Fanconi anemia (FA) pathway is required for the repair of DNA interstrand crosslinks (ICL). ICLs are caused by genotoxins, such as chemotherapeutic agents or reactive aldehydes. Inappropriately repaired ICLs contribute to hematopoietic stem cell (HSC) failure and tumorigenesis. While endogenous acetaldehyde and formaldehyde are known to induce HSC failure and leukemia in humans with FA, the effects of other toxic metabolites in FA pathogenesis have not been systematically investigated. Using a metabolism-focused CRISPR screen, we found that ALDH9A1 deficiency causes synthetic lethality in FA pathway-deficient cells. Combined deficiency of ALDH9A1 and FANCD2 causes genomic instability, apoptosis, and decreased hematopoietic colony formation. Fanca-/-Aldh9a1-/- mice exhibited an increased incidence of ovarian tumors. A suppressor CRISPR screen revealed that the loss of ATP13A3, a polyamine transporter, resulted in improved survival of FANCD2-/-ALDH9A1-/- cells. These findings implicate high intracellular polyamines and the resulting 3-aminopropanal or acrolein in the pathogenesis of FA. In addition, we find that ALDH9A1 variants may be modifying disease onset in FA patients.

Fanconi anemia (FA) is the most common inherited bone marrow failure (BMF) syndrome and is caused by impaired DNA interstrand crosslink repair. FA patients usually develop BMF during the first decade of life, prior to any known exposure to exogenous crosslinking agents. Therefore, endogenous sources of DNA damage likely play an important role in the pathogenesis of FA. We previously identified loss of ALDH9A1 as a significant source of endogenous DNA damage using a metabolism-focused CRISPR knockout (KO) screen. This finding was validated using Jurkat cells as well as human hematopoietic stem and progenitor cells. Here, we present updates of our project. To determine whether endogenous DNA damage was induced by the combined loss of FANCD2 and ALDH9A1, we investigated markers of DNA damage in bulk-edited cells. We found that the numbers of chromosomal breaks, 53BP1 foci, and gamma-H2AX foci were increased in FANCD2-/-ALDH9A1-/- cells, compared with single KO or wildtype (WT) controls, in the absence of an exogenous DNA damaging agent. These findings are consistent with spontaneously increased basal levels of DNA damage in FANCD2-/-ALDH9A1-/- cells. To study in vivo BMF and tumorigenesis phenotypes of ALDH9A1 deficiency in the setting of FA, we generated a mouse model. Fanca-/-Aldh9a1-/- mice showed the lowest frequency of long-term hematopoietic stem cells and lineage-negative Sca-1-positive cKit-positive cells, however the differences were not significant compared with control groups. While we did not observe aplastic anemia or leukemia, we found a higher incidence of solid tumors, most notably ovarian tumors and hepatocellular carcinoma in aged Fanca-/-Aldh9a1-/- mice. This suggests that the level of endogenous reactive aldehydes created by ALDH9A1 deficiency in mouse is not high enough to cause full-blown hematopoietic phenotypes, most likely due to redundant detoxification pathways. However, in tissues where ALDH9A1 is active and non-redundant, the low level of endogenous DNA damage accumulating over time causes solid tumors. To identify the specific reactive aldehydes responsible for DNA damage in ALDH9A1 deficiency, we performed a growth selection screen using FANCD2-/-ALDH9A1-/- cells. We found that the loss of ATP13A3 conferred survival advantage to FANCD2-/-ALDH9A1-/- cells. ATP13A3 transports endocytosed polyamines into the cytosol where polyamines can be metabolized by serum amine oxidases into 3-aminopropanal, a reactive aminoaldehyde. 3-aminopropanal also undergoes spontaneous decomposition to acrolein, a well-known reactive aldehyde carcinogen. Finding that the loss of ATP13A3 rescues that FANCD2-/-ALDH9A1-/- cells indicates that 3-aminopropanal and/or acrolein induces endogenous DNA damage requiring the Fanconi anemia pathway function for repair. Finally, to determine the contribution of ALDH9A1 variants to clinical manifestations of FA patients, we performed targeted sequencing of DNA from FA patients. We identified five missense variants, four of which had high CADD scores (>20). ALDH9A1 cDNA containing missense variants with high CADD scores expressed in ALDH9A1-/- Jurkat cells resulted in lower protein expression than the WT cDNA. Cell culture supernatant from cells expressing the variant cDNAs also had increased aldehyde levels as assessed by fluorometric assays, suggesting decreased enzymatic activity of the variant proteins. The patients with ALDH9A1 missense variants with high CADD scores had either early hematologic onset of FA (n=3; two patients before age 1 and one patient before age 4) or AML (n=1). In conclusion, we showed that the loss of ALDH9A1 generates endogenous DNA damage necessitating the FA pathway for its repair. Synthetic lethality caused by the combined loss of FANCD2 and ALDH9A1 was rescued by the loss of ATP13A3, which suggests that 3-aminopropanal is the culprit aminoaldehyde that accumulates in ALDH9A1-deficient cells and results in DNA damage. Functionally deleterious ALDH9A1 variants were observed in some FA patients with early onset of disease suggesting that ALDH9A1 could be a modifier of FA in humans. Disclosures Sridhar: Deciphera Pharmaceuticals: Current Employment; CRISPR Therapeutics: Ended employment in the past 24 months. White: Regeneron Pharmaceuticals: Current Employment. Smogorzewska: Rocket Pharmaceuticals: Research Funding.

