Complete absence of GLUT1 does not impair human terminal erythroid differentiation

Abstract

The Glucose transporter 1 (GLUT1) is one of the most abundant proteins within the erythrocyte membrane and is required for glucose and dehydroascorbic acid (Vitamin C precursor) transport. It is widely recognized as a key protein for red cell structure, function, and metabolism. Previous reports highlighted the importance of GLUT1 activity within these uniquely glycolysis-dependent cells, in particular for increasing antioxidant capacity needed to avoid irreversible damage from oxidative stress in humans. However, studies of glucose transporter roles in erythroid cells are complicated by species-specific differences between humans and mice. Here, using CRISPR-mediated gene editing of immortalized erythroblasts and adult CD34+ hematopoietic progenitor cells, we generate committed human erythroid cells completely deficient in expression of GLUT1. We show that absence of GLUT1 does not impede human erythroblast proliferation, differentiation, or enucleation. This work demonstrates for the first-time generation of enucleated human reticulocytes lacking GLUT1. The GLUT1-deficient reticulocytes possess no tangible alterations to membrane composition or deformability in reticulocytes. Metabolomic analyses of GLUT1-deficient reticulocytes reveal hallmarks of reduced glucose import, downregulated metabolic processes and upregulated AMPK-signalling, alongside alterations in antioxidant metabolism, resulting in increased osmotic fragility and metabolic shifts indicative of higher oxidant stress. Despite detectable metabolic changes in GLUT1 deficient reticulocytes, the absence of developmental phenotype, detectable proteomic compensation or impaired deformability comprehensively alters our understanding of the role of GLUT1 in red blood cell structure, function and metabolism. It also provides cell biological evidence supporting clinical consensus that reduced GLUT1 expression does not cause anaemia in GLUT1 deficiency syndrome.

Figure 1 –. CRISPR-mediated GLUT1 KO on immortalised BEL-A erythroblasts successfully generate reticulocytes.

A Sequencing of SLC2A1 (GLUT1) gene on Ctrl BEL-A, highlighting the gRNA (green) used for CRISPR editing. Sanger sequencing of clonal edited lines shows a homozygous 4bp frameshift mutation on the GLUT1 KO line, and a heterozygotic 16bp deletion on the KD line, both in the vicinity of the cutting site (red line). Flow cytometry histogram of GLUT1 staining in BEL-A erythroblasts (B) and derived reticulocytes (D) from ctrl (green), GLUT1 KD (blue), and GLUT1 KO (orange) cell lines compared with no-stain control (black). Cells were stained with anti-GLUT1 FITC conjugate. C Flow cytometry analysis of cell surface marker expression during differentiation. Cells were co-labelled with anti-band3 primary antibody used in conjunction with an IgG1 APC secondary and anti-α4-integrin FITC conjugate. For day 11, reticulocytes were identified using Hoechst as a nuclear DNA stain. E Bar graph illustrates the percentage GLUT1 expression on reticulocytes derived from indicated cell lines. Data is normalised to endogenous expression of Ctrl BEL-A from the median fluorescence intensity (n=4). Individual data points are shown. Error bars indicate standard error of mean. F Representative images of May-Grünwald and Giemsa-stained cytospins depicting expanding BEL-A erythroblasts (Day 0) and corresponding filtered reticulocytes after 10 day differentiation protocol. 40X magnification. Scale bars = 20μm, shown for each image. G Bar graphs illustrate expression of various membrane proteins on reticulocytes derived from indicated cell lines (n=3). Reticulocytes were identified based on Hoechst negativity. Data is normalised to endogenous expression of Ctrl BEL-A and represents the median fluorescence intensity (n=3). Individual data points are shown. Error bars indicate standard error of mean. H Western blots of lysates obtained from indicated cell lines at day 0 of differentiation and reticulocytes filtered after 10 day protocol, incubated with antibodies to α-adducin, GLUT1, Stomatin, and GAPDH (loading control). Multiple Mann-Whitney U tests were used to test for differences between groups. *p < 0.05. Error bars indicate standard deviation.

