Thiol-Reactive Arylsulfonate Masks for Phenolate Donors in Antiproliferative Iron Prochelators

Abstract

Tridentate thiosemicarbazones, among several families of iron chelators, have shown promising results in anticancer drug discovery because they target the increased need for iron that characterizes malignant cells. Prochelation strategies, in which the chelator is released under specific conditions, have the potential to avoid off-target metal binding (for instance in the bloodstream) and minimize unwanted side effects. We report a prochelation approach that employs arylsulfonate esters to mask the phenolate donor of salicylaldehyde-based chelators. The new prochelators liberate a tridentate thiosemicarbazone intracellularly upon reaction with abundant nucleophile glutathione (GSH). A 5-bromo-substituted salicylaldehyde thiosemicarbazone (STC4) was selected for the chelator unit because of its antiproliferative activity at low micromolar levels in a panel of six cancer cell lines. The arylsulfonate prochelators were assessed in vitro with respect to their stability, ability to abolish metal binding, and reactivity in the presence of GSH. Cell-based assays indicated that the arylsulfonate-masked prochelators present higher antiproliferative activities relative to the parent compound after 24 h. The activation and release of the chelator intracellularly were corroborated by assays of cytosolic iron binding and iron supplementation effects as well as cell cycle analysis. This study introduces the 1,3,4-thiadiazole sulfonate moiety to mask the phenolate donor of an iron chelator and impart good solubility and stability to prochelator constructs. The reactivity of these systems can be tuned to release the chelator at high glutathione levels as encountered in several cancer phenotypes.

Graphical Abstract

Arylsulfonate esters serve to mask the phenolate donors of prochelators that are activated by intracellular glutathione. Release of a tridentate thiosemicarbazone elicits iron sequestration and antiproliferative activities at submicromolar concentrations in a panel of cancer cell lines.

Introduction

Small-molecule chelators are employed routinely in the clinical practice to scavenge and excrete metal ions in cases of metal overload.^{1, 2} Chelation approaches are also undergoing intense investigation for their potential to target the roles of essential transition metals, such as iron, copper, and zinc, in neurodegeneration³ and cancer.⁴ Systemic chelation, however, poses safety concerns because these metal ions affect a broad spectrum of physiological functions, and indiscriminate sequestration is prone to unwanted side effects.

These considerations motivate the design of prochelator systems, in which activation under specific conditions is required to liberate the metal-binding species.^{5, 6} For instance, the reactivity of arylboronates with hydrogen peroxide has been employed to liberate phenolate donors under conditions of oxidative stress (Fig. 1, a).⁷ Alternatively, prochelators have been designed to act as substrates of specific enzymes, such as β-glucosidase⁸ and acetylcholinesterase⁹ (Fig. 1, b). We have developed disulfide-masked prochelators that are reductively activated in the presence of high concentrations of reduced glutathione (GSH) upon cellular uptake (Fig. 1, c).^{10, 11} Disulfide masks were employed in glucose^{12, 13} and albumin¹⁴ conjugates of improved selectivity and antiproliferative activity in cancer cells. Furthermore, we found that thiosemicarbazone prochelators affect the iron phenotype of tumor-associated macrophages and therefore have the potential to impact the tumor microenviroment.^{15, 16}

Figure 1.

Examples of prochelator activation strategies (R = alkyl or aryl; X = H, halogen, or NO₂; Ar = aryl or heteroaryl).

Because they require higher iron levels to sustain their rapid proliferation,¹⁷ cancer cells are susceptible to iron deprivation, and iron chelators are being pursued as anticancer agents.^{1, 4, 18} Prochelation approaches that rely on biological thiols are particularly attractive for cancer applications because of the high concentration of thiols in malignant cells relative to the surrounding tissue.¹⁹ In this study, we sought to incorporate oxygen donors in a prochelator design for activation by biological thiols. Specifically, we examine the use of arylsulfonate esters as thiol-reactive masks of phenolate donors in antiproliferative chelators (Fig. 1, d).

The reactivity of arylsulfonate esters with thiolates via nucleophilic aromatic substitution has been employed in biological settings for GSH detection by fluorescent sensors^{20, 21} and, less frequently, for the activation of prodrugs.^{22, 23} Within a prochelator design, we reasoned that this approach would prevent deprotonation and unwanted extracellular metal chelation of a phenolic oxygen, while also protecting it from metabolic processing. Unlike common masks for phenols in prodrugs, such as carboxylic and phosphate esters,²⁴ arylsulfonate esters do not require enzymatic cleavage and can be tuned to react in the presence of high intracellular concentrations of thiols.

