An erythroid-specific lentiviral vector improves anemia and iron metabolism in a new model of XLSA

Abstract

X-linked sideroblastic anemia (XLSA) is a congenital anemia caused by mutations in ALAS2, a gene responsible for heme synthesis. Treatments are limited to pyridoxine supplements and blood transfusions, offering no definitive cure except for allogeneic hematopoietic stem cell transplantation, only accessible to a subset of patients. The absence of a suitable animal model has hindered the development of gene therapy research for this disease. We engineered a conditional Alas2-knockout (KO) mouse model using tamoxifen administration or treatment with lipid nanoparticles carrying Cre-mRNA and conjugated to an anti-CD117 antibody. Alas2-KOBM animals displayed a severe anemic phenotype characterized by ineffective erythropoiesis (IE), leading to low numbers of red blood cells, hemoglobin, and hematocrit. In particular, erythropoiesis in these animals showed expansion of polychromatic erythroid cells, characterized by reduced oxidative phosphorylation, mitochondria's function, and activity of key tricarboxylic acid cycle enzymes. In contrast, glycolysis was increased in the unsuccessful attempt to extend cell survival despite mitochondrial dysfunction. The IE was associated with marked splenomegaly and low hepcidin levels, leading to iron accumulation in the liver, spleen, and bone marrow and the formation of ring sideroblasts. To investigate the potential of a gene therapy approach for XLSA, we developed a lentiviral vector (X-ALAS2-LV) to direct ALAS2 expression in erythroid cells. Infusion of bone marrow (BM) cells with 0.6 to 1.4 copies of the X-ALAS2-LV in Alas2-KOBM mice improved complete blood cell levels, tissue iron accumulation, and survival rates. These findings suggest our vector could be curative in patients with XLSA.

Graphical abstract

Figure 1.

Characterization of the new lentiviral vector X-ALAS2-LV. (A) Schematic diagram of X-ALAS2 lentiviral vector. (B) VCN analysis of X-ALAS2-LV in NIH-3T3 and HUDEP-2 cells. N = 3 for each LV concentration; the results are shown as median ± SD. (C) Chimerism analysis after BM transplantation (N = 9 for each experimental group). Data are shown as mean ± standard error of the mean (SEM). (D) X-ALAS2-LV VCN analysis 25 weeks after transplantation in the BM and spleen (SPL) samples from TAM and LNP^CD117Cre models (N = 6 in all experimental groups). Data are shown as mean ± SEM. (E) Quantification of human Alas2 gene expression (quantitative reverse transcription polymerase chain reaction [qRT-PCR]) in Lin⁻ cells from Alas2^fl/Y BM. P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗∗∗∗P < .0001. (F) Correlation of VCN (X-ALAS2-LV) and Hb (g/dL) values in the TAM and LNP^CD117Cre models in X-ALAS2 treated mice. P values, R squared, and curve equations are obtained by simple linear regression analysis. Treatment with X-ALAS2-LV extended the survival of animals with VCN <0.6, although these mice eventually showed progressive anemia that required their sacrifice. In animals with VCN >0.6 instead, treatment with X-ALAS2-LV improved the disease phenotype.

Figure 2.

Evaluation of the use of Tamoxifen and LNP^CD117Cre in a conditional Alas2-KO model. (A) Values of Hb, RBC, HCT, and reticulocytes (RETs) in R26CreER^T2 and R26CreER^T2-Alas2^fl/Y animals after administration of TAM. N = 4 for each group. Data are shown as mean ± standard error of the mean (SEM). (B) Representative flow cytometry analysis of the erythroid populations in the BM and spleen in R26CreER^T2 and R26CreER^T2-Alas2^fl/Y animals treated with TAM. Data are shown as FSC-A/CD44 subgated on the Ter119⁺ population. P1+P2/P2 = proerythroblasts/basophilic erythroblasts; P3 = polychromatic erythroblasts; P4 = orthochromatic erythroblasts/reticulocytes; P5 = erythrocytes. (C) Quantification of Alas2 allelic deletion (droplet digital polymerase chain reaction [ddPCR] analysis) of gDNA isolated from Lin⁻ cells from Alas2^fl/Y BM after exposure to LNP^CD117Cre. Data are shown as mean ± SEM (N = 3 independent replicates). P values are determined by the Tukey multiple-comparison test after the 1-way analysis of variance (ANOVA). ∗∗∗∗ P < .0001. (D) Quantification of Alas2 allelic deletion (ddPCR analysis) of gDNA isolated from Lin⁻ cells from Alas2^fl/Y BM, after tamoxifen treatment (with or without BMT) or exposure to LNP^CD117Cre. Data are shown as mean ± SEM (N = 3 independent replicates). P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗P < .05, ∗∗P < .001, ∗∗∗∗ P < .0001.