Figure 2 –. Primary erythropoiesis is not affected by knocking out GLUT1.

A Schematic diagram of human CD34+ 3-step culture method. PBMCs are isolated from apheresis cones, followed by CD34+ magnetic separation. Cells are then nucleofected on day 3 with Non-Targeting (NT) or GLUT1 specific gRNAs. B Flow cytometry histograms show GLUT1 expression of 3 donors, nucleofected with NT or SLC2A1 targeting gRNAs, on days 5, 6, 14, and 21 of differentiation. For days 14 and 21 Hoechst was used to identify the reticulocytes. Cells were stained with anti-GLUT1 FITC conjugate (n=3) or a no-stain control (black). C Day 21 filtered reticulocytes stained with anti-GLUT1 FITC conjugate (n=3) or a no-stain control (black). D Percentage of GLUT1 negative population on GLUT1-targeted KO (n=3) on days 6, 14 and 21 of differentiation. E Percentage of reticulocytes (Hoechst negative) at day 21 of differentiation on NT control and negative and positive GLUT1 populations of the GLUT1-targeted KO. F Percentage of GLUT1 negative population on GLUT1-targeted KO on reticulocytes pre- and post-filtration. G Bar graph illustrates GLUT1 expression on reticulocytes from NT and GLUT1 negative and positive populations of the GLUT1-targeted KO. Data represents the median fluorescence intensity (n=3). Individual data points are shown. Multiple Mann-Whitney U tests were used to test for differences between groups. *p < 0.05. Error bars indicate standard deviation.

Figure 3 –. Properties of CD34+ GLUT1-KO derived reticulocytes.

A Waterfall indicating progression of control and GLUT1-KO transfected cultures of single donor CD34+ differentiation. Cells were co-labelled with anti-band3 primary antibody used in conjunction with an IgG1 APC secondary and anti-α4-integrin FITC conjugate. For day 20, reticulocytes were identified using Hoechst as a nuclear DNA stain. B Representative cytospins of the same culture were obtained on day 20 post-leukofiltration. 40X magnification. Scale bars = 20μm, shown for each image. C Bar graphs indicate the expression of various membrane proteins on reticulocytes derived from NT or GLUT1-KO primary cultures (n=6, for SMVT n=3, with 2 technical repeats each). D Anti-transferrin receptor (CD71) labelling of on reticulocytes (n=6, open or filled dots indicate 2 separate cultures, each with 3 donors). For both, reticulocytes were leukofiltered on day 20. Data is normalised to endogenous expression of NT control and represents the median fluorescence intensity. E Osmotic resistance analysis calculated based on viable cell counts (flow cytometer) after incubation with decreasing concentrations of NaCl (n=3, 2 technical replicates, ∘ p≤0.0021). F Deformability and G cell area (μm²) were measured under shear stress by Automated Rheoscope Cell Analyser (n=3, N>2000 cells), which elongates cells and measures length over width as deformability parameter. Shaded region represents the standard deviation. H Quantitative analysis of lipid peroxidation detected by a shift in the fluorescence signal after treatment with 25mM cumene hydroperoxide. Data normalized to NT-control (n=3, 3 technical replicates) I Bar graph quantifying Plasmodium falciparum reticulocyte invasion efficiency. Invasion was assessed by flow cytometry using a SYBR-green DNA stain (3 separate parasitaemia percentages, 3 technical replicates) and data was normalised per donor (n=3) to the NT-control. Figure 3 comprises data obtained from two independent cultures, each of 3 donors with figures 3A and B presenting representative data from cultures with FACS sorted GLUT1-KO purity of 99%, E-I 94% and C-D combined data from all 6 donors. A non-parametric Mann-Whitney U test or Kruskal-Wallis test with Bonferroni correction were used to test for differences between groups. *p < 0.05, **p < 0.01, ***p < 0.001. Error bars indicate standard deviation.

