Modifying aroylhydrazone prochelators for hydrolytic stability and improved cytoprotection against oxidative stress

Abstract

BSIH ((E)-N’-(2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzylidene)isonicotinohydrazide) is a prodrug version of the metal chelator SIH ((E)-N’-(2-hydroxybenzylidene)isonicotinohydrazide) in which a boronate group prevents metal chelation until reaction with hydrogen peroxide releases SIH, which is then available for sequestering iron(III) and inhibiting iron-catalyzed oxidative damage. While BSIH has shown promise for conditionally targeting iron sequestration in cells under oxidative stress, the yield of SIH is limited by the fact that BSIH exists in cell culture media as an equilibrium mixture with its hydrolysis products isoniazid and 2-formylphenyl boronic acid. In the current study, several BSIH analogs were evaluated for their hydrolytic stability, reaction outcomes with H₂O₂, and prochelator-to-chelator conversion efficiency. Notably, the para-methoxy derivative (p-OMe)BSIH ((E)-N’-(5-methoxy-2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzylidene)isonicotinohydrazide) and the meta-, para- double substituted (MD)BSIH ((E)-N’-((6-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzo[d][1,3]dioxol-5-yl)methylene)isonicotinohydrazide) showed 1.3- and 1.9-fold improved hydrolytic stability compared to BSIH, respectively, leading to a 22 and 50 % increase in chelator released. Moreover, both prochelators were found to protect retinal pigment epithelial cells stressed with either H₂O₂ or paraquat insult.

Graphical Abstract

1. Introduction

Iron plays essential roles in the proper functioning of the body, but excessive or otherwise misregulated iron is also implicated in pathological progression of several diseases.^1–3 Chelation therapy with high-affinity iron chelating agents can be part of treatment plans for patients with systemic iron overload, for example thalassemia-major patients who require frequent blood transfusions.⁴ There are a number of diseases, however, that present signs of iron-promoted oxidative injury but do not involve systemic iron overload, for example neurodegenerative and cardiovascular disorders.^5–9 Several types of cancers also exhibit significantly perturbed aspects of iron uptake, utilization and regulation associated with cancer cell proliferation.^10–12 For these cases, conditional chelation strategies that direct iron chelators preferentially to sites of localized need without adversely perturbing normal metal homeostasis are desirable.^11–14 Our lab has pursued a prodrug strategy for chelating agents in which the resulting “prochelator” has negligible metal binding affinity until a specific stimulus produces a chelation site.¹⁴ Installation of a boronate masking group in place of a phenol of a multidentate chelating agent creates prochelators that are responsive to hydrogen peroxide, a reactive oxygen species that contributes to oxidative stress.^15–21 The conditional removal of the boronate mask by hydrogen peroxide thus conscripts iron sequestration specifically under oxidative conditions in which silencing redox-active iron can minimize further cell damage.

Our first prochelator BSIH ((E)-N’-(2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzylidene)isonicotinohydrazide) contains an aryl boronic ester in place of the phenol of the well-studied chelator salicylaldehyde isonicotinoyl hydrazone (SIH). The key steps involved in BSIH activation are shown in Scheme 1. BSIH is synthesized as the pinacol boronic ester, which equilibrates in water to give the boronic acid version, BASIH (step 1). Both BSIH and BASIH react with H₂O₂ in the second step to produce SIH, which is then available in step 3 to bind available metal ions, in the case of Fe³⁺ it forms a 2:1 ligand:metal complex.¹⁵ BSIH has been shown to minimize hydroxyl radical production in vitro by inhibiting iron-catalyzed Fenton chemistry,²² protect cells in culture from damage induced by hydrogen peroxide insult,²³ and be well tolerated by cells that are otherwise compromised by long-term exposure to conventional iron chelators.²⁴ Despite these favorable properties, the yield of SIH released from BSIH in cellular contexts is significantly lower than predicted, requiring higher concentrations of BSIH to protect cells against peroxide stress compared to SIH itself.²⁵

Scheme 1.

Key steps involved in BSIH hydrolytic degradation, peroxide activation, and Fe³⁺ chelation.

The low cellular BSIH-to-SIH conversion originates from the inherent equilibrium that exists in aqueous solutions between BSIH and its hydrolysis products isoniazid and (2-formylphenyl)boronic acid (FBA) (Scheme 1, step 4), which consequently reduces the amount of active SIH chelator released upon exposure to H₂O₂.²⁶ In spite of this equilibrium, BSIH itself maintains its integrity in media and plasma, with little degradation seen over 10 h incubation in plasma.²⁷ SIH, on the other hand, remains intact in buffer at pH 7.4 but degrades within hours in cell culture media and plasma due to amino acid-catalyzed hydrolysis reactions (Scheme 1 step 5; further reactions likely).^27,28

In an effort to minimize undesired hydrolysis, we previously developed a later-generation prochelator based on the SIH analogue, HAPI ((E)-N’-(1-(2- hydroxyphenyl)ethylidene)isonicotinohydrazide),²⁹ which demonstrates increased resistance to hydrolysis in plasma by replacing the salicylaldehyde component of SIH with the respective methyl ketone.¹⁸ Prochelator BHAPI harbors a self-immolative linker that undergoes 1,6-benzyl elimination upon H₂O₂ activation to release HAPI. While BHAPI demonstrates superior prochelator-to-chelator conversion rates and was shown to protect cardiomyoblast cells against oxidative injury induced by catecholamines, high-dose or long-time incubation of either BHAPI or HAPI was found to elicit cytotoxicity.^24,30 In fact, our original design, BSIH, was shown to have the most favorable therapeutic ratio of low inherent toxicity to cytoprotection against oxidative damage in comparison with several Fe chelators and prochelators.²⁴ As a result of the more favorable non-toxic yet cytoprotective outcomes for BSIH compared to BHAPI, we chose to optimize BSIH-like aroylhydrazones for improved activity.

The insights of BSIH hydrolysis shown in Scheme 1 provide a foundation for tuning hydrolytic stability and cytoprotective efficacy of BSIH analogs. In principle, new prochelators with improved stability can be created by tweaking the aroylhydrazone structure to shift the hydrolysis constant K (Scheme 1 step 4) to favor the intact prochelator, which in turn would increase the efficiency of prochelator-to-chelator conversion upon peroxide exposure and the efficacy of cytoprotection against oxidative stress. The ultimate goal of this study was to identify aroylhydrazone prochelators that maintain the favorably low inherent cytotoxicity of BSIH, yet show improved hydrolytic stability in aqueous solutions so as to maximize the amount of active chelator available for cytoprotection upon cellular oxidative stress.