Figure 3.

Characterization of anemia, survival and ineffective erythropoiesis of Alas2-KO^BM mice in presence of absence of X-ALAS2-LV. (A) Complete blood cell (CBC) panel at the end point in each experimental group. Each model includes animals treated (+) and not treated (−) with X-ALAS2-LV (N = 8 in all experimental groups). Data are shown as mean ± standard error of the mean (SEM). P values are determined by the Tukey multiple-comparison test after the 1-way analysis of variance (ANOVA). ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. Although there is not a significant difference in the Alas2-KO^BM animals treated with tamoxifen or LNP^CD117Cre, the LNP-treated animals did show a reduced absolute reticulocyte count (ARC) compared with the tamoxifen model. Compared with tamoxifen, the LNP technology is more efficient in deleting Alas2 (as shown in Figure 2D), and the cells engrafted already lack the Alas2 gene. In the tamoxifen approach, the process to delete the target gene is slower (as this requires repeated administration of tamoxifen), and more undeleted cells may be present, which still possess the ability to regenerate and mature into reticulocytes. However, these cells are insufficient to support normal erythropoiesis and rescue the animals. (B) Kaplan-Meier analysis of experimental cohorts in TAM and the LNP^CD117Cre models. Each model includes animals treated (LV) and not treated (No LV) with X-ALAS2-LV (N = 9 in all experimental groups). Curve comparison in the TAM model: P < .001, χ² = 17.50 in the Mantel-Cox test. Curve comparison in the LNP^CD117Cre model: P < .01, χ² = 14.18 in the Mantel-Cox test. (C) Representative flow cytometry analysis of the BM and spleen erythroid cells defined by FSC-A/CD-44 within the Ter119⁺ pregated population in TAM and LNP^CD117Cre models. An example of an X-ALAS2 lentiviral vector and nontreated samples (Alas2-KO^BM) are shown for each model. P1+P2/P2 = proerythroblasts/basophilic erythroblasts; P3 = polychromatic erythroblasts; P4 = orthochromatic erythroblasts/reticulocytes; P5 = erythrocytes. (D) Representative blood smears of each experimental group, in animals treated (+X-ALAS2-LV) or not treated (No Vector) with X-ALAS2-LV, from TAM and LNP^CD117Cre cohorts.

Figure 3.

Characterization of anemia, survival and ineffective erythropoiesis of Alas2-KO^BM mice in presence of absence of X-ALAS2-LV. (A) Complete blood cell (CBC) panel at the end point in each experimental group. Each model includes animals treated (+) and not treated (−) with X-ALAS2-LV (N = 8 in all experimental groups). Data are shown as mean ± standard error of the mean (SEM). P values are determined by the Tukey multiple-comparison test after the 1-way analysis of variance (ANOVA). ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. Although there is not a significant difference in the Alas2-KO^BM animals treated with tamoxifen or LNP^CD117Cre, the LNP-treated animals did show a reduced absolute reticulocyte count (ARC) compared with the tamoxifen model. Compared with tamoxifen, the LNP technology is more efficient in deleting Alas2 (as shown in Figure 2D), and the cells engrafted already lack the Alas2 gene. In the tamoxifen approach, the process to delete the target gene is slower (as this requires repeated administration of tamoxifen), and more undeleted cells may be present, which still possess the ability to regenerate and mature into reticulocytes. However, these cells are insufficient to support normal erythropoiesis and rescue the animals. (B) Kaplan-Meier analysis of experimental cohorts in TAM and the LNP^CD117Cre models. Each model includes animals treated (LV) and not treated (No LV) with X-ALAS2-LV (N = 9 in all experimental groups). Curve comparison in the TAM model: P < .001, χ² = 17.50 in the Mantel-Cox test. Curve comparison in the LNP^CD117Cre model: P < .01, χ² = 14.18 in the Mantel-Cox test. (C) Representative flow cytometry analysis of the BM and spleen erythroid cells defined by FSC-A/CD-44 within the Ter119⁺ pregated population in TAM and LNP^CD117Cre models. An example of an X-ALAS2 lentiviral vector and nontreated samples (Alas2-KO^BM) are shown for each model. P1+P2/P2 = proerythroblasts/basophilic erythroblasts; P3 = polychromatic erythroblasts; P4 = orthochromatic erythroblasts/reticulocytes; P5 = erythrocytes. (D) Representative blood smears of each experimental group, in animals treated (+X-ALAS2-LV) or not treated (No Vector) with X-ALAS2-LV, from TAM and LNP^CD117Cre cohorts.

Figure 4.

