Alarmin High Mobility Group Box 1 (HMGB1) Activates Alternative NFkB Signaling in Bone Marrow Macrophages in the Leukemia Microenvironment

High mobility group box 1 (HMGB1) is a damage-associated molecular pattern (DAMP) that signals through pattern recognition receptors on monocyte/macrophages to recruit them to sites of tissue injury. In models of sepsis and autoimmunity, HMGB1 can have both immune-activating and regulatory effects depending on the receptor system engaged and the downstream pathway activated. Although chemotherapy-stressed tumor cells are known to release HMGB1, little is known about the effects of tumor-derived HMGB1 on macrophages in the tumor microenvironment.

We have previously shown that primary human acute lymphoblastic leukemia cells release high levels of HMGB1 after treatment with anthracycline chemotherapy, and that leukemia-associated HMGB1 induces expression of the IL-1 converting enzyme (caspase-1), consistent with activation of the inflammasome. Paradoxically, inflammatory cytokine production (including IL-1beta) is reduced in this setting. In the present study, we test the hypothesis leukemia-associated HMGB1 activates alternative NFkB signaling through the repressor protein RelB which has been recently shown to silence inflammatory cytokine production via epigenetic effects in sepsis models.

To establish proof-of-principle, we first treated primary human bone marrow macrophages (BMM) and THP1 monocytes with 200 ng/ml recombinant HMGB1 (r-HMGB1) in the presence or absence of bacterial lipopolysaccharide (LPS). As predicted, LPS alone increased baseline expression of total NFkB, total p65/RelA and phosphorylated p65/RelA mRNA (20.6 + 7.9-fold by qPCR ddCT, p<0.0001) and protein (measured by Western blot, in BMM and THP1. In contrast, r-HMGB1 treatment produced a significant increase in the expression of RelB mRNA (2.5 + 0.7 fold increase p<0.05) and protein without a significant change in the expression of tNFkB or p65/RelA. Treatment with both r-HMGB1 and LPS maintained this pattern of increased RelB expression without increased expression of p65/RelA, suggesting HMGB1 activity as the dominant effect.  Increased RelB expression was unaffected by the addition of an antibody to TLR4, the primary receptor for LPS and a candidate receptor for HMGB1.

Next, we determined whether induction of RelB expression also occurs in response to leukemia-derived HMGB1. Immunomagnetically selected ALL cells from the bone marrow of 3 pediatric patients and NALM 6 ALL cells were treated with 20 nM doxorubicin, washed, and placed in culture for 24 hours prior to the addition of BMM in direct coculture. In our model, ALL cells typically release 100-450 ng/ml HMGB1 under these conditions. Again, significantly increased expression of RelB mRNA (6.7 + 2.9 fold, p<0.001) and protein was observed in BMM cultured with doxorubicin-treated ALL cells compared to ALL cells treated with media control, with a corresponding decrease in phospho-IkBa protein expression and unchanged total NFkB, p65/RelA and phosphorylated p65/RelA. Pre-treatment of ALL cells with quercetin, a commercially available flavonoid antioxidant shown to inhibit HMGB1 release, inhibited protein expression of RelB in BMM in a dose-dependent fashion.

These data indicate that tumor-derived HMGB1 activates alternative NFkB signaling and is capable of reprogramming the inflammatory response to LPS. Studies are ongoing to the determine downstream effects on macrophage antitumor function to optimize therapeutic reversal of HMGB1 effects and improve the potential for bone marrow macrophages to eradicate residual leukemia that remains after chemotherapy.