Background

Cytosine arabinoside (ARA-C) is a major active drug against AML cells, and is widely used also in this era of targeted treatments. Resistance to ARA-C remains a significant obstacle to achieving CR and long-term remissions. We previously reported an association between CNN-LOH in chromosome 1p that was more common in samples from older AML patients compared to younger patients (Koren-Michowitz M et al. Leuk Res. 2012 Apr;36(4):467-73). Since AML in older patients often exhibit primary therapy resistance, we hypothesized that genes in the commonly deleted chromosome 1p region could be associated with chemotherapy resistance.

Methods

We performed a CRISPR-Cas9 pooled loss-of-function screen using a library of sgRNAs targeting all the coding genes located on the commonly deleted chromosome 1p region to identify candidate genes whose disruption confers resistance to ARA-C in AML cells. The CRISPR-Cas9 pooled screen was designed using 1,000 non-targeting sgRNAs and 10 sgRNAs per target gene, along with three positive controls (DCK, p53, and SLC29A1). OCI-AML2-Cas9 stable cells were transduced with the pooled sgRNA library and treated with 1 μM ARA-C for 3 days. MAGeCK computational analysis was performed to identify genes whose knockout altered sensitivity to ARA-C. Genes were considered positively selected if p ≤ 0.008 and FDR ≤ 0.4 and negatively selected if p ≤ 0.05 and FDR ≤ 0.002.

Results

Analysis of the pooled screen revealed 13 negatively selected and 8 positively selected candidate genes. Gene ontology (GO) functional enrichment analysis revealed statistically significant enrichment (p < 0.05) among negatively selected targets for RNA binding and DNA-binding transcription factor activity. No significant enrichment was observed among the positively selected genes. Expression of the negatively selected genes in AML samples was retrieved from the TARGET initiative (phs000465). SRSF10 showed the highest differential expression, with a 52-fold increase (p = 1.07 × 10⁻73) in AML samples compared to normal CD34+ cells.

SRSF10 CRISPR-Cas9 single knockout (KO) clones of OCI-AML2 and MOLM13 cell lines exhibited increased sensitivity to ARA-C compared to non-targeting sgRNA controls (IC₅₀ 97 nM and 138 nM for OCI-AML2 cells, 53 nM and 82 nM MOLM13 cells, respectively). This was accompanied with an increase in the proportion of early apoptotic cells and accumulation of cells in the G2/M checkpoint in the KO cells compared to control cells. Co-culture of AML cell lines with HS5 bone marrow stromal cells protected AML cells against ARA-C cytotoxicity, as was previously shown with other anti-AML drugs. Interestingly, SRSF10 KO cells were more sensitive to ARA-C also in this setting.

Conclusions

SRSF10 could be a target for enhancing cytotoxicity of ARA-C in AML. Studies in primary AML cells are being conducted.

This content is only available as a PDF.
2025
Sign in via your Institution