As precursors for our prochelation design, we selected a small cohort of salicylaldehyde-derived thiosemicarbazone (STC) and carbazone (SSC) chelators (Fig. 2). Tridentate thiosemicarbazones are among the first classes of iron chelators to be investigated extensively as anticancer agents, and multiple compounds (e.g., Triapine, DpC) have been tested in clinical trials for cancer indications.^{25, 26} In several cases, their mechanism of action, particularly for α-N-heterocyclic thiosemicarbazones, involves not only iron deprivation but also the formation of redox-active iron and copper complexes resulting in oxidative stress.^{25, 26} Through simple modular synthesis, these compact scaffolds can be prepared with a variety of binding units and exhibit high affinity for transition metals as well as tunable lipophilicity and redox potentials. In this study, we designed new prochelators for thiosemicarbazones featuring the (O,N,S) and (O,N,O) donor sets, for which promising biological activities have been reported in various settings.^{1, 25, 26}

Figure 2.

Structures of salicylaldehyde-derived thiosemicarbazone (STC) and semicarbazone (SSC) chelators tested in this study.

Results and discussion

Synthesis and characterization of chelators.

We synthesized a series of STC and SSC chelators (Fig. 2) by condensation of 4-phenyl-thiosemicarbazide or 4-phenyl-semicarbazide with substituted salicylaldehydes (see ESI for synthetic details and characterization data). In particular, we revisited the effect of electron-withdrawing groups at the 5 position because it was previously found to increase antiproliferative activity in related compounds.^27-29 Promising antiproliferative activities of unsubstituted salicylaldehyde-derived chelator STC1 have been previously noted,³⁰ and substituted compounds STC4 and STC5 have been evaluated in leukemia cells albeit with a primary focus on the biological activity of their copper(II) complexes.³¹ Here, we examine the effect of the electron-withdrawing groups at the 5 position in a broader cancer cell panel in the context of iron chelation. Within this cohort, we sought to select an antiproliferative salicylaldehyde chelator for incorporation in a prochelation strategy.

The ability of the STC and SSC chelators to inhibit proliferation was assessed in a panel of malignant cells by MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay (Table 1). We have chosen two breast (MDA-MB-231, MCF-7), two ovarian (A2780, SK-OV-3), and two colon (Caco-2, HT-29) cancer cell lines because an altered iron metabolism has been observed in these cancer types, including in several clinical studies.^32-34 FDA-approved iron scavenger DFO was included in this assay as a reference, and its IC_S0 values in our panel were consistent with those in previous reports (IC₅₀, 2–20 μM; Table 1).^{35, 36} The salicylaldehyde-derived thiosemicarbazone chelators, STC1 to STC5, exhibited moderate to very good antiproliferative activity in all six cell lines (IC₅₀, 0.4–10 μM; Table 1). In general, these chelators are more active in breast and ovarian cancer cells than in colon cancers, and they are more effective than DFO. Conversely, as we previously observed,¹¹ the semicarbazone compounds SSC1 and SSC2 have lower activity than their thiosemicarbazone analogs.

Table 1.

Antiproliferative activities and phenolic pK_a values of STC and SSC chelators

Compounds	IC50 values (μM)[a]						Phenolic pKa[b]
	MDA-MB- 231 (Breast)	MCF-7 (Breast)	A2780 (Ovary)	SK-OV-3 (Ovary)	Caco-2 (Colon)	HT-29 (Colon)
STC1	4.8 ± 0.2	1.74 ± 0.04	1.9 ± 0.1	2.3 ± 0.2	7.3 ± 0.8	9.5 ± 0.5	8.25 ± 0.05
STC2	3.8 ± 0.7	1.73 ± 0.07	1.2 ± 0.1	2.1 ± 0.1	4.0 ± 0.4	8.9 ± 0.3	8.26 ± 0.03
STC3	1.0 ± 0.2	0.42 ± 0.01	0.85 ± 0.09	0.96 ± 0.03	2.6 ± 0.2	9.1 ± 0.1	7.91 ± 0.04
STC4	1.5 ± 0.3	0.41 ± 0.07	0.6 ± 0.2	0.9 ± 0.1	4.5 ± 0.1	6 ± 1	7.68 ± 0.05
STC5	1.2 ± 0.1	2.5 ± 0.3	1.04 ± 0.02	2.0 ± 0.1	3.0 ± 0.2	10.4 ± 0.3	5.96 ± 0.03
SSC1	48 ± 5	8.6 ± 0.4	19 ± 2	37 ± 1	55 ± 4	44 ± 8	N.D.
SSC2	12 ± 1	5.7 ± 0.4	5.5 ± 0.6	10 ± 2	9 ± 2	13 ± 1	N.D.
DFO	4.58 ± 0.03	22.5 ± 0.3	1.9 ± 0.1	9.5 ± 0.5	3.9 ± 0.5	2.3 ± 0.2	N.D.

