CHAF1B (Chromatin Assembly Factor 1B) functions as a histone chaperone within the DNA replication machinery. It is almost universally overexpressed in cancers, and we have previously shown that this upregulation is necessary to enforce the differentiation block found in acute myeloid leukemia (AML) through CHAF1B's chromatin binding function. In this study, we found that CHAF1B creates a pro-leukemic state in hematopoietic cells by repressing the expression of the E3 ubiquitin ligase TRIM13. Our data suggest that TRIM13 is a strong promoter of entry into the cell cycle functioning through novel nuclear localization and activating ubiquitination of key cell cycle genes. We propose that repression of TRIM13 by CHAF1B is a critical event for leukemogenesis because unchecked entry into cell cycle leads to loss of self-renewal and eventual exhaustion of the tumor. First, we measured the expression levels of CHAF1B and TRIM13 in several bone marrow aspirates from primary pediatric AML patients using immunohistochemistry. TRIM13 was expressed in healthy marrow and almost undetectable upon leukemic progression. CHAF1B, on the other hand, had low to moderate expression in healthy tissues and substantially higher expression upon leukemic progression and relapse. We found restoration of healthy CHAF1B and TRIM13 expression levels upon resolution of the tumor in matched patient aspirates upon remission. To determine if CHAF1B is capable of directly repressing TRIM13 expression, first we observed CHAF1B occupancy at the TRIM13 promoter by ChIPseq in multiple AML cell lines. Then we found loss of CHAF1B by shRNA resulted in increased TRIM13 transcription, and overexpression of CHAF1B by dox-inducible lentivirus drove TRIM13 transcriptional repression. Overexpression of a CHAF1B chromatin binding deficient mutant (CHAF1BRRAA) was unable to repress TRIM13, suggesting that CHAF1B chromatin binding is necessary for TRIM13 repression. This led us to hypothesize that TRIM13 may be a novel tumor suppressor in AML. To test this hypothesis, we overexpressed TRIM13 in AML cell lines and several patient-derived xenografts (PDXs) spanning many subtypes of AML and observed consistent repression of colony formation, increase in apoptosis, and reduction of leukemic progression in vivo. Conversely, loss of TRIM13 expression by shRNA led to an increase in colony formation in vitro and enhanced leukemogenicity of PDXs and AML cell lines in vivo. To better understand the mechanistic contribution of TRIM13 to myeloid leukemogenesis, we started by identifying potential catalytic targets of TRIM13. Since TRIM13 is a RING-type E3 ubiquitin ligase, we used CRISPR to generate in-frame RING domain deleted alleles of TRIM13 (T13RINGdel) and performed ubiquitin proteomics comparing our T13RINGdel and TRIM13 WT clones. We expected that loss of the catalytic function would lead to stabilization of pro-leukemic proteins as has been shown in solid tumor models. We did not find that, but we did observe a profound reduction in ubiquitination of proteins associated with the cell cycle, including CCNA1, CCNA2, CCNB2, and CDK1. Additionally, the total levels of these cell cycle-promoting proteins was reduced in T13RINGdel cells suggesting the ubiquitination event leads to their stability and activity. We confirmed that our T13RINGdel clones of AML cell lines had reduced entry and progression through cell cycle. Using imaging cytometry, we found a novel nuclear localization of TRIM13 and an expression pattern characterized by increased TRIM13 expression as cell cycle progressed. We confirmed the stabilization/activation relationship between TRIM13 and CCNA1 by restoring CCNA1 expression and confirming loss of the colony formation benefit in T13RINGdel AML cell lines. This study is the first to link CHAF1B, a cell cycle gene, with transcriptional repression of genes that promote cell cycle progression. We believe this relationship is critical for AML cells to support high levels of CHAF1B expression, thereby promoting an immature leukemic state throughout replication, without premature entry into multiple damaging rounds of proliferation.
Read full abstract