Category: Preclinical Development
Purpose: Synthetic aldehydes are a class of AEH with potential use in the treatment of sickle cell disease (SCD). Their primary anti-sickling mechanism of action is the “transiently covalent” modification of hemoglobin (Hb) by Schiff-base interaction (Hb-AEH adduct), preventing Hb polymerization and red blood cell sickling. 5-HMF (a naturally occurring furaldehyde) was studied until phase II, while voxelotor (a benzaldehyde derivative) is currently in phase III clinical testing. TD-7, TD-8, TD-9 and VZHE-39 are synthetic benzaldehydes, derived from (naturally occurring) vanillin. To allow for high-throughput screening, a novel, universal weak-cation exchange HPLC method to quantitate HbA-AEH adduct was used to screen AEH compounds and assess their structure-activity relationships in order to prioritize them for further development as potential SCD therapeutics.
Methods: Varying initial AEH concentrations were added to 0.2 or 0.1 mM HbA solutions ([HbA]0) on a 96-well plate and incubated at 37ºC. Serial aliquots were taken at different time points until steady-state (SS) was achieved (based on pilot studies) to characterize both SS binding properties and binding kinetics. The reaction was stopped by adding a mixture of NaBH3CN/NaBH4 (1:1v/v, 50 mM) to samples. An HPLC-UV/Vis assay method with a weak cation-exchange column (PolyCAT A) at 26°C with a linear mobile phase gradient and a detection wavelength of 410 nm was employed; it had been validated for the quantitation of HbA-AEH adduct and HbA concentrations (eLOQ for HbA of 0.006 mM). A sigmoidal Bmax-model, characterized by maximal AEH-HbA adduct formed/AEH bound (Bmax), dissociation equilibrium constant (KDss) and Hill coefficient/sigmoidicity factor (n), was fit to the AEH concentration-dependent SS profiles; a bimolecular kinetic binding model, characterized by association rate constant (kon) and dissociation rate constant (koff), was fit to the final HbA-AEH adduct concentration-time profiles. Nonlinear regression was used to estimate primary binding parameters (see above) as well as secondary binding parameters, i.e., KDkinetic=koff/kon, and equilibration half-life, i.e., t1/2eq = ln(2)/([HbA]0*kon+koff).
Results: All AEH show the expected (“transiently covalent”) time- and (saturable) concentration-dependent HbA binding. HbA-AEH adduct concentrations ultimately achieve SS under the experimental conditions. The SS and kinetic profiles are well fit by their respective binding models (results presented in Table 1): Both methods result in similar point estimates for KD. All AEH, other than vanillin (the parent compound for the benzaldehyde derivatives), show 100% HbA modification, i.e., Bmax/HbA0~1, at high concentrations. Consistently positive values for n in SS binding indicate allosteric effects - which is supported by X-ray crystallography studies, where two AEH molecules (e.g., TD-7 or VZHE-39) bind as Schiff-bases to the N-terminal Val1 of the α-globin chain. TD-8, TD-9 and 5-HMF equilibrate relatively slowly with HbA because of their small kon and koff. On the other hand, vanillin, TD-7 and VZHE-39 with large kon and/or koff values equilibrate rapidly with HbA. All AEH, except vanillin, show a strong correlation between their kon and koff values (r = 0.99, n=5). Vanillin - without a bulky side-chain available for HbA interactions - exhibits a small kon and relatively large koff, resulting in poor HbA affinity (KD). For benzaldehydes, the addition of the methoxy-pyridine side chain to the benzene ring, in general, leads to a (desirable) increase of kon. However, TD-7 and VZHE-039 also show, to a lesser extent, an (undesirable) increase of koff. Overall, KD values are improved from a low millimolar to a medium to high micromolar range. The three TD compounds are chemical isomers, but show large differences in their binding kinetics as a result of the location of the side chains on the benzene ring. VZHE-39, which is structurally similar to TD-7, except an ortho-positioned hydroxyl group instead of meta-positioned methoxy group relative to the aldehyde moiety, shows the fastest binding kinetics and highest binding affinity among the tested AEH; these results confirm findings from X-ray crystallography studies where additional interactions with the HbA molecule were observed.
Conclusion: Vanillin shows lower HbA binding affinity than 5-HMF (our lead compound), due to its faster dissociation from HbA. However, compared to parent vanillin, all benzaldehyde derivatives exhibit enhanced HbA binding affinity, primarily due to their faster association and, to a lesser extent, slower dissociation - indicating that the additional pyridinyl-methoxy side chains increase HbA binding affinity/interactions. In addition, the positions of the pyridinyl-methoxy group or/and methoxy group on the benzene ring determine KD (TD-7< TD-9< TD-8), mainly by enhancement of kon with little change in koff. Finally, a single replacement of the meta-positioned methoxy group (in TD-7) with an ortho-positioned-hydroxy group (in VZHE-039) results in the fastest and most potent HbA binder among all AEH tested, namely VZHE-039, our current top candidate for further SCD/ADME screening.