^[a]IC₅₀ values from MTT assays in the indicated cell lines after exposure to the tested compounds for 72 h; values are presented as mean ± SDM, n = 3.

^[b]Apparent pK_a value and standard deviation determined via pH titrations monitored by optical absorption spectroscopy. N.D.: not determined.

The calculated octanol/water partition coefficients (logP_o/w, Swiss ADME tool, iLOGP method)³⁷ of several compounds in this series are between 2.5 and 4.0 (Table S1), a range that was previously found to be optimal for thiosemicarbazone chelators.³⁸ These calculated values, however, pertain to the neutral chelators, whereas a fraction of each species is expected to be deprotonated at neutral pH depending on the protonation equilibrium constants. Because the presence of negative charges affects cell permeability and iron affinity, the pK_a values for all STC chelators were determined through pH titrations monitored by optical absorption spectroscopy (Fig. S1, see ESI for experimental details). The deprotonation of the thiosemicarbazone moiety at high pH (i.e., above pH 11) was not investigated as it is not biologically relevant. Conversely, the deprotonation of the phenolic hydroxy group at central pH values (i.e., pH range 5–9) corresponded to a decrease in the absorption band at ~330 nm and the appearance of a new band between 360 and 390 nm, as previously reported.³⁹ The pK_a values in this series (Table 1) are lower than that of the unsubstituted (i.e., lacking the N4 phenyl) salicylaldehyde thiosemicarbazone (pK_a 8.8),³⁹ and the observed trend is as expected based on para substituent effects (i.e., NO₂ > Br ≈ Cl > F ≈ H).⁴⁰ Collectively, the data in Table 1 show that chelators STC3 and STC4, which feature intermediate lipophilicity and deprotonation constants in this compound cohort, exhibit the highest antiproliferative activities in the chosen panel of six malignant cell lines. These compounds likely combine effective cellular uptake of their neutral form with high-affinity binding of labile iron in the deprotonated form.

Given its antiproliferative activity and chemical properties (Table 1), we selected the bromo-substituted chelator STC4 for further characterization and for the new prochelation design. The iron complex of STC4 was prepared by combining the chelator with Fe(BF₄)₂·6H₂O (0.5 equiv.) in THF under aerobic conditions. The complex was crystallized from an acetone/pentane mixture, and single-crystal diffraction analysis (Table S2, Figs. 3, S2) indicated the isolation of a 2:1 complex associated to a [BF₄]⁻ counterion. The identity of this species was also confirmed by mass spectrometry and elemental analysis data. As observed for other (O,N,S) and (S,N,S) thiosemicarbazone chelators,^{35, 39} STC4 favors Fe(III) coordination and only the ferric species was isolated in the presence of oxygen. The Fe(III) complex is therefore expected to form intracellularly.

Figure 3.

Crystal structure of [Fe(STC4−H)₂]⁺ showing a partial atom labeling scheme and selected bond lengths (Å). Thermal ellipsoids are scaled to the 50% probability level. Carbon-bound hydrogen atoms in calculated positions as well as a [BF₄]⁻ counterion and two acetone molecules are not shown (CCDC 2178894).

The iron center in [Fe(STC4-H)₂]⁺ resides in a pseudo-octahedral geometry with two monoanionic tridentate ligands: for instance, the dihedral angle between the two chelator planes is 86°, and the bond angle between the opposite nitrogen donors (N1A-Fe-N1B) is 156.30°. The Fe-N and Fe-S bond lengths are characteristic of a high-spin configuration in iron thiosemicarbazone complexes.^{41, 42} For instance, in high-spin species, the Fe-N bond is typically longer than 2.1 Å and the Fe-S bond is generally longer than 2.35 Å as we observed in [Fe(STC4-H)₂]⁺ (Figs. 3, S2). The effective magnetic moment of the complex at room temperature is 5.6 ± 0.1 μ_B (Evans method), confirming a high-spin configuration for the d⁵ Fe(III) center.