In the current study, several boronate-masked aroylhydrazone prochelators were evaluated for their peroxide-activated prochelator-to-chelator conversion efficiency, hydrolytic stability, and cytoprotective efficacy in oxidatively stressed retinal pigment epithelial cells.

2. Results

2.1. Design and Synthesis

BSIH-like derivatives are readily synthesized via Schiff base condensation reactions between hydrazide and borylated aryl aldehyde building blocks in acidic conditions, as shown in Chart 1. This straightforward procedure provides access to a wide range of structural variation. The compounds for the current study were selected based on knowledge gained from prior structure-function studies. As we previously reported, replacing the pyridine ring of the hydrazide building block with an aryl ring diminished Fe³⁺ binding affinity of the resulting chelator, which in turn lowered its efficacy to inhibit iron-promoted hydroxyl radical formation.¹⁵ The current compounds therefore retain a pyridyl ring to maintain high Fe³⁺ affinity, a consequence of electronic effects, as the pyridyl N is not directly part of the metal coordination sphere. Substitutions on the aryl aldehyde building block, on the other hand, had little effect on the Fe³⁺ affinity, but did impact the rate of H₂O₂-dependent conversion from prochelator to chelator, with electron-donating substituents favoring faster reactions than electron-withdrawing substituents.¹⁵

Chart 1.

Synthetic scheme and structures of prochelators used in this study. Structures of the corresponding chelators replace the boronic ester or boronic acid with an OH functional group; names of chelators mirror those of the prochelators, just without the “B”.

Based on these prior observations, which were done in methanol or water-methanol mixtures, we selected the aroylhydrazone prochelator derivatives shown in Chart 1 for further study in aqueous solution and to test their efficacy at minimizing oxidative stress in cell culture. Compared to the original BSIH structure, (p-OMe)BSIH and (m-OMe)BSNH contain an electron-donating methoxy substituent installed at either the para- or meta position to the boronic ester, respectively. Nicotinic hydrazide was used for (m-OMe)BSNH in place of the isoniazid building block for the other compounds because the chelator version (m-OMe)SIH was not soluble above 50 µM in aqueous buffers, whereas (m-OMe)BSNH and (m-OMe)SNH can be prepared at concentrations as high as 300 µM. The methylenedioxy derivative (MD)BSIH was also prepared in order to evaluate the effect of double substitution at both para-and meta- positions.

Inspired by the precedence of hydrolytically stable chelators and prochelators based on thiosemicarbazone structures,^31,32 BSMTH and BSPTH embed thiosemicarbazide components to incorporate a sulfur donor into the metal coordination sphere. Finally, the control compound (m-OMe)BIIH repositions the boronic ester to test the hypothesis that its placement contributes to the hydrolytic susceptibility of the hydrazone framework and so that reaction with H₂O₂ will produce a phenol that is not positioned to provide a metal chelation site.

Based on the previously reported crystal structure of BSIH, the prochelators prepared here are predicted to prefer the E configuration about the C=N double bond, with the boronic ester positioned anti to the imine N.²² As was the case with BSIH, however, additional small peaks in the ¹H NMR spectra of clean samples reveal the presence of other isomers (Supp Info). The ¹³C NMR spectra for these prochelators appear to be missing a peak, an observation consistent with the known line-broadening effects of ¹¹B on the aromatic carbon to which it is bonded.³³ An additional subtlety of these compounds is the potential for boronic acids at the ortho position of aroylhydrazones to dehydrate and form cyclic diazaborine monomers and anhydrated dimers.³⁴ While the prochelators reported here were prepared as pinacol esters to minimize this transformation, the pinacol diol is labile and can be lost under high vacuum or upon hydrolysis. In all cases, the ¹H NMR spectra obtained in DMSO-d₆ confirm that the compounds as prepared exist in solution as the open aroylhydrazone form, as evidenced by the presence of a 12-H singlet at 1.3 ppm corresponding to a pinacol coordinated to boron, and a D₂O-exchangeable NH peak ~12 ppm (Supp Info). Given the propensity of these compounds to transform and interconvert by the addition or removal of H₂O, it is not surprising that some of the diazaborine forms can be observed during chromatographic analysis (see Supp Info).

2.2. Conditional Iron Chelation

A prerequisite for a prochelator is the correct transformation into a form that binds its target metal ion under a specific triggering condition. Here, a fluorometric competitive binding assay was utilized to assess the efficiency of prochelator-to-chelator conversions of the BSIH derivatives triggered by the presence of H₂O₂. Calcein is a fluorescent probe with a multi-carboxylate metal-binding site that binds Fe(III) with an affinity and 1:1 stoichiometry that corresponds to a pFe of 20.3.³⁵ The term pFe (−log[Fe_aq³⁺]) is a quantitative expression that facilitates direct comparison of chelating agents of different binding stoichiometries by calculating the amount of uncomplexed iron under defined conditions, typically pH 7.4 with total ligand and iron concentrations of 10 µM and 1 µM, respectively.³⁶ As shown in Figure 1, the fluorescence emission of calcein was quenched by ~70% upon complexation with 1 equivalent of Fe³⁺ (basal level, gray dashed line). Addition of 2 equivalents of chelator SIH to the calcein-Fe³⁺ solution resulted in an almost full recovery of the original fluorescence, indicating that SIH readily out-competes calcein for Fe³⁺ chelation, as predicted based on pFe values of 23 and higher for this ligand class.^15,37,38

Figure 1.

Calcein fluorescence assay with chelators and H₂O₂-activated prochelators. Buffered solutions of prochelators were pre-incubated overnight with excess H₂O₂ then diluted to 4 μM and equilibrated with 2 μM calcein-Fe³⁺ for 2 h prior to measurement of calcein fluorescence emission. Chelation efficiency was expressed as a percentage of the calcein emission of the free calcein control. The grey dashed line indicates the residual fluorescence response of 2 μM calcein quenched by 1 equiv of Fe³⁺. The red dashed line indicates the ~50% calcein emission restored by the reaction mixture of BSIH and H₂O₂. Data are presented as mean ± SD, n = 3; statistical significance (t-test), * p < 0.05, compared to the BSIH group. # p < 0.05, compared to the SIH group.

