Boronic Acids and Beyond: ROS-Responsive Prodrugs as Tools for a Safer and More Effective Cancer Chemotherapy

Abstract

Despite significant scientific advances and the wide variety of available treatments, cancer remains a major cause of death worldwide. Chemotherapy, which is frequently one of the first-line treatments, frequently suffers from low selectivity to cancer cells, leading to the appearance of important side effects. Thus, it becomes imperative to develop a new generation of targeted alternatives that spare the healthy tissues by delivering the cytotoxic payloads safely and selectively to cancer cells. In this respect, prodrugs that are activated by tumor-specific stimuli have attracted significant attention. Despite being a hallmark of cancer and present in high concentrations in cancer cells, reactive oxygen species (ROS) have been rather underexplored as a stimulus for the preparation of targeted prodrugs, particularly when compared with an acidic pH or glutathione. Despite their lower expression, ROS have recently been gaining substantial consideration, with various ROS-responsive prodrugs already reported with meaningful performances both in vitro and in vivo. This review aims to provide critical insights into this strategy by discussing the various available functional groups (with an important focus on boronic acids and their esters), their mechanisms of action, examples of their applications, advantages, limitations, and future challenges.

1 Introduction

2 Boronic Acids and Boronate Esters

2.1 Histone Deacetylase Inhibitors

2.2 DNA Alkylating Agents

2.3 Selective Estrogen Receptor Modulators and Selective Estrogen Receptor Degraders

2.4 ROS Inducers

2.5 Prodrugs Based on Other Types of Anticancer Drugs

3 Other ROS-Responsive Moieties

3.1 Thiazolidinones

3.2 1,3-Oxathiolanes

3.3 Selenium Ethers

3.4 Sulfur-Containing ROS-Responsive Moieties

4 Summary and Future Perspectives

Introduction

According to the World Health Organization, cancer was responsible for over 10 million deaths in 2020, making it one of the leading causes of death worldwide. Early detection and effective treatment are crucial to improve the prognoses of cancer patients. However, for this to happen, a correct diagnosis of the disease is also essential, as the treatment provided will differ depending on the cancer phenotype.[1] Surgery is frequently the first line of treatment for solid tumors, but additional treatments are often required, such as radiotherapy, chemotherapy, hormone therapy, or targeted biological therapies.[2]

Although the conventional treatment strategies have demonstrated considerable success in a variety of cancers, important issues such as drug resistance, recurrence, and metastasis still plague their success. On top of that, lack of selectivity is still a major concern in chemotherapy. Most chemotherapeutic agents are cytotoxic molecules that interfere with cell division by microtubule disruption, blocking DNA replication, or direct DNA damage. Although these agents are very effective in killing cancer cells, their lack of selectivity leads to the death of other fast-replicating cells, such as blood cells, skin cells, and liver cells, among many others. It is this nonspecific targeting that promotes the appearance of the well-known chemotherapy side effects: hair loss, fatigue, nausea, etc.[3] [4] [5] Hence, the development of novel technologies featuring unique and alternative mechanisms of action is essential to decrease the disease burden and reduce its side effects.

One of the strategies with the potential to overcome this problem and to increase drug accumulation in cancer cells is the development of prodrugs.[6] Prodrugs are molecules with no pharmacological activity that are converted into the active parent drug in vivo by specific enzymatic or chemical reactions. These prodrugs can be classified as bioprecursor prodrugs or as carrier-linked prodrugs. Bioprecursor prodrugs are compounds that result from a direct molecular modification of the active agent itself, and do not contain a carrier or promoiety, being transformed into the corresponding active molecule through metabolic or chemical alterations. In carrier-linked prodrugs, the active drug is covalently linked to a carrier (also known as a promoiety) and, upon biotransformation at the desired site, both the active drug and the carrier are released.[6] [7] [8] Moreover, prodrugs can offer a simple and straightforward way to overcome other challenges encountered in drug formulation and delivery, such as low solubility, chemical instability, and insufficient oral absorption.[8]

In the case of tumor-targeting prodrugs, the concept is to use specific properties of the tumor microenvironment to control the local activation of the prodrug and the consequent decaging of the active drug specifically in tumor cells.[6] [9] [10] Fortunately, the tumor microenvironment displays various specific properties, such as an acidic pH, overexpression of proteases, and high concentrations of glutathione, which have enabled the development of a considerable number of tumor-targeting prodrugs.[6,9]

In this review, the focus will be placed on well-defined prodrugs that are responsive to reactive oxygen species (ROS). ROS are a family of highly reactive molecules with unpaired electrons, formed from molecular oxygen; examples include the superoxide anion (O₂ ^•–), hydrogen peroxide (H₂O₂), and hydroxyl radicals (HO^•).[6] [10] [11] Of these, H₂O₂ is the most stable and the most common ROS in cancer cells, being found in relatively higher concentrations in cells compared with the other species.[6] [12] As such, H₂O₂ is very often used as a representative type of ROS to assess the ROS-responsive performance of prodrugs.[13]

The major source of endogenous ROS in cells is the mitochondrial respiratory chain, and ROS play an important part in cellular signaling and metabolism.[11] [13] [14] In healthy cells, there is an intracellular reduction–oxidation balancing system that allows ROS levels to be kept low (~20 nM). However, cancer cells have been found to possess a dysfunctional redox balance that leads to an increased generation of these species, reaching concentrations of up to 1 mM.[12–14] This oxidative stress phenomenon is an outcome of a disruption in the tumor’s redox homeostasis, due to either an elevation of ROS production or a decline in ROS-scavenging capacity, and it contributes to DNA alterations, cancer-cell proliferation, apoptosis, angiogenesis, metastasis, and cellular sensitivity to anticancer drugs.[7,14]

Regarding their structure, ROS-activated prodrugs can be composed of three functional units, an ROS-accepting moiety (trigger), an effector, and an optional self-immolative linker that connects the trigger and the effector (Scheme [1]).[12] [15] The trigger unit is the precursor for the activation mechanism. Once this group is specifically metabolized by ROS in tumor cells, the linker should quickly transmit the trigger’s signal to activate the effector and produce a significant cytotoxic effect.[7,15,16]