Whereas thiosemicarbazone chelators derived from mercaptobenzaldehyde have so far resulted in low-spin ferric complexes,^{14, 35, 43} the presence of a phenolic donor leads to a high-spin configuration in [Fe(STC4-H)₂]⁺. Interestingly, the differences in ligand field are likely small, and chelators with similar binding units exhibit spin-crossover behavior.^{41, 42}

Synthesis and reactivity of prochelators.

Given its pK_a value of 7.68, a significant fraction of STC4 is in the anionic phenolate form at physiological pH values (e.g., ~58% in neutral form and ~42 % in phenolate form at pH 7.40). The negatively charged phenolate is less permeant through cellular membranes and more available for unwanted metal coordination in the extracellular space (and potentially in the bloodstream in more complex in-vivo settings). In addition, phenolic moieties are prone to phase II metabolism, for instance through O-glucuronidation or O-sulfation, leading to rapid clearance from the organism.⁴⁴ To avoid these complications, we sought to incorporate a prochelation strategy for the intracellular activation of the salicylaldehyde derivative STC4.

We developed a series of sulfonate ester prochelators (Fig. 4) designed for intracellular activation by GSH, which is present at millimolar concentrations in the cytosol and only micromolar levels in the extracellular space.⁴⁵ The generally higher GSH levels in rapidly proliferating malignant cells¹⁹ make this thiol-reactive approach particularly appealing to target iron in cancer cells with antiproliferative chelator STC4.

Figure 4.

Structures of sulfonate ester prochelators of STC4.

Prochelator STC4-S1 features the common 2,4-dinitro-benzenesulfonate (DNS) ester,²⁰ which has been previously employed in a variety of fluorescent probes for the detection of biological thiols.²¹ In STC4-S2, we included a benzothiazole-2-sulfonate (BTS) ester, which takes advantage of an electron-withdrawing heteroaromatic group and was employed successfully in coumarin- and fluorescein-based sensors.⁴⁶ In addition, we synthesized prochelators STC4-S3 and STC4-S4 carrying the 1,3,4-thiadiazole moiety, an important heteroaromatic motif found in a broad variety of pharmaceuticals and agrochemicals.^47-49 Its favourable solubility and stability profiles in aqueous solutions, as well as its electron-deficient nature, make the 1,3,4-thiadiazole ring an attractive moiety for the design of thiol-reactive sulfonate ester prodrugs. Furthermore, we reasoned that substitution at the 5 position (i.e., Me vs Ph in STC4-S3 and STC4-S4, respectively) would allow tuning of the electrophilic properties. To the best of our knowledge, 1,3,4-thiadiazole-2-sulfonate (TDS) esters have not been previously enlisted in a thiol-reactive prodrug strategy. Finally, prochelator STC4-S5 was included as a potential control of lower reactivity because its benzenesulfonate mask features a less electron-withdrawing 2-methoxycarbonyl substituent when compared to the DNS analog STC4-S1.

In all cases, the prochelator mask was installed on the salicylaldehyde precursor through reaction with a sulfonyl chloride reagent (see ESI for experimental details). The final product was then obtained through condensation with 4-phenylthiosemicarbazide in a simple step that is amenable to the future syntheses of sulfonate ester prochelators with a variety of iron-binding moieties from the same masked salicylaldehyde precursor. As expected for compounds with an additional aromatic ring, the calculated lipophilicity increased for all prochelators (relative to the parent STC4) with the exception of STC4-S1, which carries two nitro substituents. The new 1,3,4-thiadiazole-2-sulfonate (TDS) esters STC4-S3 and STC-S4 maintain an intermediate lipophilicity compatible with good solubility and cellular uptake through passive diffusion (Table S1).

In preparation for testing in our cancer cell panel, the prochelators were evaluated with respect to (i) their stability in cell growth media, (ii) ability to bind iron (or lack thereof), and (iii) reactivity with GSH in biologically relevant conditions.