All borylated compounds in Chart 1 contain boronate masks blocking potential metal chelation sites such that they demonstrate negligible affinity for Fe³⁺ with calcein-Fe emission remaining at the basal level (Figure S1). In contrast, reaction of the prochelators with H₂O₂ should in principle convert them into active chelators capable of chelating Fe³⁺ and restoring calcein fluorescence. As indicated by the red dashed line in Figure 1, BSIH reacted with H₂O₂ to incur a ~50% recovery in calcein fluorescence. This result reinforces the conclusion that prochelator BSIH in aqueous solution is only partially converted to its corresponding chelator SIH for concomitant Fe³⁺ chelation.^25–27 Among the new prochelators, (p-OMe)BSIH and (MD)BSIH restored calcein emission to ~60% and 75% chelation efficiency, respectively, suggesting enhanced content of active chelator. Peroxide activation of (m-Cl)BASIH, BSPTH, and (m-OMe)BSNH, however, resulted in approximately the same or lower levels of calcein fluorescence compared to that of BSIH. Incubation of the control compound, (m-OMe)BIIH with H₂O₂ showed no change in fluorescence compared to the calcein-Fe³⁺ solution, confirming that the reaction products fail to extract Fe³⁺, as anticipated because the resulting phenol product is ill-positioned to create a multidentate metal binding pocket. The calcein emission of the methyl thiosemicarbazide-based prochelator BSMTH after reaction with H₂O₂ also remained at the basal level of the calcein-Fe³⁺.

The conclusions from the calcein assay with peroxide-reacted prochelators were reinforced by results of the calcein assay with their corresponding chelator counterparts. As shown in Figure 1, addition of SIH, (p-OMe)SIH and (m-OMe)SNH lead to an increase in calcein fluorescence by more than 90%, indicating that these chelators readily compete with calcein for Fe³⁺ binding. Meanwhile, the S-containing chelators SMTH and SPTH failed to fully restore calcein emission; indeed, SMTH had no effect on the quenched signal of calcein-Fe³⁺ solution. The consistent results of weak Fe³⁺ affinity for both thiolated chelators and activated prochelators imply that replacement of the aroylhydrazone chelation pocket with thiosemicarbazone scaffold significantly reduces the binding affinity for Fe³⁺.

2.3. Hydrolytic Stability

Recognizing (p-OMe)BSIH and (MD)BSIH as the most promising derivatives with respect to prochelator-to-chelator conversion, these compounds were selected for further assessment of hydrolytic stability in aqueous solutions. The control compound (m-OMe)BIIH was included, based on our hypothesis that the position of the boronic functionality might influence the hydrolytic stability of the hydrazone bond.

According to our previous observations for BSIH hydrolysis, intact aroylhydrazone prochelators are anticipated to establish equilibria in aqueous solutions with corresponding degradation fragments: hydrazides and boronate-masked salicylaldehydes. To quantify the equilibrium ratios of intact prochelator vs. degradation components, aqueous samples of these selected prochelators were analyzed by NMR. Figure 2a shows the aromatic region of the ¹H NMR spectra of (p-OMe)BSIH in DMSO-d₆, with the hydrazone proton from the intact prochelator clearly visible at 9.01 ppm (peak c). Upon dilution and equilibration in D₂O for 1 h, however, the hydrazone peak disappeared, with concomitant appearance of two extra peaks corresponding to the aromatic protons of isoniazid (indicated by red arrows), Figure 2b. Integration of peak a and a’ in Figure 2b gave 85% of intact (p-OMe)BSIH vs. 15% of INH at equilibrium in a 500-μM sample of (p-OMe)BSIH. Comparable solutions of BSIH resulted in 80% intact chelator.²⁶ A similar set of spectra for (MD)BSIH are shown in Supp Info Figure S2. Because of the lower aqueous solubility of (MD)BSIH, these samples were prepared at 200 μM. Integration indicates a composition of 80% intact (MD)BSIH vs. 20% hydrolysis fragments in the equilibrated solution. These numbers are not directly comparable with BSIH though, due to the different initial concentrations of the prochelators.

Figure 2.

Hydrolysis of (p-OMe)BSIH observed by ¹H NMR spectroscopy. The aromatic regions of ¹H NMR spectra of (a) (p-OMe)BSIH (10 mM in DMSO-d₆) (b) (p-OMe)BSIH (500 μM in D₂O, with 5% DMSO-d₆ after 1-h equilibration) and (c) INH (500 μM in D₂O, with 5% DMSO-d₆). The arrows indicate the peaks corresponding to the aromatic protons of isoniazid (red).

Unlike the observations for the BSIH-based prochelators, the hydrazone proton signal of the control compound (m-OMe)BIIH detected in DMSO-d₆ (δc = 8.38, Figure 3a) persists upon dilution in D₂O (δc = 8.17, Figure 3b). Meanwhile, no additional peaks corresponding to potential hydrolysis products were observed, substantiating the superior hydrolytic resistance of (m-OMe)BIIH.

Figure 3.

Hydrolytic stability of (m-OMe)BIIH analyzed by ¹H NMR spectroscopy. The aromatic regions of ¹H NMR spectra of (a) (m-OMe)BIIH (10 mM in DMSO-d₆) and (b) (m-OMe)BIIH (500 μM in D₂O, with 5% DMSO-d₆).

The hydrolysis equilibria of (p-OMe)BSIH and (MD)BSIH in PBS, pH 7.4 were further monitored by UV-Vis spectrophotometry. The spectra in Figure 4a and c show the time-dependent changes of intact prochelators (p-OMe)BSIH and (MD)BSIH upon dilution in PBS from corresponding stock solutions prepared in DMSO. Similar alterations of spectral features were also observed in the reaction mixtures of the respective hydrolysis fragments (Figure 4 b and d). These results reinforce the conclusion that these BSIH analogs also undergo hydrolysis equilibria in aqueous solutions with their corresponding hydrazide and aldehyde degradation products.

Figure 4.

The hydrolysis equilibria of 100 μM (a) (p-OMe)BSIH and (c) (MD)BSIH observed by UV-Vis spectrophotometry; spectra of reaction mixtures of 100 μM INH with100 μM of (b) (p-OMe)FBA or (d) (MD)FBA, pH 7.4 over the course of 30 min.

The hydrolysis equilibrium constants for these two prochelators were calculated from the spectra of equilibrated reaction mixtures at various concentrations, which gave K values of (1.1 ± 0.1) × 10⁻⁵ M for (p-OMe)BSIH and (8.0 ± 0.6) × 10⁻⁶ M for (MD)BSIH. These numbers are smaller compared to that of BSIH [(1.5 ± 0.1) × 10⁻⁵ M], indicating that the hydrolysis reactions of (p-OMe)BSIH and (MD)BSIH slightly favor the intact prochelators.

In contrast, the spectrum of the (m-OMe)BIIH control compound remains unchanged over time in PBS buffer (Figure S3). This observation in combination with the signal integrity in ¹H NMR speciation represents a unique feature of (m-OMe)BIIH that distinguishes it from other BSIH derivatives and implies that installation of the boronic ester mask at a different position prevents the hydrolysis equilibrium in aqueous buffered solutions observed in BSIH-like prochelators. However, this modification also disables the metal-binding capacity of the activation product, as the critical phenol functionality after deboronation is inappropriately positioned for metal chelation.