The present Account aims to summarize and provide insight into recent strategies employed in the design of ROS-responsive prodrugs. A recent review has also been published that describes the use of boronic acids in the treatment of cancer, including their use as ROS-responsive prodrugs.[17] That review, which is organized primarily by the type of cancer and the available boronated therapeutic options, is particularly focused on the biological properties of prodrugs, including their pharmacokinetics and pharmacodynamics. Differently, this review will discuss the different classes of ROS-responsive moieties that have been developed, with a focus on the activation mechanism of the prodrug by ROS and the release of the active drug; it also discusses the advantages and limitations of each methodology. As a side note, ROS-responsive units have also found considerable success in the design of supramolecular drug-delivery systems. These systems, however, are not considered well-defined prodrugs and, therefore, are not included in this review and have been highlighted elsewhere.[15] ^, [18] [19] [20] [21]

Boronic Acids and Boronate Esters

Boron is a metalloid element present ubiquitously in nature, with a special presence in fruits, vegetables, and nuts. Although for a long time it was believed that boron-containing compounds were toxic, this belief has been set aside as boron has been revealed to be an essential nutrient in plants, even though its physiological functions are still poorly understood.[22] [23] [24] Organoboron compounds, in particular, have good physiological tolerances as they are degraded through metabolic oxidation to the stable boric acid, which is known to be a safe and benign molecule that is readily eliminated by the body (Figure [1]).[23,24] Moreover, being an element of group 13 of the Periodic Table, boron is characterized by having three valence electrons and an empty p-orbital with an electrophilic character, which promotes a dynamic interaction with Lewis bases.[23] [25] [26]

Among the various organoboron compounds, it is possible to highlight boronic acids (BAs) as the most common objects of study in organic chemistry owing to their stability towards air and moisture, and the wide availability of methodologies for their preparation in high purity.[23] [25] [26]

BAs are trivalent boron-containing organic compounds that possess one carbon-based substituent (i.e., a C–B bond) and two hydroxy groups (Figure [1]) positioned in a trigonal planar geometry as a consequence of the sp² hybridization of the boron atom.[25] Due to their vacant p-orbital, BAs, like other boron-containing compounds, behave as Lewis acids.[27] Thus, they display pK _a values in the range 4–10, depending on the substituents on the BA. Since the electrodeficiency on the boron center correlates directly with its Lewis acidity, aryl BAs and BAs featuring electron-withdrawing groups tend to have lower pK _a values than their alkyl congeners. Hence, at physiological pH values, these compounds are in their uncharged trigonal form, converting into the anionic tetrahedral form when the pH of the medium exceeds the pK _a value of the compound (Scheme [2]).[23]

Because they are not found in nature, BAs must be prepared from primary sources of boron. The first synthesis of a boronic acid was performed by Frankland and Duppa in 1860, who treated triethyl borate with an excess of diethylzinc, resulting in triethylborane, which, due to its instability, was slowly converted into a more stable product, ethylboronic acid.[27] [28] On the other hand, boronic esters (BEs) can be prepared by reacting the corresponding boronic acid with an alcohol or a diol.[25]

BEs are usually less reactive than BAs because these species have reduced Lewis acidity. This effect results from the σ-donating ability of the carbon, which makes the lone pairs of oxygen in boronic esters more readily conjugated to the electron-deficient boron center.[29] In addition, they are also less polar and are easier to handle because they lack hydroxy groups, known for their hydrogen-bond donor capability.[25] [29] Acyclic boronic esters are hydrolytically unstable but the cyclic five- and six-membered esters, formed from the reaction between BAs and 1,2- or 1,3-diols, respectively, are more stable.[30] Among those, pinacol esters are the most popular and these are frequently used as surrogates for boronic acids, with improved stability and handling.

The use of BAs and their esters flourished in organic chemistry with their application in the Suzuki–Miyaura reaction, in which they undergo cross-coupling with organohalides (or triflates) in the presence of a palladium catalyst and a base. This exceptionally important reaction in organic chemistry permits the straightforward formation of C–C bonds, giving rise to functionalized alkenes, styrenes, or biaryl compounds.[31] Nonetheless, due to their aforementioned low toxicity, high stability, adjustable Lewis acidity, and dynamic coordination profile, BAs have become attractive assets in other areas, such as medicinal chemistry, analytical chemistry, materials science, and pharmaceutical chemistry.[27]

In medicinal chemistry, these compounds, especially arylboronic acids and their esters, have shown great potential in the design of ROS-responsive prodrugs and self-immolative groups. That is because the electron-deficient boronate group can undergo oxidation by H₂O₂ (the most common ROS in cancer cells) to generate the corresponding phenol.[6] [24] [32] When designing a boron-based ROS-responsive prodrug, the active drug can be directly connected to the BA/boronate groups or joined through a self-immolative spacer. In the direct attachment of the active drug to the BA/boronate functionality, it is only possible to mask alcohol-containing drugs (including enols or phenols), whereas the use of a self-immolative spacer permits the use of masking effectors containing alcohol, amine, or hydroxamic acid functions through ether, carbonate, carbamate, or hydroxamate bonds (Scheme [3]).[10,12]

The reaction between the BA/boronate group and H₂O₂, which is iorthogonal and biocompatible, is illustrated in Scheme [4]A. This reaction begins with the attack of the oxygen of H₂O₂ (or its deprotonated anionic form) on the electron-deficient boronic acid to form a tetrahedral boronate intermediate. Then, a 1,2-rearrangement occurs, and the product obtained spontaneously undergoes hydrolysis to form a deboronated alcohol derivative and boric acid. When the alcohol derivative obtained is part of a self-immolative linker (Scheme [4]B), the process continues, and a spontaneous 1,6-elimination reaction takes place, leading to the release of the active drug and para-quinone methide (para-QM).[12] [24] [33] The latter can undergo hydrolysis or can react with glutathione (GSH) present in the cytosol, producing 4-hydroxybenzyl alcohol (4-HBA) or a GSH–QM adduct, respectively.[34] Because GSH has an important role as an intracellular antioxidant, the GSH-QM adduct formed can deplete intracellular reserves of GSH and weaken the cell’s antioxidative abilities. This depletion leads to an elevation of oxidative stress and increased ROS generation, with the consequent death of cancer cells due to their vulnerability to further ROS-associated toxicity.[11,35]

In addition, QMs have already been reported as being able to alkylate DNA, and there are various examples in which ROS-activated QM prodrugs with a boronic pinacol ester as a trigger unit have been developed. These will be detailed later in this Account.[36] [37] [38] [39]

The following sections summarize various ways of exploiting this ROS-responsive behavior of BAs in the design of prodrugs with increased specificity and selectivity toward cancer cells. The boron-based ROS-activatable prodrugs covered in this work are precursors of histone deacetylase inhibitors, DNA alkylating agents, selective estrogen receptor modulators and selective estrogen receptor degraders, ROS inducers, and other miscellaneous classes.