In full growth media (containing 10% fetal bovine serum) at 37 °C, within the first 4 h the DNS-based prochelator STC4-S1 showed significant degradation (~40%), likely due to reactivity with thiols at micromolar concentrations in the media (vide infra). Conversely, the other prochelators remained essentially intact for the full observation period of 24 h (Fig. S3).

For all the prochelator structures, we found that the sulfonate ester moieties mask the phenolic oxygen donor in the tridentate binding units and hence prevent the thiosemicarbazone prochelators from coordinating iron. Whereas STC4 (25 μM) promptly binds Fe(II) (i.e., Fe(BF₄)₂-6H₂O, 12.5 μM) in phosphate-buffered saline (PBS) solution at pH 7.40, no changes are observed in the absorption spectra of the prochelators in the same conditions (Fig. S4).

The reactivities of the prochelators (100 μM) in the presence of physiological levels of GSH (10 mM) were evaluated at 37 °C in a mixture (1:1, v/v) of PBS solution (pH 7.40) and DMF, which ensured that all reactants and products remained in solution for the duration of the experiment. The progress of the reaction was monitored by HPLC through the consumption of the prochelators over a period of 24 h (Figs. 5, S5). The chromatograms indicated that the reaction proceeds cleanly: the generation of the chelator was confirmed by a retention time in line with that of authentic STC4, and the other products of the reaction (i.e., GS-Ar) were detected by LCMS, thus demonstrating the expected ipso attack by GSH on the arylsulfonate prochelators (Figs. S5-S6, ESI).

Figure 5.

The reactivity of the prochelators of STC4 (100 μM) with GSH (10 mM) in PBS solution (pH 7.4) and DMF (1:1, v/v) was monitored via HPLC analysis. The chelator release was determined from the peak area of prochelators relative to the peak area at 0 h.

Whereas the methoxycarbonyl-substituted arylsulfonate prochelator STC4-S5 remained unreactive throughout the observation period, the other compounds presented varying rates of chelator release in the tested conditions (Fig. 5). As observed in sensors and prodrugs featuring the DNS group, prochelator STC4-S1 is stable in neutral aqueous solutions^{50, 51} but has a fast reaction rate with thiols therefore complete conversion to the phenolic species is achieved within 2 h.^{22, 23} For the prochelators with heteroaryl sulfonate esters, the reactivity is slower, and the amount of STC4 released for STC4-S2 to STC4-S4 was 47%, 18% and 87% after 24 h, respectively. As expected, the STC4-S3 prochelator with a 5-methyl group on the 1,3,4-thiadiazole ring is less reactive than STC4-S4 with a more electron-withdrawing 5-phenyl group.

Collectively, our in-vitro characterization experiments indicated that the synthesized prochelators have stability and reactivity characteristics suitable for further biological testing in cultured cells.

Biological activity of prochelators.

We assessed the ability of the prochelators to inhibit proliferation in our panel of six malignant cell lines relevant to the study of iron deprivation in cancer. Within a 72 h timeframe, the antiproliferative activities of reactive prochelators STC4-S1 to STC4-S4 are at low micromolar values, as good as or slightly better than those of STC4 (Table S3, Fig. S7). In all cases, as expected for a largely unreactive prochelator, STC4-S5 presents much lower activity relative to the parent chelator STC4. Overall, the MCF-7 and A2780 cancer cells are the most susceptible to iron sequestration in this panel, with submicromolar IC₅₀ values for chelator STC4 and prochelators S1-S4.

The advantage of the prochelation design becomes more apparent in a 24 h timeframe, which is expected to highlight differences in cellular uptake and reactivity. In A2780 cells, prochelators STC4-S1, S2 and S4 present significantly higher antiproliferative activities (i.e., lower IC₅₀ values) relative to the parent chelator (Fig. 6 top panel, Table S4). The slow activation of STC4-S3 (Fig. 5) is consistent with the lower antiproliferative activity of this prochelator in the 24-h timeframe, and unreactive STC4-S5 is again basically inactive (Fig. 6 top panel, Table S4). As a comparison in non-malignant cells, we assessed the antiproliferative activities at 24 h in MRC-5 lung fibroblasts (Fig. 6 bottom panel, Table S4). The chelator and all the prochelators present lower antiproliferative activities in the normal cell line, and the difference is more pronounced for the prochelators. For instance, the 24-h IC₅₀ values of STC4-S2 and STC4-S4 are ~7 times higher in the non-malignant fibroblasts, possibly reflecting lesser iron demand and/or slower thiol-dependent activation in these cells.