Given that (p-OMe)BSIH and (MD)BSIH exhibited improved hydrolytic stability, we speculated that the higher content of intact prochelators in aqueous solutions would allow for enhanced release of corresponding active chelators upon peroxide activation. Figure 5a shows the ¹H NMR spectra of a 500-μM sample of (p-OMe)BSIH in D₂O after reaction with excess H₂O₂. The intact chelator (p-OMe)SIH as well as the activation byproducts INH (indicated in red arrows) and (p-OMe)salicylaldehyde (indicated in green arrows) were clearly identified. Integration of the aromatic peak a derived from (p-OMe)SIH and peak a’ from INH indicates ~67% of intact chelator vs. ~33% of byproduct formation, which is significantly improved compared to the 51% of SIH generated from peroxide activation of BSIH at the same concentration.²⁶ Likewise, Figure 5d shows the spectrum of the reaction mixture of (MD)BSIH (200 μM) with H₂O₂ in D₂O, with byproduct INH indicated in red arrows. Integration of a” and a’ gives ~45% formation of intact chelator under this condition. Since the prochelator solution of (MD)BSIH was prepared at a lower concentration, the content of chelator formation was lower in comparison with BSIH. The reduced equilibrium constant K obtained from data in Figure 4 and the highest prochelator-to-chelator efficiency shown in Figure 1, however, both identify (MD)BSIH as the most promising analog regarding improved hydrolytic stability and maximal chelator release.

Figure 5.

The reactions of prochelators (p-OMe)BSIH and (MD)BSIH with H₂O₂ observed by ¹H NMR. The aromatic regions of ¹H NMR spectra of (a) the reaction mixture of (p-OMe)BSIH (500 μM in D₂O, with 5% DMSO-d₆) with H₂O₂ (25 mM) for 12 h (b) (p-OMe)SIH (500 μM in D₂O, with 5% DMSO-d₆) (c) INH (500 μM in D₂O, with 5% DMSO-d₆) (d) the reaction mixture of (MD)BSIH (200 μM in D₂O, with 2% DMSO-d₆) with H₂O₂ (10 mM) for 12 h. See main text for description of arrows and peak labels.

2.4. Cytotoxicity and Cytoprotection against Oxidative Stress

Having shown that selected BSIH derivatives (p-OMe)BSIH and (MD)BSIH fulfill desirable prochelator attributes, we further explored the structure-activity relationship of these aroylhydrazone derivatives and investigated their ability to protect cells subjected to oxidative damage. The cell culture experiments were conducted in human retinal pigment epithelial ARPE-19 cells as they suffer from iron-induced oxidative stress during age-related macular degeneration.^39,40

Initial cytotoxicity studies revealed that all the new BSIH-based prochelators were non-toxic at concentrations up to 100 μM for 24-h incubation (data not shown). BSIH and (MD)BSIH remained non-toxic after 72 h incubation, however, cells were less tolerant of elevated doses of (p-OMe)BSIH when incubation was extended over 72 h, with a 20% reduction of cell viability at the concentration of 100 μM (Figure 6).

Figure 6.

Inherent cytotoxicity of selected aroylhydrazone prochelators after 72-h incubation in ARPE-19 retinal pigment epithelial cells. Error bars represent standard deviations from triplicate runs (n = 3).

The cytotoxicity of our self-immolative prochelator BHAPI was also assessed for comparison, as it was previously found to elicit toxicity in H9c2 cardiomyoblasts, with a TC₅₀ value of 37.1 ± 7.7 μM after 72 h.²⁴ A relatively moderate toxic effect was observed in ARPE-19 cells exposed to BHAPI, with a 25% viability reduction at the highest concentration tested (100 μM).

Figure 7 shows the efficacy of the new aroylhydrazone prochelators to protect cells from peroxide damage. To avoid direct reaction of prochelators with H₂O₂ in the treatment medium, cells were first preincubated with the test compound prior to exposure to a toxic 200-μM bolus of H₂O₂. After 19 h, cells were found to be significantly protected from death by most of the prochelators tested, with the exceptions of (m-OMe)BIIH and BSMTH. The two prochelators with improved hydrolytic stability, (p-OMe)BSIH and (MD)BSIH, demonstrated the most potent cytoprotective activity, being effective at 25 μM. Similar dose-dependent protection was also observed in peroxide-stressed cells exposed to BHAPI.¹⁸ Meanwhile, the thiolated prochelator BSPTH demonstrated comparable protective potency to BSIH, being effective at concentrations between 25 and 50 μM. In contrast, the lack of cytoprotection observed in the cases of BSMTH and the (m-OMe)BIIH control is likely due to the non-chelating property of their corresponding activation products.

Figure 7.

Cytoprotective effects of aroylhydrazone prochelators from oxidative damage induced by H₂O₂ in ARPE-19 cells. ARPE-19 cells were pre-incubated with various concentrations of aroylhydrazone prochelators for 5 h prior to exposure to H₂O₂. Cell viability was measured 19 h after peroxide treatment and expressed as percentage of the untreated control group (100%). Error bars represent standard deviations from triplicate runs (n = 3).

To further explore the role of iron chelation in the mode of protection for aroylhydrazone prochelators, we also tested the cytoprotective effects of corresponding chelators against oxidative stress. As shown in Figure S4, the thiosemicarbazone SMTH that lacks affinity for Fe³⁺ had no protective effect on H₂O₂-stressed ARPE-19 cells, while other chelators with adequate metal binding affinity protected cells from the oxidative damage induced by H₂O₂, being effective at 5 μM.

In addition to stressing cells with exogenous H₂O₂ added in bolus form, cells were also subjected to paraquat, a viologen herbicide that is a known environmental risk factor for Parkinson’s disease. Upon reduction in mitochondria, paraquat elicits overproduction of superoxide with concomitant release of H₂O₂ and liberation of Fe from Fe-S clusters.^41,42 The cytoprotective effect of aroylhydrazone prochelators was tested in ARPE-19 cells stressed with 10 mM paraquat for 48 h, the timepoint at which cell death by paraquat exposure was observed under these conditions. BHAPI was previously shown to be protective from paraquat-induced cell death, with a 40% viability restoration at 100 μM.¹⁸ Figure 8 shows that the cytoprotection of (MD)BSIH tracked very closely with BHAPI to give a dose-dependent recovery of cell viability. (p-OMe)BSIH also displayed adequate protective effects on paraquat-stressed cells, with 100 μM prochelators capable of restoring viability to ~30% of the untreated cells. BSIH and other prochelators, however, failed to protect cells from paraquat damage. The observation that the most stable prochelators are the only ones that protect cells against paraquat exposure highlights the importance of prolonged prochelator stability for maximal cellular protection.