Histone Deacetylase Inhibitors

The synthesis of proteins occurs through a process called translation, which involves the transcription of DNA into RNA. After being synthesized, and at any stage of their lifespan, proteins can still undergo reversible or irreversible chemical changes in their polypeptide chain, including enzymatic cleavage of peptide bonds or covalent additions of particular chemical groups, lipids, carbohydrates, or whole proteins to amino acid side chains. These events, referred to as posttranslational modifications (PTMs), are catalyzed by specific enzymes and permit a broadening of the variety of amino acid structures and properties, diversifying the structures and functions of proteins.[40] These modifications, called epigenetic modifications, have a crucial role in gene regulation, as they are able to influence the structure of chromatin and gene expression without modifying the DNA sequence.[41]

In particular, PTMs in the histones of chromatin play a key role in regulating gene expression.[42] Histones are proteins found in eukaryotic cells and are associated with DNA, as well as nonhistone proteins, to form a complex called chromatin. The structure of this complex, particularly its compression and relaxation, has an important effect on gene transcription, which is the first step in the process of gene expression. In turn, the structure of chromatin is influenced by the PTMs of histones, including the methylation, phosphorylation, and acetylation of lysine residues and the N-terminal regions present in their core.[43] [44]

Acetylation of histones occurs in the ε-amino group of lysine residues and is a highly dynamic process controlled by the activity of two groups of enzymes: histone acetyltransferases (HATs) and histone deacetylases (HDACs).[43] [45] The HATs catalyze the transfer of an acetyl group from acetyl-CoA to the lysine residue. When this happens, the net positive charge on the histone proteins is neutralized, which causes a reduction in the binding strength between the histone and the negatively charged DNA phosphodiester backbone. Consequently, a relaxation of the nucleosome structure at the modified site is observed, exposing DNA to transcriptional proteins. Conversely, HDACs are responsible for reverting the acetylation, which means that the lysine sidechain becomes positively charged again, making it possible to restore DNA packaging and, as in this case, the chromatin is more condensed and ceases to be accessible to the cell’s transcriptional machinery, leading to a suppression of transcription.[42,45]

Abnormal alterations in histone acetylation and the overexpression of various HDAC isoforms (HDACi) are related to the development of various pathologies, including cancer.[46] [47] [48] Hence, inhibitors that target this enzyme family have been designed, and these small molecules usually have thiol, benzamide, or hydroxamic acid functionalities in their pharmacophores. Four HDACi have been approved by the United States Food and Drug Administration (FDA), three of which have a hydroxamic acid residue that chelates the Zn(II) ion present in the active site of the HDACi. The three hydroxamic acids, vorinostat, belinostat, and panobinostat, are indicated for cutaneous T-cell lymphoma, relapsed/refractory peripheral T-cell lymphoma, and multiple myeloma, respectively.[46,48] However, following an accelerated approval process, panobinostat was removed from the market in 2021 after the pharmaceutical company responsible for its commercialization announced that it was unable to complete the required post-approval clinical studies.[49]

Despite appearing promising, these drugs have some drawbacks that have negative impacts on their activity, including chemical instability and rapid metabolic elimination. These problems arise because the hydroxamate group of HDACi can either undergo glucuronidation and sulfation or can be hydrolyzed to the carboxylic acid analogue. These reactions give rise to inactive metabolites before the drug can reach its intended target.[46] [48] Furthermore, these compounds not only have poor tissue penetration, caused by the highly polar hydroxamic acid residue, but also display side effects and toxicity in normal cells by nonselectively inhibiting most isoforms of the metal-dependent HDACs. Accordingly, these types of drugs appear to be good candidates for the design of prodrugs.[46,50]

To improve the selectivity toward cancer cells over normal cells, in 2018 Bhagat et al. developed compound 1, a prodrug of vorinostat (Scheme [5]), by masking the hydroxamic acid residue with an H₂O₂-responsive unit, a 4-[(hydroxymethyl)phenyl]boronic pinacol ester.[46] As expected, the prodrug and vorinostat exhibited differences in their physicochemical properties, such as lipophilicity, with prodrug 1 showing an increased lipophilicity (clogP = 4.337 ± 0.3) in comparison with the parent drug (clogP = 0.989 ± 0.5). This increase in lipophilicity also led to a markedly accelerated cellular drug uptake of 1. Moreover, 1 was shown to be stable in serum, but it was able to react with H₂O₂ with breakage of the C–B bond and rapid release of the parent drug. Furthermore, the researchers compared the activity of the two compounds and, based on the half-maximal effective concentration (EC₅₀) value, they found that compound 1 was substantially less toxic to HEK-293T cells (healthy cells) (EC₅₀ > 1 μM) than was vorinostat (EC₅₀ ≈ 100 nM). The prodrug was also more active than the parent drug in three out of the four cancer cell lines used in this study (cervical carcinoma HeLa, breast cancer MCF-7, and musculus skin melanoma B16-F10 cells). For instance, in HeLa cells, the prodrug reached an EC₅₀ = 115 nM, whereas vorinostat has an EC₅₀ = 210 nM. In addition, the prodrug reduced the tumor size by about 80% (at a concentration of 200 nM), thereby displaying improved effectiveness compared with the parent drug, which only achieved a size reduction of about 50%.[46]

In the same year, compound 2, another prodrug of vorinostat (Scheme [5]) for the treatment of acute myeloid leukemia was reported by Liao et al.[50] Unlike prodrug 1, prodrug 2 has a BA instead of a BE as a masking group. Besides increasing the lipophilicity of the parental hydroxamic acid HDACi, it retains water solubility. Another difference of this prodrug is that it can be activated by both H₂O₂ and peroxynitrite (PNT). However, in the cancer cell lines tested (human acute myeloid leukemia U937 and MV4-11 cells), this compound was less cytotoxic than the parent drug (IC_50,2 = 6.13 μM versus IC_{50,vorinostat} = 0.91 μM for U937 cells and IC_50,2 = 3.07 μM versus IC_{50,vorinostat} = 0.50 μM for MV4-11 cells).[50]