Pathological evaluation of spleen, liver and bone marrow of Alas2-KO^BM animals in presence or absence of X-ALAS2-LV. (A) Quantification of the spleen/body weight ratio in animals treated (+) or not treated (−) with X-ALAS2-LV (N = 9 in all experimental groups) from TAM (TAM) and LNP^CD117Cre (LNP) models at the end point of each experimental group. Data are shown as mean ± standard error of the mean (SEM). P values are determined by the Tukey multiple-comparison test after 1-way analysis of variance (ANOVA). ∗∗∗∗P < .0001. Representative pictures of spleens were collected at the end point of each experimental group. (B) Representative images of hematoxylin and eosin (H&E) and Prussian blue staining in spleen tissues were collected at the end point of each experimental group. Magnification ×10. (C) Representative images of H&E and Prussian blue staining in the bone marrow and liver tissues were collected at the end point of each experimental group. Magnification ×4.

Figure 5.

Spleen, bone marrow and liver iron quantification with clear formation of ring sideroblasts and iron accumulation in the mitochondria. (A) Quantification of iron accumulation using the Aperio Versa 200 slide scanner and a positive pixel count algorithm to quantify the level of iron accumulation in the different organs. Data are shown as mean ± standard error of the mean (SEM). P values are determined by the Tukey multiple-comparison test after 1-way analysis of variance (ANOVA). ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. (B) Representative images of Prussian blue staining in bone marrow tissues and cell smears. Magnifications ×20 and ×100. (C) Representative electron microscopy images of isolated P3 cells from wild type, Alas2-KO^BM, and X-ALAS2–treated animals. Iron deposits in the mitochondria are highlighted by the orange arrows. Direct magnification is ×100 000 and printed magnification is ×129 000 at 7.0 in.

Figure 6.

Iron metabolism and metabolomic analysis of the Alas2-KO^BM model in presence or absence of X-ALAS2-LV. (A) ERFE, HAMP, EPO, and iron values in blood serum of Alas2-KO^BM animals transduced or not with +/− X-ALAS2-LV at the end point of each experimental group (N = 4-9 in all experimental groups). Data are shown as mean ± standard error of the mean (SEM). P values are determined by the Tukey multiple-comparison test after the 1-way analysis of variance (ANOVA). ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. WT is a C57BL/6J used as a reference. (B) Metabolomic analysis of polychromatic erythroid cells, sorted from WT and Alas2-KO^BM animals (N = 6 for each group). Data are shown as mean ± SEM. P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. (C) Glycolysis analysis of polychromatic erythroid cells, sorted from WT and Alas2-KO^BM animals (N = 6 for each group). Data are shown as mean ± SEM. P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. (D) Relative gene expression analysis of glycolytic pathway (quantitative reverse transcription polymerase chain reaction [qRT-PCR]) in polychromatic erythroid cells, sorted from WT and Alas2-KO^BM animals (N = 6 for each group). P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗P < .05, ∗∗P < .01.

Figure 6.

Iron metabolism and metabolomic analysis of the Alas2-KO^BM model in presence or absence of X-ALAS2-LV. (A) ERFE, HAMP, EPO, and iron values in blood serum of Alas2-KO^BM animals transduced or not with +/− X-ALAS2-LV at the end point of each experimental group (N = 4-9 in all experimental groups). Data are shown as mean ± standard error of the mean (SEM). P values are determined by the Tukey multiple-comparison test after the 1-way analysis of variance (ANOVA). ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. WT is a C57BL/6J used as a reference. (B) Metabolomic analysis of polychromatic erythroid cells, sorted from WT and Alas2-KO^BM animals (N = 6 for each group). Data are shown as mean ± SEM. P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. (C) Glycolysis analysis of polychromatic erythroid cells, sorted from WT and Alas2-KO^BM animals (N = 6 for each group). Data are shown as mean ± SEM. P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. (D) Relative gene expression analysis of glycolytic pathway (quantitative reverse transcription polymerase chain reaction [qRT-PCR]) in polychromatic erythroid cells, sorted from WT and Alas2-KO^BM animals (N = 6 for each group). P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗P < .05, ∗∗P < .01.

Figure 7.

Apoptosis and inflammation gene expression analysis in the Alas2-KO^BM model and effect of the treatment with X-ALAS2-LV. (A) Relative gene expression analysis of apoptosis pathway (quantitative reverse transcription polymerase chain reaction [qRT-PCR]) in polychromatic erythroid cells, sorted from WT and Alas2-KO^BM animals (N = 6 for each group). P values are determined by the Tukey multiple-comparison test after the 1-way analysis of variance (ANOVA). ∗P < .05. (B) Relative gene expression analysis of Tnf, mIl-6, and Hmox1 (qRT-PCR) in polychromatic erythroid cells, sorted from WT and Alas2-KO^BM animals (N = 6 for each group). P values are determined by the Tukey multiple-comparison test after the 1-way ANOVA. ∗∗P < .01.