Figure 8.

Cytoprotective effects of aroylhydrazone prochelators against oxidative damage induced by 10 mM paraquat in ARPE-19 cells. Cell viability was measured after 48 h and expressed as percentage of the untreated control group (100%). Error bars represent standard deviations from triplicate runs (n = 3).

To determine whether the lack of cytoprotection by BSIH, BSMTH and BSPTH was due to insufficient prochelator-to-chelator conversion or inability of the corresponding chelator to protect cells against paraquat damage, the cytoprotective effects of corresponding chelators were assessed in paraquat-stressed ARPE-19 cells. As shown in Figure S5, SIH, (p-OMe)SIH, HAPI and SPTH all demonstrated ~40% cytoprotection against paraquat-induced cell death, whereas the non-chelating SMTH failed to protect cells. Combined, these results substantiate the importance of iron chelation in the protective activity against this model of oxidative stress.

3. Discussion

Structural modification to the hydrazide and aryl aldehyde building blocks of the parent BSIH compound tunes its chemical properties and biological activity. Changing the electronic nature of the aldehyde-containing ring has substantive effects on the hydrolytic stability of the aroylhydrazone prochelators. The electron-donating p-OMe derivative is more stable with a higher content of intact prochelator present in aqueous solutions compared to the parent compound. The m- and p- double substitution with electron-donating methylenedioxy functionality affords prochelator (MD)BSIH with even greater hydrolytic stability and a hydrolysis equilibrium constant K ~2-fold lower than the value of parent BSIH. We reason that the electron-donating functionalities installed on the aldehyde-containing ring increase the electron density of the hydrazone carbon, reducing the susceptibility of C=N to the nucleophilic attack by water. Consistent with this reasoning, the prochelator with m-Cl substitution that is electron withdrawing with respect to the hydrazone carbon was less effective in the peroxide-triggered calcein assay, regardless of its chelator’s similar Fe affinity.¹⁵ This result suggests the electron withdrawing substituent aggravates hydrazone hydrolysis.

Additionally, modifications on the aldehyde-containing ring were previously found to significantly influence the kinetics of prochelator-to-chelator conversion activated by H₂O₂.¹⁵ In moving from the electron-withdrawing m-Cl to the electron-donating p-OMe compound, a nearly 5-fold increase in the H₂O₂-dependent conversion rates was observed. The accelerated kinetics of peroxide activation was further shown to associate with increased inhibition of metal-promoted ROS formation in vitro.¹⁵ The current study shows that improved hydrolytic stability is another benefit of the electron-donating derivatives (p-OMe)BSIH and (MD)BSIH, which ultimately leads to higher yields of active chelator upon peroxide activation and enhanced cytoprotective efficacy against cellular oxidative damage.

The exception to these structure-activity relationships is the control compounds, (m-OMe)BIIH. Moving the boronic ester functionality to the meta-position of the aldehyde/hydrazone group yields a hydrolytically stable aroylhydrazone framework. Despite the desired structural integrity of this compound, selective deborylation initiated by H₂O₂ generates a product with a hobbled metal chelation site. Therefore, we surmise that the aqueous susceptibility of other BSIH-derivatized prochelators likely derives from the electron-deficient boronic ester functionality installed at the ortho-position to the hydrazone bond. While this ortho-borylation is necessary for peroxide-triggered metal binding, the electron-withdrawing effect of boron reduces the electron density of the adjacent hydrazone carbon sufficiently to make it susceptible to nucleophilic attack by water.

Replacing the aroyl oxygen donor with a sulfur donor to create an O,N,S coordination sphere was unfavorable in our system, as it significantly decreased the Fe(III) affinity of the SPTH and SMTH chelators compared to SIH, as evidenced by the lack of competition with calcein under the conditions tested. While BSPTH retained modest cellular protection against peroxide exposure, BSMTH was ineffective. This difference in activity likely arises from electronic differences between the phenyl vs. methyl substituents of BSPTH vs BSMTH, something that has been observed in disulfide-bridged prochelators built on a similar scaffold.^31,43 Interestingly, the disulfide prochelators introduced by Tomat and coworkers are reductively-activated and create S,N,S or S,N,O iron chelators with interesting antiproliferative activity against cancer cells.^12,43 The boronate prochelators discussed here, on the other hand, are oxidatively activated and create O,N,O coordination spheres. The combination of these two approaches reveals the opportunity available to tune prochelators for desirable biological applications by adjusting their coordination sphere, activation mechanism and stability.

4. Conclusions

The structure-activity relationships of the new aroylhydrazone derivatives studied here reveal that the potency of cytoprotection against oxidative stress correlates with prochelator-to-chelator conversion efficiency as well as the Fe-chelating capacity of the resulting chelators. In particular, the electron-donating substituted prochelators (p-OMe)BSIH and (MD)BSIH with improved hydrolytic stability provided the best cytoprotection to cells stressed with either H₂O₂ or paraquat. The overall trends in cytoprotection observed for this set of aroylhydrazone compounds suggest that the major source of their protective effects derives from preferential Fe chelation in response to oxidative stress. These advances are important steps in developing this promising class of conditionally active iron chelators to potentially mitigate iron-associated oxidative damage in vivo.

5. Experimental Section

5.1. Materials and Instrumentation

BSIH ((E)-N’-(2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2- yl)benzylidene)isonicotinohydrazide),²² (m-Cl)BASIH) (E)-(5- chloro-2-((2- isonicotinoylhydrazineylidene)methyl)phenyl)boronic acid,¹⁵ and SIH ((E)-N’-(2-hydroxybenzylidene)isonicotinohydrazide)⁴⁴ were prepared as reported in the literature. The starting materials isoniazid and (2-formylphenyl)boronic acid pinacol ester were purchased in reagent grade from Thermo Fisher Scientific Inc and used without further purification; (5-chloro-2- formylphenyl)boronic acid and (2-formy-4-methoxy-lphenyl)boronic acid were purchased from Matrix Scientific; (4,5-dimethoxymethylene-2-formylphenyl)boronic acid and (2-formyl-5-methoxyphenyl)boronic acid were purchased from Frontier Scientific. Other reagents were purchased from Sigma-Aldrich. All solvents were reagent grade and all aqueous solutions were prepared from nanopure water. Compound stock solutions (10 mM) were prepared in DMSO and stored at −20 °C. ¹H NMR and ¹³C NMR spectroscopy was performed on a Varian Inova 400 or 500 MHz spectrometer. High-resolution mass spectra (HRMS) were recorded on an Agilent G6224 LCMS-TOF system. Purity of all final compounds was determined by high performance liquid chromatography (HPLC) to be 95% or higher. The instrument was an Agilent 1100 Series LCMS apparatus with a Supelco Ascentis C18 (50 × 1.0 mm) column. UV-Vis spectra were recorded on a Cary 50 UV-Vis spectrophotometer. Absorbance or fluorescence readings were measured using clear, flat-bottom 96-well plates and a Perkin Elmer Victor 3V multi-label counter maintained at 25 °C.