Furthermore, in 2018, Zheng et al. designed and synthesized the belinostat prodrug 3 (Scheme [5]) to enhance the therapeutic efficacy of this HDACi against solid tumors.[48] However, on evaluating the antiproliferative activity of the prodrug through in vitro assays in four cancer cell lines (human lung carcinoma A549, cervical carcinoma HeLa, and human breast adenocarcinoma MDA-MB-231 and MCF-7 cells), the researchers found that prodrug 3 was about three-to-six times less potent than belinostat under the same conditions. To rationalize this result, the concentrations of the prodrug and of the active form in the MDA-MB-231 cell-culture medium were analyzed. After one day of incubation of MDA-MB-231 cells with the prodrug, the researchers found a complete conversion of 3 into the product through hydrolysis of the pinacol ester, as this compound was present in a concentration of 331.8 ng/mL. On the other hand, the prodrug was only partially converted into belinostat, whose measured concentration was 30.7 ng/mL. The HDAC inhibitory activity of the prodrug was also measured and was shown to be about ten times lower than that of belinostat, which was consistent with their difference in antiproliferation activities against cancer cell lines and is rationalized in terms of the incomplete oxidation and release of the active drug. However, in mice, prodrug 3 inhibited tumor growth and reduced tumor volume after three weeks of treatment with a dosage of 10 mg/kg/day, displaying a significantly increased efficacy over the original compound, contrary to what was observed in all the in vitro assays. This might be explained by the fact that the prodrug has a better bioavailability than belinostat, because it was found that, after 22 days of treatment, mice that had been administered with the prodrug had a higher concentration of the active drug in their tumor tissues than those that had been administered the active drug directly. Taken together, these findings suggest a promising clinical application for compound 3.[48]

DNA Alkylating Agents

Alkylating agents are a class of anticancer drugs that interfere with the DNA of cancer cells across multiple stages of the cell cycle, inhibiting their growth and eventually leading to their death.[51] [52] These drugs are characterized by their high electrophilicity because, under physiological conditions, they form transient positively charged groups that can react with nucleophilic biological molecules such as DNA or proteins.[3,51] Consequently, the biological function of DNA and proteins is altered as a result of the irreversible covalent addition of exogenous molecules to their nucleophilic centers (oxygen, nitrogen, phosphorus, or sulfur atoms).[50]

These antitumor agents exert their activity by reacting with the bases of DNA, with a special preference for guanine, which is normally alkylated at the N7 position (Figure [2]). However, other nucleobases can be alkylated, and this process can occur in various positions in the following preferred order: the N1 and O6 positions of guanine; the N1, N3, and N7 positions of adenine; the N3 position of cytosine; and the O4 position of thymine (Figure [2]).[51]

Depending on the number of alkylating groups they possess, these agents can be classified as either monofunctional alkylating agents or bifunctional alkylating agents. The former have a single reactive group, so they only react with one strand of DNA, whereas the latter, by having two reactive groups, can either crosslink two strands of DNA (interstrand crosslinking), when the electrophilic groups react with nucleic acid bases from opposite strands of the DNA helix, or covalently connect two nucleotide residues present on the same DNA strand (intrastrand crosslinking) (Figure [3]A).[3] [53] In the presence of interstrand crosslinking, transcription and replication are disrupted, and the cell dies by apoptosis or necrosis due to its ineffective attempts to repair the DNA. Intrastrand crosslinking prevents the portion of DNA that is alkylated from being recognized by the enzymes needed to catalyze its replication and transcription.[3,54] Alkylating agents can also modify the structure and function of proteins through alterations in their amino acid sequence through miscoding. This is because, after undergoing alkylation, guanine favors the enol tautomer over the keto tautomer and is therefore more likely to base-pair with thymine rather than with cytosine (Figure [3]B).[3] However, intrastrand crosslinking is the primary mode of action for most alkylating agents.[53] [55]

Alkylating agents can be grouped into various families, including nitrogen mustards; nitrosoureas; alkyl sulfonates; hydrazine and triazine derivatives; aziridines and epoxides; ethylenimines; benzoquinone-containing agents; illudins; and platinum-containing agents.[51] [52]

Due to their nonspecific mechanism of action, alkylating agents exert their cytotoxic activity not only in cancer cells, but also in other fast-replicating cells such as bone marrow and gastrointestinal cells. This off-target toxicity is responsible for the appearance of classical chemotherapy side effects. Additionally, alkylating agents can even lead to the onset of other types of cancer or other diseases.[3] [52] [53] For these reasons, the design of tumor-targeted prodrugs could significantly improve the therapeutic efficacy and safety profile of these compounds.

Consequently, the sections below address how alkylating agents, namely nitrogen mustards and quinone methides, have been used as bases for the design of BA-containing ROS-responsive prodrugs.

Nitrogen Mustards

DNA alkylating agents with nitrogen mustards were one of the first effective chemotherapeutic agents discovered and, even more than 70 years after their discovery, they remain important drugs for the treatment of many forms of cancer.[56] [57] A common feature of the nitrogen mustards is that all possess a bis(chloroethyl) group attached to a nitrogen atom, which is essential for their mechanism of action (Scheme [6]). The rest of the prodrug is responsible for modulating its physicochemical properties and its transport, distribution, and reactivity.[57]

The process of DNA alkylation by this type of drugs is illustrated in Scheme [6], and it occurs in two steps. First, at a neutral or alkaline pH, an intramolecular cyclization of the bis(2-chloroethyl)amine moiety proceeds through an S_N2 reaction, forming an aziridinium ion. Due to its ring strain and positive charge, the aziridinium ring that is formed is an unstable and highly reactive species that can be attacked by a DNA nucleophile to form a covalent bond and yield a monoalkylated adduct through an S_N2 mechanism. Then, a second intramolecular cyclization can occur with the remaining chloroethyl group, followed by another DNA alkylation step to generate a crosslink between two complementary strands of DNA.[56] This crosslink forms preferentially between the N7 positions of two guanine residues located in sequences 5′-GNC-3′, where N represents any nucleobase (Scheme [6]).[51]

In 2011, Kuang et al. reported the two prodrugs 4 and 5 (Figure [4]) that use mechlorethamine as the nitrogen mustard; the difference between these two compounds lies in the type of boronate derivative used.[58]