Figure 6.

Cell viability assessed by MTT assays in A2780 cells (top) and MCR-5 cells (bottom) after exposure to tested compounds for 24 h.

The ability of STC4 and its prochelators to elicit iron sequestration was assessed in A2780 cells: variations in the intracellular labile iron pool were detected using the fluorescent sensor calcein-AM (i.e., calcein-acetoxymethyl ester).^{52, 53} After reaction with intracellular esterases that hydrolyze the acetoxymethyl groups, the fluorescein-based calcein sensor is partially quenched by iron binding in the cytosol (pH ~7). The presence of a competing iron chelator, which liberates a fraction of fluorescent calcein, is in turn detected as an increase in fluorescence. Within an incubation time of 3 hours, we found that all the salicylaldehyde-derived chelator STC4 resulted in an increase of calcein fluorescence similar to that of the high-affinity chelator SIH employed as a positive control (Fig. 7).⁵⁴ Within the STC4 prochelator series, only STC4-S5 failed to demonstrate iron-binding ability, likely because this compound is not activated effectively by intracellular thiolates (Fig. 5). All the other arylsulfonate prochelators elicited intracellular iron sequestration. The lower fluorescence increase for methyl-substituted TDS prochelator STC4-S3 is consistent with its slower reactivity with thiolates and decreased ability to liberate the STC4 chelator in the given timeframe (Fig. 5). These data indicated that prochelators S1-S4 are activated in cells and liberate an iron-binding species that successfully competes with calcein for intracellular iron.

Figure 7.

Intracellular iron binding reported by the calcein assay. A2780 cells were incubated with calcein-AM (0.2 μM in media without Phenol Red) for 30 min, washed, and then incubated with chelators or prochelators (10 μM) for 3 h. Significant fluorescence increases indicate effective competition for intracellular iron. Data presented are the mean ± SDM for three independent experiments (T-test relative to vehicle only (DMF 0.1% v/v), *** for p < 0.001 and ns for p >0.05).

The importance of iron sequestration for the antiproliferative activities of the prochelators was investigated through an iron supplementation assay. The viability of A2780 cells was evaluated for each compound in the presence and absence of ferric ammonium citrate (FAC), which alone had no significant effect on cell viability or growth (Fig. 8). The concentrations for the 24 h incubations were chosen close to the IC₅₀ values: 5 μM prochelators, 10 μM STC4, and 100 μM DFO as a positive control. Except for the inactive prochelator STC4-S5, iron supplementation of the growth media with equimolar FAC rescued the cells from the antiproliferative effects of the iron-binding compounds (Fig. 8). These experiments therefore confirm that iron sequestration is a significant component of the antiproliferative activity of the new prochelator systems.

Figure 8.

Assessment of cell viability (MTT assay, 24 h) in the presence or absence of ferric ammonium citrate (FAC). A2780 cells were incubated with DMF (0.1% v/v) in the presence or absence of FAC (100 μM) for 24 h as a control. Incubations with DFO (100 μM), STC4 (10 μM) or its prochelators (5 μM) were conducted in the presence or absence of equimolar FAC for 24 h. Data presented are the mean ± SDM for three independent experiments (*** for p < 0.001, ** for p < 0.01, * for p < 0.05 and ns for p >0.05).

The impact of STC4 and its prochelators on cell cycle progression was investigated by flow cytometry. It is well established that iron chelation has pronounced effects on cell cycle through multiple iron-dependent processes.⁵⁵ For instance, the inhibition of ribonucleotide reductase (RNR), a critical enzyme in DNA biosynthesis, is well documented in the case of tridentate thiosemicarbazone chelators.^{10, 56, 57} In general, iron chelation often leads to halting of the cell cycle in the early phases.⁵⁵ In our cell cycle analysis in A2780 cells, DFO was employed as an established iron scavenger and positive control.⁵⁸ To capture subtle changes in dynamic cell populations, we analyzed cell cycle distributions in A2780 cells upon 24-h incubations with DFO and STC4 at two concentrations close to their respective IC₅₀ values (for 24 h) (Fig. 9a). The higher concentrations resulted in a clear accumulation of cells in the G₁ phase with concurrent decrease of the percentage of cells in the G₂ phase compared to the untreated control. At the lower concentrations, a larger fraction of cells was able to proceed to the S phase but remained unable to reach the G2 phase.