5.2. Synthesis

5.2.1. (p-OMe) BSIH

(E)-N’-(5-methoxy-2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzylidene)isonicotinohydrazide)

A portion of 4-methoxy-2-formylphenylboronic acid (1mmol, 0.180 g) was added to a nearly saturated solution of isoniazid (1 mmol, 0.137 g) in 0.1 M sodium acetate buffer, pH 4.5. The reaction mixture was stirred over an oil bath at 100 °C for 4 min. The white insoluble product was collected via vacuum filtration, washed with diethyl ether, and dried in vacuo to give an off- white powder corresponding to (E)-(2-((2-isonicotinoylhydrazineylidene)methyl)-4-methoxyphenyl)boronic acid ((p-OMe)BASIH) in 88% yield. Portions of (p-OMe)BASIH (0.5 mmol, 0.150 g) and pinacol (0.5 mmol, 0.059 g) were then dissolved in DMF and added to a round bottom flask. The reaction stirred for 21 h at rt then dried via rotary evaporation to give an off-white powder in 82% yield. ¹H NMR (DMSO-d₆, 500 MHz) δ (ppm): 12.24 (s, 1H), 9.01 (s, 1H), 8.79 (d, J = 5.6 Hz, 2H), 7.82 (d, J = 5.6 Hz, 2H), 7.71 (d, J = 8.4 Hz, 1H), 7.56 (d, J = 2.6 Hz, 1H), 7.04 (dd, J = 8.3, 2.6 Hz, 2H), 3.84 (s, 3H), 1.32 (s, 12H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 162.24, 161.50, 150.28, 149.20, 141.54, 140.76, 137.82, 121.78, 115.80, 109.55, 83.72, 55.21, 24.61. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₂₀H₂₄BN₃O₄] + H)⁺, 382.1933, found 382.1933.

5.2.2. (MD) BSIH

(E)-N’-((6-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzo[d][1,3]dioxol-5-yl)methylene)isonicotinohydrazide

A portion of 2-formyl-4,5-methylenedioxyphenylboronic acid (1 mmol, 0.194 g) and pinacol (1 mmol, 0.118 g) were dissolved in DMF and added to a round bottom flask. The reaction stirred for 21 h at rt then dried by rotary evaporation to give an off-white powder corresponding to 6-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzo[d][1,3]dioxole-5-carbaldehyde (MD-FBA_pin) in 80% yield. A portion of isoniazid (0.5 mmol, 0.068g) was dissolved in 1 mL methanol and added to a solution of MD-FBA_pin (0.5 mmol, 0.138 g) dissolved in 2 mL of methanol. The reaction was stirred for 2 h over an oil bath at 100 °C and cooled over ice. The liquid phase was then collected via vacuum filtration and dried by rotary evaporation. A yellow powder was isolated in 45% yield. ¹H NMR (DMSO-d₆, 500 MHz) δ (ppm): 12.15 (s, 1H), 8.95 (s, 1H), 8.79 (d, J = 5.2 Hz, 2H), 7.80 (d, J = 5.2 Hz, 2H), 7.52 (s, 1H), 7.14 (s, 1H), 6.12 (s, 2H), 1.33 (s, 12 H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 162.07, 150.24, 150.19, 148.83, 148.61, 140.83, 135.40, 121.73, 113.76, 104.71, 101.75, 83.97, 24.54. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₂₀H₂₂BN₃O₅] + H)⁺, 396.1725, found 396.1720.

5.2.3. (m- OMe) BSNH

(E)-N’-(4-methoxy-2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzylidene)nicotinohydrazide

A similar procedure to the one described above for (MD)BSIH was followed, but replacing 2-formyl-4,5- methylenedioxyphenylboronic acid with 2-formyl-5-methoxyphenylboronic acid. An off-white powder was isolated in 9% yield. ¹H NMR (DMSO-d₆, 400 MHz) δ (ppm): 12.02 (s, 1H), 9.04 (d, J = 1.9 Hz, 1H), 8.87 (s, 1H), 8.76 (dd, J = 4.7, 1.5 Hz, 1H), 8.23 (dt, J = 8.0, 1.9 Hz, 1H), 8.00 (d, J = 8.7 Hz, 1H), 7.57 (dd, J = 7.8, 4.7 Hz, 1H), 7.19 (d, J = 2.7 Hz, 1H), 7.15 (dd, J = 8.7, 2.8 Hz, 1H), 3.82 (s, 3H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 161.96, 159.84, 152.11, 148.74, 135.61, 131.98, 129.60, 127.58, 123.59, 119.52, 117.32, 84.09, 55.29, 24.61. HR- MS (ESI) (m/z): [M + H]⁺ calcd for ([C₂₀H₂₄BN₃O₄] + H)⁺, 382.1933, found 382.1932.

5.2.4. BSMTH

(E)-N-methyl-2-(2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzylidene)hydrazine-1-carbothioamide

One equivalent of (2-formylphenyl)boronic acid pinacol ester (1 mmol, 0.232 g) was added to a 2-mL solution of 4-methyl-3-thiosemicarbazide (1 mmol, 0.105 g) in ethanol and heated to 60 °C. After stirring for 10 min, 1 mL of chilled diethyl ether was added and the reaction was chilled over ice. A white precipitate was collected via vacuum filtration, washed with diethyl ether and dried in vacuo to give a white powder in 81% yield. ¹H NMR (DMSO-d₆, 400 MHz) δ (ppm): 11.74 (s, 1H), 8.68 (s, 1H), 8.48 (d, J = 4.5 Hz, 1H), 8.29 (d, J = 7.9 Hz, 1H), 7.69 (d, J = 7.3 Hz, 1H), 7.50 (t, J = 7.5 Hz, 1H), 7.38 (t, J = 7.3 Hz, 1H), 3.02 (s, 3H), 1.33 (s, 12H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 177.81, 142.47, 139.69, 135.46, 130.90, 128.60, 125.23, 83.94, 30.82, 24.54. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₁₅H ₂₂BN₃O₂S] + H)⁺, 320.1599, found 320.1598.