On investigating the ability of compounds 4 and 5 to form DNA interstrand crosslinks, it was found that, in the absence of H₂O₂, neither was able to form a DNA interstrand crosslink (ICL), confirming that the toxicity of the nitrogen mustard is masked in the designed prodrugs. The researchers suggest that the masked cytotoxicity of these compounds is associated with the establishment of a positive charge on the nitrogen, which decreases the electron density of the nitrogen mustard, hindering the generation of the aziridinium ring. On the other hand, in the presence of H₂O₂, both compounds achieved an efficient crosslink formation with a yield of 43%, whereas the nitrogen mustard alone reached a value of 47%. These results showed that the compounds under study are toxic to DNA only when they are activated by H₂O₂ and release mechlorethamine. Furthermore, the activation of the prodrugs was selective for H₂O₂ against other ROS. These compounds showed an inhibition of cell growth of between 57 and 90% at a concentration of 10 μM against SR (leukemia cells), NCI-H460 (non-small cell lung cancer cells), and CAKI-1 and SN12C cell lines (renal cancer cells), but no increase in apoptosis when similar assays were performed on normal lymphocytes.[58]

Despite these promising results, charged molecules are notorious for their poor membrane diffusion and low intracellular accumulation, making them unsuitable for drug development. This prompted the researchers to report a novel strategy for creating neutral H₂O₂-activated prodrugs. For this purpose, they designed and synthesized compounds 6a–f and 7 (Figure [4]) in which the nitrogen mustard is directly attached to an aromatic ring. Due to the weak electron-withdrawing nature of the boronyl group, the lone pair of the nitrogen mustard is delocalized to the aromatic ring and is less prone to forming an electrophilic aziridinium ring (Scheme [7]). Upon oxidation by H₂O₂, the electron-rich hydroxy group that is formed stabilizes the electron pair on the nitrogen and facilitates the intramolecular cyclization.[55]

To understand the influence of the leaving group, the researchers prepared compounds 6a–f and evaluated their performance in the presence and absence of H₂O₂. Most compounds tested displayed no ICL in the absence of H₂O₂, with the notable exception of 6f, which features two mesylate groups (22% DNA crosslinking). Due to their very high electrophilic nature, the electron-withdrawing effect of the BAs is unable to mask the nitrogen effectively to prevent the formation of the aziridinium ring. Compound 6f is, therefore, unsuitable for the development of prodrugs. However, compounds bearing only one mesylate group (6d and 6e) displayed adequate stability in the absence of H₂O₂ and achieved better ICL results than those bearing halogen leaving groups (6a–c). Hence, mesylate groups were shown to be viable alternatives to classical halogen-based mustards in the design of ROS-responsive prodrugs.

Also, the activity of prodrugs 6a and 7 was triggered more efficiently by H₂O₂ in comparison with the other ROS tested. In an in vitro cytotoxicity assay, these compounds were able to inhibit the growth of the sixty tested cancer cell lines. With a single dose of 10 μM, the growth percentages of the cancer cell lines were lower than 50%, and most compounds presented a half-maximal growth inhibition (GI₅₀) value of less than 5 μM. Finally, 6a and 7 also displayed adequate selectivity against leukemia cells, with a percentage of apoptosis between 40 and 80%, while being less toxic toward normal lymphocytes, with a cell death percentage lower than 25%.[55]

Later in 2014, this group reported three new nitrogen mustard prodrugs 8–10 with an H₂O₂-cleavable trigger (Figure [5]). In these prodrugs, the effector is once again an aromatic nitrogen mustard and the trigger moiety is an arylboronate. In this case, however, an electron-withdrawing group was used as a linker to merge the two units. These compounds were designed to evaluate whether the aromatic nitrogen mustard reactivity could be tuned through the variation of the aromatic substituents. Thus, prodrugs containing a quaternary ammonium salt (8), a carbamate (9), or a carbonate group (10) were prepared and evaluated.[59]

The newly prepared prodrugs performed satisfactorily, displaying over 20% ICL in the presence of H₂O₂, whereas a negligible ICL was observed in its absence. Once more, H₂O₂ was found to be the ROS that activates these compounds most efficiently, inducing about 20% ICL formation, whereas less than 5% ICL was observed with other ROS. Among the three compounds, prodrug 10 displayed the highest cytotoxicity, as it inhibited growth in most of the 25 cancer cell lines studied at a concentration of 10 μM. These results suggest that a neutral carbamate linker is preferable in the design of this type of ROS-activated prodrug.[59]

In 2018, the same research group reported a novel family of aromatic nitrogen mustard prodrugs. This time, using the two previously reported compounds 6a and 6f as a starting point, they introduced substituents onto the aromatic ring and evaluated their effects on the DNA alkylation ability.[55] [60] The studied compounds (Figure [6]), featured electron-donating (11 and 12) or electron-withdrawing groups (13) to study the relationship between electronic effects and biological activity. Amino acid side chains (14–17) and pinacol boronate groups were also included to improve the physicochemical properties of the prodrugs.[60]

After being incubated at 10 μM with sixty cancer cell lines, compound 16 exhibited the most significant cytotoxicity toward most of the cell lines tested, with the lowest GI₅₀ value among the tested compounds. This compound contains an amino acid side chain and a BA group that could enhance its druglike properties. When compared with clinically used drugs, compound 16 achieved an IC₅₀ value of 3.43 μM, and was, therefore, 10- to 14-fold more toxic than chlorambucil (IC₅₀ = 48.7 μM) or melphalan (IC₅₀ = 34.44 μM). Finally, it was able to efficiently suppress tumor growth in an MDA-MB-468-derived xenograft mouse model without obvious side effects. These observations led to the belief that this compound is a promising drug candidate to be developed for cancer therapy.[60]

Quinone Methides

Quinone methides (QMs) are electrophilic transient Michael acceptors that possess high reactivity and are known to have various biological activities.[39] [61] [62] These species, which can occur naturally during various biological processes, might also be intermediates that form during the metabolism of drugs and xenobiotics and could be responsible for the activities of many antitumor drugs, DNA alkylators, insecticides, and antibiotics.[38,39] QMs can also be formed during the conversion of biologically inactive compounds through cellular processes such as enzymatic oxidation or reduction, or in the presence of high levels of ROS.[39] Due to their extremely elevated electrophilicity, QMs can react swiftly with various endogenous nucleophiles, such as GSH or DNA, leading to potentially cytotoxic effects.[62] Together with the fact that they can be generated in the presence of high levels of ROS, QMs present themselves as excellent candidates for the development of new ROS-responsive prodrugs.[38] Accordingly, this section focuses on the various studies that have been carried out with the aim of optimizing these compounds to improve their activity and expand their therapeutic applicability.

After developing an ROS-activated BA-nitrogen mustard prodrug in 2012, Cao et al. decided to explore the potential of the quinone methides that were formed in the process.[36] They reported three DNA bisalkylating and crosslinking agents 18–20 (Scheme [8]A) that, after H₂O₂-mediated activation of the boronate, generate an active biquinone methide.[36] [58]

On evaluating the activity of these compounds, no formation of ICLs was observed in the absence of H₂O₂, meaning that, under these conditions, there is no formation of a QM. However, upon addition of H₂O₂, an efficient formation of ICLs by compound 19 was observed, whereas the other two compounds failed to generate ICLs. The differences in the crosslinking efficiency of the three compounds can be explained by observing the structure of the corresponding phenol derivative formed after their reaction with H₂O₂. After treatment with H₂O₂, compound 19 gives rise to compound 19a, which has two hydroxy groups attached to the same aromatic ring, whereas compounds 18a and 20a possess only one hydroxy group attached to the aromatic ring. Accordingly, compound 19a possesses a more electron-rich aromatic ring, which favors QM formation and regeneration, resulting in the establishment of more efficient formation of DNA ICLs (Scheme [8]B).[36] [37] Additionally, at a concentration of 2 mM, compound 19 achieved a crosslinking yield of 24%, and this value was higher under basic conditions. Furthermore, compound 19 can be selectively activated by H₂O₂ over other ROS.[36]

Later, the same group performed additional studies in an attempt to understand how the aromatic substituents and the benzylic leaving groups affect not only the QM formation but also the generation of DNA ICLs. For this purpose, compounds 21–34 (Figure [7]) were designed.[37]

Regarding the oxidative cleavage step of the arylboronic esters, the researchers found that this is facilitated by electron-withdrawing groups, whereas the rate of the reaction decreases if the substituent on the aromatic ring is an electron-donating group. For example, compound 29, which possesses a strongly electron-withdrawing nitro group, showed about a 23-fold increase in the reaction rate compared with that of its parent compound 26. On the other hand, the presence of a methoxy group in compound 28 decreased the reaction rate about fourfold, and a methyl substituent in compound 27 slowed down the reaction rate twofold. The benzylic leaving group also influences the oxidative cleavage step, with quaternary ammonium salts (compounds 26–30) displaying 1.5–4.6 times faster rates than compounds 21–25, which have a bromide leaving group. On the other hand, QM generation is favored by electron-donating substituents, such as methyl or methoxy, whereas electron-withdrawing substituents strongly prevent this process. This process is also favored when a good leaving group, such as bromine, is present at the benzylic substituent position.[37] These results suggest that it is essential for QMs to have a precise balance between aromatic substituents that are electron donating and electron withdrawing, as these will increase the rates of oxidative cleavage step and QM generation, respectively.[37]

In 2014, the same researchers conducted another study that focused on the effects of leaving groups in H₂O₂-induced DNA crosslinking by arylboronates and, to this end, they used compounds 31–34 (Figure [7]). They found an efficient ICL formation with 2 mM of compounds 31a (25%), 32a (24%), and 33a (9%), which had bromides as leaving groups, whereas this did not occur with compounds 31b, 32b, and 33b with quaternary ammonia salts. However, for compounds 34a and 34b, an opposite effect was observed. Compound 34b, with a quaternary ammonia salt, exhibited a higher crosslinking yield (23%) than compound 34a (3.5%), and this might be related to the fact that the former compound is more soluble in water, which could influence the crosslinking efficiency. To test this, they prepared compound 34c, which contains both leaving groups, and it achieved a crosslinking yield of 33.7%, considerably higher than that of the other two compounds.[38]

However, despite showing promising ICL activity in biochemical assays, most arylboronate QM precursors reported up to that point displayed insufficient activity in cell-viability experiments. To address this issue, three years later, the same group developed a new generation of prodrugs. Starting from compound 31a, which had already proved its potential as an H₂O₂-activated anticancer prodrug, a series of new compounds were prepared featuring various electron-donating methoxy groups on the aromatic ring and neutral leaving groups (Figure [8]).[39]

In the absence of H₂O₂, most of the compounds did not induce DNA crosslinking, whereas in the presence of H₂O₂, an efficient ICL formation (20–50%) was indeed observed. The compounds bearing an acetate leaving group displayed lower crosslinking yields than the corresponding bromides. Compound 38a was unable to form DNA ICLs in the presence of H₂O₂. It was possible to highlight position 4 of the aromatic ring as a suitable site for further modification, as compounds 36a and 39a, which had an electron-donating group in this position, were shown to be the best H₂O₂-activated DNA crosslinking agents with crosslinking yields of 36 and 50%, respectively. This suggests that the addition of electron-donating groups at position 4 favors ICL formation.[39]

In vitro assays showed that compounds 35b, 36b, and 38a were not cytotoxic to the ovarian cancer SKOV3 cell line, whereas compounds 31a, 36a, and 39a displayed IC₅₀ values of 6.3 μM, 5.2 μM, and 3.8 μM, respectively. Additionally, compound 36a caused considerable growth inhibition in sixty cancer cell lines and was more toxic than compound 35a. In eight renal and breast cancer cell lines, compounds 31a, 36a, and 39a were found to be more effective than chlorambucil or melphalan. Finally, these three compounds were also tested in lymphocytes obtained from chronic lymphocytic leukemia (CLL) patients, and they all demonstrated potent cytotoxicity (IC_{50, 31a} = 48.3 μM, IC_{50, 36a} = 28.4 μM, and IC_{50, 39a} = 20.8 μM) in comparison to chlorambucil or melphalan (IC₅₀ = 98 μM and 76.9 μM, respectively); moreover, they displayed no toxicity toward normal lymphocytes.[39]

Selective Estrogen Receptor Modulators and Selective Estrogen Receptor Degraders

Estrogen receptors (ERs) have two main isoforms, ERα and ERβ, responsible for regulating various physiological processes. However, the dysregulation of ER signaling in certain tissues, such as breast, ovarian, and uterine tissues, can lead to the development of cancers in these organs. Although the role of the ERβ in cancers development is unclear, it is known that the ERα is involved in the initiation and progression of breast cancer.[63]