Fanconi anemia (FA) is the most common inherited bone marrow failure (BMF) syndrome. Impaired DNA interstrand crosslink (ICL) repair is the underlying mechanism for BMF in FA. FA patients usually develop BMF during the first decade of life, prior to any known exposure to exogenous crosslinking agents. Therefore, endogenous sources of DNA damage are most likely to play an important role in the pathogenesis of FA. Metabolic by-products, such as reactive aldehydes, have been implicated in the acceleration of BMF or leukemia in both humans and mice lacking the ICL repair pathway. However, the potential contribution of other detoxifying metabolic enzymes to genome maintenance has not been systematically investigated. Identification of all sources of endogenous DNA damage will allow us to develop novel strategies to prevent DNA damage from arising, and possibly prevent BMF and leukemia in FA as well as other BMF syndromes. To determine whether detoxifying enzymes other than ALDH2 or ADH5 play a role in the protection of HSPC, we performed a metabolism-focused CRISPR/Cas9 synthetic lethality screen (3000 metabolism genes, 10 sgRNA per gene), using wild-type and FANCD2-/- Jurkat cells. From the screen, we identified ALDH9A1 as the most significantly depleted gene in FANCD2-/- Jurkat cells compared with wild-type. Eight out of ten sgALDH9A1 were significantly depleted in FANCD2-/- Jurkat cells, indicating robust effect of the ALDH9A1 knockout. ADH5, a known synthetic lethal gene with FA, was also depleted in FANCD2-/- Jurkat cells, but to a lesser degree. In vitro fluorescence-based competition assay confirmed synthetic lethal interaction between the two genes, in two independent FANCD2-/- Jurkat clones. To determine whether ALDH9A1 deficiency also caused cell death in FA-deficient human hematopoietic stem progenitor cells (HSPCs), we performed an in vitro validation assay using human umbilical cord blood (UCB). UCB CD34+ cells were edited by ribonucleoprotein delivery of Cas9 and sgRNA, in which either of sgCTRL, sgFANCD2 or sgALDH9A1, or both sgFANCD2 and sgALDH9A1 were used. Edited cells were grown on methylcellulose for 10 to 14 days, after which individual colonies were scored and harvested for sequencing. While CD34+ cells that were targeted by both sgFANCD2 and sgALDH9A1 (double KO) achieved lower editing efficiency for each gene compared with cells targeted by single guides, they produced the fewest hematopoietic colonies and the lowest frequency of GEMM (Granulocyte, Erythrocyte, Macrophage and Megakaryocyte) colonies. We observed fewer colonies targeted for both genes (biallelic double KO; observed to expected ratio 0.33) as compared to either single gene KO. These results suggest that loss of ALDH9A1 is deleterious in FANCD2-deficient HSPC. Lastly, we generated a Fanca-/-Aldh9a1-/- mouse model to examine the in vivo hematopoietic phenotype due to increased endogenous aldehydes. These mice were born at the Mendelian ratio without significant anomalies except rare cases of eye abnormalities. At three months of life, Fanca-/-Aldh9a1-/- mice had lower platelet counts than wild-type, Fanca-/- or Aldh9a1-/- control mice, but total white blood counts and hemoglobin levels were similar between groups. A follow-up result of this mouse model will be presented at the meeting. In conclusion, we identified that cells with ALDH9A1 deficiency require the FA pathway for survival. ALDH9A1 may protect human and mouse HSPC that are deficient in the FA pathway from DNA damage and cell death. Disclosures No relevant conflicts of interest to declare.