5.2.5. BSPTH

(E)-N-phenyl-2-(2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzylidene)hydrazine-1-carbothioamide

A similar procedure to the one described above for BSMTH was followed, but replacing 4-methyl-3-thiosemicarbazide with 4-phenylthiosemicarbazide. A white powder was isolated in 75% yield. ¹H NMR (DMSO-d₆, 500 MHz) δ (ppm): 12.10 (s, 1H), 10.08 (s, 1H), 8.79 (s, 1H), 8.42 (d, J = 7.9 Hz, 1H), 7.71 (d, J = 7.4 Hz, 1H), 7.57 (d, J = 7.9 Hz, 2H), 7.51 (t, J = 7.6 Hz, 1H), 7.41 (t, J = 7.6 Hz, 1H), 7.37 (t, J = 7.6 Hz, 2H), 7.20 (t, J = 7.3 Hz, 1H), 1.35 (s, 12 H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 176.09, 143.69, 139.38, 139.08, 135.44, 130.95, 128.90, 128.01, 125.85, 125.25, 83.98, 24.56. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₂₀H₂₄BN₃O₂S] + H)⁺, 382.1755, found 382.1756.

5.2.6. (m-OMe) BIIH

(E)-N’-(4-methoxy-3-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzylidene)isonicotinohydrazide

A similar procedure to the one described above for (p-OMe)BSIH was followed, but replacing 4-methoxy-2- formylphenylboronic acid with 5-formyl-2-methoxyphenylboronic acid. A white powder was isolated in 68% yield. ¹H NMR (DMSO-d₆, 500 MHz) δ (ppm): 11.95 (s, 1H), 8.78 (d, J = 5.5 Hz, 2H), 8.39 (s, 1H), 8.05 (s, 1H), 7.81 (d, J = 5.6 Hz, 2H), 7.74 (d, J = 8.7 Hz, 1H), 7.07 (d, J = 8.6 Hz, 2H), 3.81 (s, 3H), 1.30 (s, 12H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 165.54, 161.46, 150.31, 148.74, 140.58, 134.70, 133.20, 125.76, 121.50, 110.97, 83.35, 55.60, 24.64. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₂₀H₂₄BN₃O₄] + H)⁺, 382.1933, found 382.1929.

5.2.7. (p-OMe) SIH

(E)-N’-(2-hydroxy-5-methoxybenzylidene)isonicotinohydrazide

A similar procedure to the one described for SIH was followed, but replacing salicylaldehyde with 2-hydroxy-5-methoxybenzaldehyde. A yellow powder was isolated in 87% yield. ¹H NMR (DMSO-d₆, 500 MHz) δ (ppm): 12.26 (s, 1H), 10.50 (s, 1H), 8.80 (d, J = 4.9 Hz, 2H), 8.67 (s, 1H), 7.84 (d, J = 4.9 Hz, 2H), 7.17 (d, J = 2.4 Hz, 1H), 6.93 (dd, J = 8.7, 2.6 Hz, 2H), 6.87 (d, J = 8.7 Hz, 1H), 3.73 (s, 3H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 161.36, 152.17, 151.52, 150.36, 148.18, 140.07, 121.51, 118.98, 118.66, 117.34, 111.70, 55.51. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₁₄H₁₃N₃O₃] + H)⁺, 272.1030, found 272.1029.

5.2.8. (m-OMe) SNH

(E)-N’-(2-hydroxy-4-methoxybenzylidene)nicotinohydrazide

Equimolar quantities of nicotinic hydrazide (1 mmol, 0.137 g) and 2-hydroxy-4-methoxybenzaldehyde (1 mmol, 0.152 g), both in methanol were added to a 25-mL round-bottom flask equipped with a stir bar. The reaction mixture was stirred over an oil bath at 80 °C for 4 min. The precipitate was collected via vacuum filtration and washed with water to give a white powder in 85% yield. ¹H NMR (DMSO-d₆, 500 MHz) δ (ppm): 12.12 (s, 1H), 11.48 (s, 1H), 9.07 – 9.05 (m, 1H), 8.75 (d, J = 4.8 Hz, 1H), 8.54 (s, 1H), 8.24 (d, J = 8.0 Hz, 1H), 7.56 (dd, J = 8.0, 4.8 Hz, 1H), 7.46 (d, J = 8.5 Hz, 1H), 6.52 (dd, J = 8.5, 2.3 Hz, 1H), 6.49 (d, J = 2.3 Hz, 1H), 3.75 (s, 3H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 162.24, 161.18, 159.43, 152.37, 149.12, 148.57, 135.39, 131.05, 128.77, 123.64, 111.71, 106.57, 101.17, 55.34. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₁₄H₁₃N₃O₃] + H)⁺, 272.1030, found 272.1031.

5.2.9. SMTH

(E)-2-(2-hydroxybenzylidene)-N-methylhydrazine-1-carbothioamide

Equimolar quantities of 4-methyl-3-thiosemicarbazide (1 mmol, 0.105 g) and salicylaldehyde (1 mmol, 0.122 g), both in methanol were added to a 25-mL round-bottom flask equipped with a stir bar. The reaction mixture was stirred over an oil bath at 80 °C for 4 min. The precipitate was collected via vacuum filtration and washed with water to give a white powder in 72% yield. ¹H NMR (DMSO-d₆, 500 MHz) δ (ppm): 11.41 (s, 1H), 9.88 (s, 1H), 8.42 (s, 1H), 8.36 (s, 1H), 7.93 (d, J = 7.5 Hz, 1H), 7.21 (t, J = 7.5 Hz, 1H), 6.86 (d, J = 8.1 Hz, 1H), 6.82 (t, J = 7.4 Hz, 1H), 3.00 (s, 3H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 177.52, 156.31, 138.92, 130.94, 126.57, 120.48, 119.17, 116.03, 30.83. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₉H₁₁N₃OS] + H)⁺, 210.0696, found 210.0696.

5.2.10. SPTH

(E)-2-(2-hydroxybenzylidene)-N-phenylhydrazine-1-carbothioamide

A similar procedure to the one described above for SMTH was followed, but replacing 4-methyl-3-thiosemicarbazide with 4-phenylthiosemicarbazide. A white powder was isolated in 82% yield. ¹H NMR (DMSO-d₆, 500 MHz) δ (ppm): 11.76 (s, 1H), 10.04 (s, 1H), 8.49 (s, 1H), 8.08 (d, J = 8.1 Hz, 1H), 7.57 (d, J = 7.7 Hz, 2H), 7.36 (t, J = 7.6 Hz, 2H), 7.24 (t, J = 7.3 Hz, 1H), 7.19 (t, J = 7.3 Hz, 1H), 6.88 (d, J = 8.1 Hz, 1H), 6.84 (t, J = 7.5 Hz, 1H). ¹³C NMR (DMSO-d₆, 125 MHz) δ (ppm): 175.70, 156.59, 139.14, 131.31, 128.02, 127.93, 127.05, 125.67, 125.15, 120.25, 119.22, 116.03. HR-MS (ESI) (m/z): [M + H]⁺ calcd for ([C₁₄H₁₃N₃OS] + H)⁺, 272.0852, found 272.0852.

