Loss of ATG5 in Humans Causes Syndromic Congenital Dyserythropoietic Anemia with Impaired Mitophagy in Late Stages of Terminal Erythropoiesis

Autophagy is a vital process for cellular homeostasis wherein degradation of intracellular components and organelles is achieved via engulfment into double-membrane vesicles, autophagosomes. It plays a key role in terminal erythropoiesis and reticulocyte maturation, where timely clearance of mitochondria and ribosomes is necessary for the formation of healthy red blood cells. Deletion of autophagy-related (ATG) genes in mouse models and ex vivo models of human erythropoiesis have shown delayed terminal differentiation, impaired enucleation and reduced red blood cell production. Interestingly, hematopoietic-specific conditional deletion of Atg7 in mice led to lethal anemia (Mortensen et al, PNAS 2010) while humans born with severe loss or complete absence of ATG7 presented with neurodevelopmental disorder, but no reported anemia (Collier et al, NEJM 2021), demonstrating differences between mouse and human erythropoiesis. Atg5-null mice suffer from neonatal lethality but can be rescued by transgenic ATG5 expression in neurons, allowing investigation of the effects of ATG5 deficiency in other tissues (Yoshii et al, Developmental Cell 2016). These mice were found to develop hypochromic anemia due to iron malabsorption, correctable by iron injection. However, two siblings with a homozygous missense mutation in ATG5, that when modeled in yeast, caused a 30-50% reduction of induced autophagy, were described to have congenital ataxia, mental retardation, and developmental delay but no anemia (Kim et al. eLife 2016).

We present here a patient with compound heterozygous pathogenic variants in ATG5, leading to severe ATG5 protein deficiency, associated with syndromic congenital dyserythropoietic anemia, phenocopied in ex vivo erythropoiesis cultures using Bristol Erythroid Adult Line 2 (BEL-A2) cells after CRISPR-Cas9 mediated deletion of ATG5.

A 4-month old male infant with transfusion-dependent anemia and signs of dyserythropoiesis since birth was enrolled in the Congenital Dyserythropoietic Anemia (CDA) Registry of North America (NCT02964494). He was born prematurely after 32 weeks gestation, complicated by intrauterine growth retardation and prenatal hydrops fetalis. His phenotype after delivery included respiratory failure, hepatosplenomegaly, and severe functional and anatomic neurological features (seizure disorder, lissencephaly, hypogenesis of the corpus callosum and a hypoplastic vermis). He remained in intensive care, with no improvement until death at 5 months of age, attributable to panlobular acute necrotizing bronchopneumonia superimposed on severe bronchopulmonary dysplasia, with CDA, adrenal insufficiency, and neurodevelopmental abnormalities.

Microarray analysis detected a maternally inherited interstitial deletion of 1.8 Mb of DNA from the long arm of chromosome 6 (6q21), encompassing fully ATG5, QRSL1, RTN4IP1 and the final exon of PDSS2, all associated with autosomal recessive Mendelian diseases. Whole Genome Sequencing (WGS) of the proband and parents with trio-analysis revealed a paternally-inherited splicing variant in ATG5 (c.236+1G>T) which appeared homozygous due to the complete deletion of the gene in trans. Western blot of cell lysate from cultured patient-derived fibroblasts confirmed complete absence of ATG5 as well as of LC3-II. We have performed CRISPR/Cas9-mediated ATG5 deletion in the BEL-A2 cell line. Differentiation of BEL-A2 ATG5-knock-out (ko) clones versus control demonstrated that ATG5 loss did not significantly affect the rate of erythroblast differentiation or survival but did cause dyserythropoiesis with binucleation and nuclear atypia, progressively worsening with maturation. Most importantly, impaired mitophagy was noted in the stage of orthochromatic erythroblasts, with mitochondria still present in ATG5-ko red cells after enucleation. Studies utilizing patient and healthy control-derived induced pluripotent stem cells (iPSCs) are ongoing to further evaluate the role of ATG5 in another model of human erythropoiesis in vitro.

This case and the associated in vitro studies demonstrate that ATG5 is indispensable for terminal erythropoiesis in humans in contrast to the finding on murine studies that ATG7 is more essential to red cell development, providing another example of differences between murine and human erythropoiesis.