Acute lymphoblastic leukemia (ALL) in infants carries a poor prognosis and is characterized by cytogenetic rearrangements producing abnormal MLL fusion genes. Clinically effective targeting of the MLL fusion heterocomplex remains challenging, and therapeutic options remain limited. We have observed that the reduced isoform of HMGB1, a chromatin architectural protein that stabilizes DNA and facilitates transcription, is selectively over-expressed in the nuclei of infant MLL-ALL cells. In this study, we generated an HMGB1 siRNA knockdown in primary MLL-ALL cells from 3 infants to test our hypothesis that HMGB1-MLL interactions regulate pro-leukemic gene expression and represent a rational therapeutic target.

CD19-selected leukemic blasts were isolated from the cryopreserved bone marrow or peripheral blood specimens of 3 infants with cytogenetically confirmed MLL-AF4 rearrangements. HMGB1 knockdown was confirmed by comparing HMGB1 mRNA and protein expression, by qPCR and Western Blot, in cells transfected with HMGB1 vs. control sequence siRNA. First, determined whether HMGB1 knockdown affected expression of the MLL fusion gene itself, by comparing MLL-AF4 mRNA and protein levels 72 hours after siRNA transfection. HMGB1 knockdown produced a 2.8 (± 0.55)- fold decrease in MLL-AF4 mRNA expression by qPCR (p<0.05), with a corresponding decrease in MLL-AF4 fusion protein expression by Western Blot, in each of the 3 specimens.

Next, we determined whether HMGB1 binds functionally relevant regions of the MLL gene. We developed an electrophoretic mobility assay (EMSA) to compare the mobility of lysates from control vs. HMGB1 siRNA-treated infant MLL-ALL cells when mixed with biotinylated oligonucleotides spanning the transcriptionally active domains of MLL1. In each of 3 primary infant MLL-ALL cells, we detected a consistent gel-shift pattern on SDS-PAGE, in wild-type and control siRNA lysates, with oligonucleotides spanning exons 6-9- where many MLL-AF4 fusions occur. The gel-shift was completely abrogated in HMGB1 siRNA lysates. We then compared the expression of MLL target genes involved in leukemic transformation, by qPCR, in infant MLL-ALL cells treated with HMGB1 vs. control siRNA. We observed a significant (p<0.01) reduction in expression of MEIS1 (5.8 ± 2.2-fold decrease), HOXA7 (4.3 ± 0.4-fold decrease) and HOXA9 (3.7 ± 1.5-fold decrease) in infant MLL-ALL cells treated with HMGB1 vs. control siRNA. These data confirmed a role for HMGB1 in MLL gene/target gene regulation at the DNA level.

Finally, we considered whether HMGB1, as a scaffold protein, could interact directly with the MLL fusion heterocomplex at the protein level. We immunoprecipitated HMGB1 from the nuclear fraction of wild-type primary infant MLL-ALL cells (n=3 patients), then probed the pull-down for N-terminal MLL (MLLn), C-terminal MLL (MLLc), the MLLn-AF4 fusion, the MLLn-ENL fusion, and the MLL-associated histone 3 methyltransferase DOT1L. MLLn and MLLn-AF4 were strongly detected in all HMGB1 immunoprecipitates. Individual and sequential co-immunoprecipitation of HMGB1 with MLL-AF4 and DOT1L in revealed loss of known complex formation between MLL-AF4 and DOT1L following HMGB1 knockdown. This was accompanied by a 3.4 (± 0.9)-fold decrease in DOT1L mRNA expression (p<0.001) by qPCR and a complete loss of histone 3k79me2 protein expression by Western blot.

Taken together, these data suggest a central role for the fully reduced isoform of HMGB1, found in high abundance in infant ALL nuclei, in the formation of the MLL-AF4 transcription complex- including for the stable recruitment of DOT1L and H3K79me2, and in the regulation of MLL target genes such as HOXA9 and MEIS1. We are currently conducting chromatin immunoprecipitation and sequencing studies to identify methylation marks, particularly at H3K79me2, impacted by HMGB1 knockdown in infant ALL cells. We hope these studies will directly inform the development of small molecule inhibitors that specifically disrupt the binding sites and capacities of HMGB1 with MLL, which could synergize with the effects of methyltransferase inhibitors to more completely silence leukemic gene expression in infant ALL and improve the prognosis of this devastating disease.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution