Non-covalent dual HDAC6/proteasome inhibitors with anti-multiple myeloma activity

Multiple myeloma (MM) remains an incurable hematological malignancy, even with the clinical success of proteasome inhibitors (PIs). The compensatory upregulation of HDAC6 following proteasome inhibition offers a strong rationale for dual targeting. However, a systematic comparison of different zinc-binding groups (ZBGs) within a non-covalent scaffold has been lacking. Here, we designed a series of non-covalent dual HDAC6/proteasome inhibitors based on a carfilzomib-derived scaffold, incorporating either a hydroxamic acid or an ortho-aminoanilide ZBG with linkers of varying lengths. Our SAR analysis revealed a clear division of labor: the R¹ position dominates proteasome inhibition (IC₅₀ = 2.6-8.1 nM, comparable to MG-132 at 10.8 nM), whereas HDAC6 activity is highly dependent on the choice of ZBG and cap group. For instance, the ortho-aminoanilide ZBG confers potent HDAC6 inhibition only when paired with a 2-chlorobenzyl cap (IC₅₀ = 29.8 nM); with a tetralin cap, it is essentially inactive (IC₅₀ > 5 μM). In contrast, the hydroxamic acid ZBG delivers moderate HDAC6 activity (130-280 nM) across both cap groups. Among the nine compounds, H-07 emerged as the most balanced dual inhibitor, with proteasome IC₅₀ = 4.2 nM and HDAC6 IC₅₀ = 29.8 nM. In RPMI-8226 cells, H-07 increased ac-α-tubulin and ac-histone H3 levels and induced PARP cleavage-findings consistent with engagement of both HDAC6 and class I HDACs, in addition to the expected effects from proteasome inhibition. Molecular dynamics simulations provided plausible binding modes for both targets. Despite this promising in vitro profile, a substantial loss of potency was observed in cell-based functional assays: the cellular proteasome IC₅₀ of H-07 was 6.768 μM, representing a > 1600-fold increase relative to its enzymatic IC₅₀ (4.2 nM). This marked shift is consistent with the limited membrane permeability frequently associated with polar peptidomimetic scaffolds, and it underscores a key limitation that must be addressed in future optimization efforts. Nevertheless, this work establishes a clear SAR framework and identifies key determinants for the rational optimization of this class of non-covalent dual inhibitors.