measured by HPLC analysis on days 9 and 12 of in vitro erythroid differentiation culture. (C) HPLC analysis following transduction of ZF constructs in two independent CD34⁺ donor cells on day 12 of culture. HPLC elution tracings (left) and quantification of HPLC peaks (right) illustrate high levels of HbF induction. Vector copy number (VCN D12) and percent Venus/mEmerald positivity (%Fluor D12) are displayed for each vector tested. n = 2 independent donors.

Figure 4

Robust ex vivo γ-globin induction in rhesus macaque progenitor cells transduced with optimized ZF constructs (A) Hemoglobin electrophoresis (top) and RP-HPLC (bottom) in differentiated rhesus macaque CD34⁺ HSPCs; n = 1. (B) Flow cytometric analysis of mEmerald or Venus vector expression in differentiated rhesus macaque CD34⁺ HSPCs. Cells were transduced with a lentivirus (MOI = 50 for CD34⁺ and MOI = 10 for PBMCs) at high cell density (2 × 10⁶ cells/mL) in XVIVO-10 medium + SFT (100 ng/mL each); n = 3. Statistical analyses were done using one-way ANOVA. Error bars represent standard deviation. ∗∗∗p < 0.001. (C and D) Ex vivo analysis of rhesus macaques receiving transplants, showing (C) hemoglobin electrophoresis and (D) RP-HPLC analysis of non-transduced cells (NTCs) and GG38-expressing lentivirus-transduced cells on day 14 of erythroid differentiation of infused products; n = 1. HbF and flow cytometry assays were performed on unsorted bulk erythroid cells. HbC, hemoglobin C; HbS, hemoglobin S/sickle hemoglobin; HbF, hemoglobin F/fetal hemoglobin; HbA, hemoglobin A/adult hemoglobin.

Figure 5

Long-term stable HbF expression was not seen in analysis of bulk peripheral blood from transplanted rhesus macaques (A–D) Analysis of PB from rhesus macaques (n = 2) transplanted with the optimized ANK1-GG38 ZF construct: (A) VCN, (B) Venus (percent), (C) F-cells (percent), and (D) γ-globin (percent) in transplanted animals. n = 2 independent transplanted rhesus macaques. Control: red blood cells (RBCs) from non-transplanted animals.

Figure 6

Significant in vivo γ-globin induction in Venus⁺ RBCs from transplanted rhesus macaques (A and B) Sorted Venus⁺ red blood cells (RBCs) displayed high γ-globin protein expression compared with Venus⁻ RBCs. Shown are the percentage of γ-globin expression (expressed as γ-globin / [β-globin + γ-globin]) in sorted RBCs at (A) 1 year post transplantation and (B) 2 years post transplantation. n = 2 independent transplanted rhesus macaques.