5.3. Calcein Fluorescence Assay for Prochelator-to-Chelator Conversion Efficiency

The relative Fe³⁺ binding affinity of chelators and peroxide-activated prochelators was assessed by a calcein fluorometric assay. The fluorescence probe calcein binds Fe²⁺ or Fe³⁺ to form 1:1 Fe-calcein complexes, which readily oxidize to the Fe³⁺ state in aerobic environments.^29,45 Experiments were performed at 25 °C using black-walled 96-well plates with clear bottoms in a Wallac Victor 1420 plate reader equipped with F485 and F535 incident and emission filters, respectively. The fluorescence emission of calcein is quenched when coordinated to Fe³⁺. Stocks of prochelators and chelators were prepared immediately before use at 100 μM concentrations in PBS buffer at pH 7.4. The reaction mixtures of prochechlators with H₂O₂ were generated by incubating a 100 μM working solution of prochelators with 5 mM H₂O₂ for 12 h. The Fe(calcein) stock solution was prepared by equilibrating 4 μM ferric ammonium citrate and 4 μM calcein (from a 1 mM stock in 1M NaHCO₃) in PBS buffer for 4 h to ensure formation of the ferric calcein complex. In each well, Fe³⁺-calcein at a final concentration of 2 μM was combined with 4 μM of chelators, prochelators, or reaction mixtures of prochelator with H₂O₂ in PBS buffer. Emission values from triplicate wells were averaged for each condition. Wells containing only calcein were used to determine the 100% chelation efficiency.

5.4. NMR Characterization of Hydrolytic Stability and Activation Products of Aroylhydrazone Prochelators

Stock solutions of prochelators (10 mM) in DMSO-d₆ were prepared for ¹H-NMR investigations of hydrolytic stability and reaction with H₂O₂. Portions of prochelator stock solution were diluted to 500 μM in D₂O with 5% DMSO-d₆, or 200 μM with 2% DMSO-d₆ in the (MD)BSIH case, and equilibrated for 1 h prior to NMR analysis. The peroxide activation of the prochelators was triggered by addition of concentrated H₂O₂ to give a final concentration of 25 mM. Products were identified after 24 h of equilibration by comparison of their ¹H NMR spectra to authentic samples. The percentage of chelator was calculated taking the ratio of integrated peak areas corresponding to pyridine H_a from the intact chelator vs. the sum of integrated peak areas for H_a of the chelator plus H_a’ of INH.

5.5. UV-Vis Spectrophotometry for Determination of Hydrolysis Equilibrium Constants of Aroylhydrazone Prochelators

UV-Vis absorption spectra were collected on a Cary 50 UV-Vis spectrophotometer using quartz cuvettes with 1-cm pathlength. For experiments tracking spectral changes of prochelator hydrolysis, the compositions of the prochelator solution in PBS buffer at pH 7.4 were monitored by collecting UV-Vis absorption spectra at 1 min intervals over 30 min.

To determine the hydrolysis equilibrium constant K of (p-OMe)BSIH and (MD)BSIH, the UV-Vis absorption spectra of equilibrated (p-OMe)BSIH or (MD)BSIH solutions in PBS buffer at various initial concentrations were recorded. The absorbance values at 263 nm and 321 nm from each spectrum was used to calculated the equilibrated concentrations of the intact prochelator and hydrolysis products INH and FBA based on Eq. 1, where ε represents the extinction coefficient at the indicated wavelength and [INH] or [FBA] refers to the equilibrated concentrations of isoniazid and substituted formyl phenyl boronic acid components. Briefly, the extinction coefficients for each prochelator, FBA and INH were determined from the UV-Vis spectra of solutions of each component at various concentrations in DMSO or PBS, pH 7.4. The absorbance at 263 nm or 321 nm vs. concentrations provides a linear plot where the slope is the corresponding extinction coefficient, ε. If [prochelator] was assumed to be x, [INH] = [FBA] = c_i – x, where c_i represents the initial concentration of BASIH added to the PBS solution. The readings of absorbance at 263 nm or 321 nm from the equilibrated UV-Vis spectra and the ε values were plugged into Eq. 1 to solve for x, which afforded the compositions of the equilibrated solution mixtures.

$${\text{A=ε}}_{\text{prochelator}}[\text{prochelator}]+{\text{ε}}_{\text{INH}}[\text{INH}]+{\text{ε}}_{\text{FBA}}[\text{FBA}]$$ (1)

The equilibrium constants K at each initial concentration of BSIH were further calculated according to Eq. 2 and averaged to give the final equilibrium constant of (p-OMe)BSIH and (MD)BSIH.

$$\text{K=}\frac{\text{[INH][FBA]}}{[\text{prochelator]}}$$ (2)

5.6. Cell Culture

All cell culture reagents, including minimal essential medium (MEM), Dulbecco’s modified eagle medium (DMEM), F12 Ham’s nutrient mix (F12), fetal bovine serum (FBS), pencillin– streptomycin (pen-strep), l-glutamine, and trypsin–EDTA were purchased from Gibco. The spontaneously immortalized human retinal pigment epithelial cell line ARPE-19 was purchased from American Type Culture Collection. The cells were grown in 1:1 DMEM and F12 medium supplemented with FBS (10%), pen-strep (1%), and glutamine (1%). Cells were cultured until confluent in 24-well Falcon plates and incubated at 37 °C in a fully humidified atmosphere containing 5% CO₂. CellTiter-Blue Cell Viability Assay was obtained from Promega. Cell viability assays were performed on a Perkin Elmer Victor³ 1420 plate reader.

For cytotoxicity assays, the growth medium was removed when cells reached confluence. Cells were then washed three times with MEM and treated with various prochelators at final concentrations ranging from 5–100 μM in MEM with up to 1% DMSO. After 72 h of incubation, cell viability was determined by the CellTiter-Blue assay. For cytoprotection against H₂O₂, washed cells were pre-treated for 5 h with various prochelators or chelators at the indicated concentrations prior to addition of H₂O₂ (200 μM final concentration). The cells were further incubated for 19 h before determination of viability. For cytoprotection against paraquat, washed cells were co-incubated with 10 mM paraquat and various prochelators or chelators for 48 h prior to determination of viability. The conditions compared in each experiment were: positive control (cells treated only with MEM), negative control (cells treated with MEM containing H₂O₂ or paraquat), and cells treated with H₂O₂ or paraquat and compounds at a range of concentrations. Each condition was run in triplicate and variability was determined as the standard deviation of the results.