Of all diagnosed breast cancer cases, 75% are associated with ER+ breast cancer, and ER signaling has been found to be crucial in the development of the disease. As such, ER continues to be explored as an important pharmacological target in the fight against breast cancer. Several inhibitors of the ER pathway already exist, with ER antagonists being the most predominantly used. Compounds that directly antagonize ER can be of two types: selective estrogen receptor modulators (SERMs) and selective estrogen receptor degraders (SERDs). Whereas SERMs bind to estrogen receptors, preventing estrogen from binding there and thereby blocking its effect, SERDs bind to estrogen receptors and signal them for destruction through ubiquitination and proteasomal degradation.[63] [64] [65]

Tamoxifen, an FDA-approved SERM, is in fact a prodrug that is converted into the more-potent metabolites 4-hydroxytamoxifen (4-OHT) and endoxifen by specific enzymes. However, it appears that a significant percentage of patients have an intrinsic resistance to tamoxifen due to poor metabolization of the drug, which can result in shorter overall survival rates.[66] [67] To overcome these problems, Zheng, Wang, and co-workers took advantage of the oxidative conditions in breast cancer cells and created the boronated prodrugs of 4-OHT 40 and 41 (Figure [9]) and of endoxifen 42 to improve the bioavailability of both compounds.[66] [67] [68]

In cell-survival assays, prodrugs 40 and 41 proved to be effective and inhibited the growth of ER+ breast cancer cell lines (MCF-7 and T47D) in a similar way to 4-OHT. A concentration of 1 μM of 4-OHT inhibited MCF-7 cells growth by 63%, whereas for the same concentration of prodrugs 40 and 41, the corresponding values were 63 and 75%, respectively. Moreover, compound 40 showed an IC₅₀ value of 0.15 μM, which was about five times higher than that of 4-OHT (IC₅₀ = 0.029 μM). Compound 41 achieved an IC₅₀ value of 0.042 μM, which was similar to that of 4-OHT. Additionally, the prodrugs were shown to be selective as, at a concentration of 0.1 μM, they did not inhibit the growth of the ER-breast cancer cell line MDA-MB-231. 4-OHT also does not exhibit an obvious effect on the growth of these cells, having a survival ratio higher than 95%.[66] Further in vivo studies on compound 40 demonstrated an improved bioavailability, with a higher plasma concentration of 4-OHT and a better in vivo efficacy in comparison with the same studies on 4-OHT or tamoxifen.[67]

Regarding compound 42, the boronated derivative of endoxifen (Figure [9]), the replacement of the hydroxy group by a pinacol ester prevents the compound from being inactivated by O-glucuronidation during first-pass metabolism, resulting in improved oral bioavailability. Like 4-OHT and endoxifen, this prodrug is a potent SERM. A dose concentration of 0.1 μM led to 69% growth inhibition of MCF-7 cells, whereas, when incubated at the same concentration, 4-OHT and endoxifen inhibited the growth of MCF-7 cells by about 74% and 67%, respectively. The IC₅₀ concentration of 42 in MCF-7 cells was found to be 0.708 μM, which is comparable to that of endoxifen (0.675 μM), but nearly 40 times more potent than tamoxifen (25.77 μM). In both in vitro and in vivo assays in mice, the prodrug was efficiently converted into endoxifen by oxidative deboronation. Finally, the prodrug proved to be more effective than endoxifen in inhibiting tumor growth in mice and also displayed a plasma concentration 40 times higher than the parent drug.[68]

After validating their prodrug strategy with SERMs, the same research group expanded the methodology to design a prodrug of fulvestrant, an FDA-approved SERD. As with the compounds discussed above, fulvestrant also has a poor oral bioavailability, as it can be metabolized by O-glucuronidation and O-sulfation, leading to inactive metabolites. Because this represents a negative impact on its widespread clinical applications, the 3-OH group in fulvestrant was substituted with a boronic acid group, giving rise to prodrug 43 (Figure [10]). With an IC₅₀ value of 3.2 nM, the in vitro potency of the prodrug on the proliferation of breast cancer cells is comparable to that of 4-OHT (IC₅₀ = 3.3 nM) and fulvestrant (IC₅₀ = 1.5 nM). Furthermore, the prodrug effectively degrades the ER in a dose-dependent manner. In mice, compound 43 showed a better oral bioavailability than fulvestrant, and improved tumor growth inhibition.[65] Additionally, the designed compound displayed an improved oral pharmacokinetic profile, and fulvestrant was found to be its major oxidative metabolite.[69] In 2016, these researchers reported compound 44 (Figure [10]), which is the prodrug of the SERD candidate GW7604, and the conclusions taken from this work were similar to those for the boronic acid prodrug of fulvestrant.[70]

Nonetheless, it should be considered that, although the boronic acid prodrugs of SERMs and SERDs discussed in this section have shown promising results, the researchers have not evaluated the activation of these compounds by H₂O₂, emphasizing only their biological activity. Moreover, in prodrugs 42–44, the researchers do not mention selective delivery to cancer cells as an important advantage of these compounds, as they were more interested in the improvement of the bioavailability of the prodrugs compared with those of the parent drugs.

ROS Inducers

As previously mentioned, cancer cells are characterized by high levels of oxidative stress, and this high production of ROS plays an important role in tumor progression, proliferation, and metastasis. The adaptation of cancer cells to this high-oxidative-stress environment is done through activation of nonenzymatic and/or enzymatic antioxidizing systems.[71] [72]

In healthy cells, abnormal concentrations of ROS cause cellular damage in proteins, lipids, and DNA, leading to apoptotic cell death. Moreover, the overproduction of ROS can also be harmful to cancer cells, depending on the concentration and duration of the ROS stress. This is because, despite being adapted to high oxidative stress, cancer cells have low ROS tolerance and can be easily damaged by extra ROS derived from exogenous stimuli. In this way, stimulated overproduction of ROS can be used as a weapon to kill cancer cells.[71] [72] A new therapeutic strategy called oxidation therapy has emerged that uses chemotherapeutic agents to modulate the redox status in tumor cells, resulting in excessive accumulation of ROS and cell death. The accumulation of ROS by these drugs can be achieved either by an assisted overproduction of ROS directly in solid tumors or by preventing the elimination of ROS in cancer cells through the suppression of the antioxidant systems responsible for this function.[71–73]