Figure 4 –. Proteomics confirms GLUT1 absence in CD34+ GLUT1-KO derived reticulocytes and Metabolite analyses reveals downregulated metabolic processes

A Simplified schematic of multi-omics cell preparation, where CD34+ from 3 donors were nucleofected with either GLUT1- or non-targeting gRNAs, expanded and differentiated into reticulocytes. 10 million filtered reticulocytes were needed for comprehensive analyses of the proteome, metabolome, and lipidome. Made using BioRender. B Box plot comparing GLUT1 protein level between NT and GLUT1 KO reticulocytes as Log₂(Fold change). Box plot analysis (mean±min to max with standard deviation) was performed by RStudio, and significance was calculated upon false discovery rate correction (***p < 0.001).C,D Hierarchical clustering of the Top 50 t-test significant proteins (C) and metabolites (D) between NT and GLUT1 KO CD34⁺-derived reticulocytes. E Schematic representation of Glycolysis, the Polyol pathway, and the Glutathione redox cycle where proteins and metabolites are colour-coded by Log₂ (Fold Change) of GLUT1-KO reticulocytes in relation to NT control. The 10 steps of glycolysis are represented, with glucose and lactate both reduced in the KO while the remaining intermediate products increased. All involved enzymes are decreased (HK, Hexokinase; GPI, Glucose-6-phosphate isomerase; PFK1, Phosphofructokinase-1; TPI1, Triosephosphate isomerase; GAPDH, Glyceraldehyde-3-phosphate dehydrogenase; BPGM, Biphosphoglycerate mutase; PGK, Phosphoglycerate kinase; PGAM, Phosphoglycerate mutase; PK, Pyruvate kinase; LDHA, Lactate dehydrogenase A). There is an imbalance in the Glutathione cycle, as a consequence of increased ROS, characterised by the depletion of the reduced form (GSH) and increase of oxidised glutathione (GSSG) and glutathione peroxidase 1 (GPX1). The Hexose monophosphate (HMP) shunt is upregulated as source of NADPH, needed to maintain glutathione in its reduced form. (G6PD, Glucose-6-phosphate dehydrogenase; PGLS, 6-phosphogluconolactonase; GSR, Glutathione-disulfide reductase). An increase in polyols (such as sorbitol and mannitol) was also detected, which can be converted into Fructose-1,6-phosphate by sorbitol dehydrogenase (SORD) and ketohexokinase (KHK). Created with Biorender.

Figure 5 –. Lipid composition highlights increased oxidant stress to the membrane of GLUT1-KO reticulocytes.

A Box plot showing the comparison of Acyl-carnitines as Log₂(FoldChange, FC), between NT and GLUT1-KO reticulocytes. Box plots analysis (mean±min to max with standard deviation) was performed by RStudio, and significance was calculated upon false discovery rate correction (p < 0.05). B Simplified schematics of the Lands Cycle, capable of repairing damaged lipids (lysophospholipids) generated from increased oxidative stress. The damaged chain is removed by a phospholipase, originating a free fatty acid which is converted into Acyl-CoAs by Acyl-CoA synthetase (ACS) in an ATP-dependent reaction. Acyl-CoAs can be converted into Acyl-carnitines by carnitine palmitoyltransferase (CPT). Lysophospholipid acyltransferases (LPLATs) incorporate undamaged fatty acid chains into membrane lysophospholipids, repairing the lipid membrane. Created in Biorender. C-E Box plots comparing Lysophosphatidylethanolamine (LPE, C), Lysophosphatidylserines (LPS, D) and Lysophosphatidylcholine (LPC, E) between NT and GLUT1-KO reticulocytes, presented as Log²(FC). F Hierarchical clustering of the Top 50 t-test significant lipids between NT and GLUT1 KO CD34⁺-derived reticulocytes.