Figure 9.

Effect of chelators (top panel) and prochelators (bottom panel) on cell cycle in A2780 cells. Cells were treated with the compounds (concentration as indicated, 24 h), harvested, fixed, pelleted, and then treated with RNAse and propidium iodide (0.5 mg/mL and 40 μg/mL, respectively, 30 min) before analysis by flow cytometry. Data presented are the mean ± SDM for three independent experiments (*** for p <0.001, ** for p <0.01, * for p <0.05 and ns for p >0.05).

For the prochelator series, we observed significant cell death when incubating at 10 μM concentrations (consistent with their low IC₅₀ values, Fig. 6) therefore accurate cell populations could not be measured. At lower concentrations (1.0 μM), however, all prochelators resulted in an S-phase increase after 24 h, accompanied by a decrease in the G₁ and G₂ fractions (Fig. 9b).

Overall, the effects of STC4 and the arylsulfonate prochelators on cell cycle are consistent with inhibition of proliferation, as expected for species that interfere with intracellular iron availability. In particular, the concentration-dependent impact on cell cycle and the S-phase population increase at low concentrations have been previously observed for iron chelators such as DFO and deferasirox (DFX).^{58, 59}

Conclusions

Tridentate thiosemicarbazone chelators continue to attract considerable attention as reliable metal-binding compounds for medicinal applications. Here, we revisited a series of thiosemicarbazones derived from salicylaldehydes featuring electron-withdrawing groups at the 5 position. The bromo-substituted chelator STC4 was found to have antiproliferative activities below 2 μM concentrations in a panel of six cancer cell lines of breast, ovary, and colon origin. STC4 stabilizes iron in a high-spin ferric configuration in a 2:1 ligand-to-metal complex that was characterized fully. The stabilization of Fe(III) is characteristic of the (O,N,S) thiosemicarbazones and is attributed to the harder oxygen donor in comparison with the (N,N,S) donor sets of α-N-heterocyclic thiosemicarbazones, which typically favor Fe(II) coordination.²⁵ Although the compounds of this STC series could also bind other transition metals in the cellular milieu (e.g., Cu(II), Zn(II)),³¹ iron is by far the most abundant transition metal chelatable in the cytosol (i.e., at micromolar levels),⁶⁰ and the disruption of iron homeostasis is an important aspect of the biological activity of thiosemicarbazone chelators.^{25, 26}

Because phenols are susceptible to phase II metabolism and are deprotonated at near-neutral pH, we sought to develop a prochelation strategy for the intracellular activation of these salicylaldehyde-derived compounds. We chose arylsulfonate esters to protect the phenolic oxygen from unwanted reactivity in the extracellular space. The aryl moieties were selected to facilitate reactivity by nucleophilic aromatic substitution with intracellular thiols. The compact 1,3,4-thiadiazole ring was employed for the first time in the context of a prochelator design in STC4-S3 and STC4-S4.

The envisioned prochelator reactivity was confirmed in vitro in the presence of millimolar GSH. Possibly owing to improved cellular uptake, the prochelators present higher antiproliferative activities relative the parent chelator after 24 h incubations in ovarian carcinoma cells. Notably, chelator STC4 and all its prochelators are significantly less active in non-malignant lung fibroblasts, likely underscoring the lesser dependance on iron of normal cells and/or the lower concentrations of GSH in their intracellular milieu.

The activation of the prochelators in cells was demonstrated by their ability to affect the availability of intracellular iron and to impact the cell cycle progression even at low concentration (1.0 μM) within 24 h. Additionally, cells could be rescued against the antiproliferative prochelators through iron supplementation. For future studies, the relatively slow reactivity of the 5-methyl-1,3,4-thiadiazole-2-sulfonate ester group could enable longer drug lifetimes and gradual release of metal-binding chelators, while also taking advantage of the more reducing environment of cancer cells. To achieve preferential activation in the tumor, the 1,3,4-thiadiazole-2-sulfonate group could be tuned for reactivity at a specific GSH threshold as provided, for instance, by a companion diagnostic sensor.²⁰ More generally, the sulfonate esters described herein could be employed to mask oxygen donors in a variety of chelators to be activated under reducing conditions in the presence of biological thiols.