However, although these drugs lead to the death of cancer cells, they also increase the concentration of ROS in normal cells. It is crucial to ensure selective and site-specific generation of ROS at the tumor site because high amounts of ROS in normal cells can lead to significant side effects, including the generation of secondary tumors.[73] [74] To circumvent this, several boronate-modified prodrugs that target cancer cells by amplification of oxidative stress after being activated by H₂O₂ have been described. This section covers prodrugs based on ferrocenes and copper chelators. There is also a reported approach of a dual stimulus-responsive prodrug activated by H₂O₂ and acidic pH to release glutathione-scavenging quinone methide and ROS-generating cinnamaldehyde, respectively.[71]

Ferrocene Derivatives

Between 2012 and 2022, Mokhir and co-workers developed several prodrugs based on aminoferrocenes.[4] ^, [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] After the reaction with H₂O₂, two synergistic cytotoxic products are formed, namely, a para-QM that alkylates GSH and inhibits the antioxidative system of cells, and a ferrocene fragment that is converted by H₂O₂ into redox-active iron-containing species [a ferrocenium derivative and iron(II) ions] that catalyze the generation of highly active ROS such as O^2– and HO^• from O₂ and H₂O₂, respectively (Scheme [9]). The para-QM and the ferrocene fragment act synergistically, leading to death of the cancer cells as a result of the severe elevation of ROS levels in the cells.[4] ^, [77] [78] [79]

Compounds 45a–e were prepared with various substituents on the aromatic ring and on the carbamate nitrogen group to evaluate their effects on cell-membrane permeability and reactivity toward H₂O₂ (Figure [11]). The presence of a high concentration of H₂O₂, caused the decomposition of compounds 45a–d, with the formation of Fe(II) and para-QM, whereas in the case of compound 45e, activation with H₂O₂ led to the formation of para-QM and ferrocenium ions. The latter compound exhibited the highest toxicity on the human promyelocytic leukemia HL-60 and human glioblastoma-astrocytoma U373 cell lines, with IC₅₀ values of 9 μM and 25 μM, respectively. The researchers also found that, after 48 hours, only 4% of the HL-60 cells and 0% of the U373 cells incubated with 100 μM of 45e remained viable. On incubating compound 45e with nonmalignant fibroblasts, 78% of the cells remained viable, and these cells were found to be resistant to the effect of the prodrug. Also, by using prodrug 45a as a reference, it was observed that compound 45b, featuring the polar fluorine substituent, is approximately three times less membrane permeable, whereas the less-polar compounds 45c, 45d, and 45e displayed higher membrane permeabilities (a 1.3- to 3.1-fold increase). Finally, the researchers verified that, upon activation, all prodrugs could be used as catalysts to convert H₂O₂ into HO^• in cancer cells.[4] Further studies on compound 45e concluded that it displayed no visible side effects and was able to extend the survival of mice carrying L1210 leukemia xenografts from 13.7 to 17.5 days with an increase in oxidative stress after administrating six daily doses of 26 μg/kg.[79]

In the following year, aiming to improve the properties of the prodrugs 45a and 45e (i.e., cell membrane permeability, suitability for in vivo applications, solubility in water, high toxicity to cancer cells, and cancer-cell specificity), the same group reported a new generation of prodrugs 46a, 46b, 47, 48a, and 48b. In the case of prodrug 47, it was found that the attachment of a carboxylic acid substituent onto the ferrocene moiety created a more water soluble but less cell-membrane permeable compound, which negatively impacted its toxicity (IC_50,47 = 68 μM versus IC_50,45a = 52 μM in HL-60 cells). In prodrugs 48a and 48b, where an alkyl substituent was introduced at the para-position of the N-benzyl fragment, no improvement in cell-membrane permeability was observed and ROS generation was also decreased compared with the original molecule 45e. On the other hand, the addition of a second arylboronic acid ester residue to the ferrocene core afforded more-potent compounds in HL-60 cells (IC_50,46a = 20 μM and IC_50,46b = 14 μM), due to increased membrane permeability. These two prodrugs were later tested in CLL cells, where they displayed IC₅₀ values of 1.4 and 1.8 μM, whereas an IC₅₀ value of 2.2 μM was observed for prodrug 45a. More importantly, these compounds hardly affected the tested healthy cells or representative bacterial cells.[77]

In 2015, the same researchers reported the activity of the aminoferrocene-based prodrugs 45a and 45e against prostate cancer, because the aggressiveness of this disease is related to the ability of cells to produce high amounts of ROS. At a concentration of 50 μM, compound 45a showed no toxicity toward both cell lines used (androgen-dependent LNCaP and androgen-independent DU-145 cells), whereas compound 45e exhibited a noteworthy cytotoxicity (IC₅₀ = 18 μM for DU-145 cells and IC_{50 =} 17 μM for LNCaP cells). These results can be explained by the fact that compound 45e possesses better cell-membrane permeability, which increases its accumulation in cancer cells.[78]

After establishing the importance of the benzyl group in the activity of these prodrugs, Mokhir’s research group developed a new class of prodrugs targeted for specific organelles, such as the lysosome (51; Figure [12]), mitochondria (52, Scheme [10]A), and endoplasmic reticulum of cancer cells (53 and 54; Figure [13]).[74] [80] [82] [83]

Prodrug 51 was developed from compound 45e through the introduction of a piperidine fragment. The piperidine residue is protonated in the acidic environment of lysosomes, enabling its accumulation there. This further favors the prodrug’s activation due to the high concentrations of ROS present on lysosomes, with particularly high fractions of HO^∙, one of the most reactive ROS. The new molecule showed higher toxicity than its parent compound in BL-2 and Jurkat cell lines (IC_50,51 = 3.5 μM versus IC_50,45e = 26 for BL-2 cells, and IC_50,51 = 7.2 μM versus IC_50,45e = 44 μM for Jurkat cells); it also displayed a good activity against CLL cells (IC₅₀ = 2.0 μM). Furthermore, low toxicity was achieved against noncancer dermal fibroblast adult HDFa cells (IC₅₀ = 30 μM) and normal mononuclear cells (IC₅₀ = 15 μM). It was also found that, at low concentrations, prodrug 51 induced cell death through apoptotic and necrotic mechanisms, whereas necrosis predominated at a high concentration.[80]

By containing an alkyl(triphenyl)phosphonium cation, compound 52 first accumulates in mitochondria; this is followed by its activation by mitochondrial ROS (Scheme [10]B). This prodrug was able to efficiently catalyze the conversion of H₂O₂ into HO^•, leading to IC₅₀ values in the range 5–10 μM in Burkitt lymphoma BL-2, human ovarian A2780, and human prostate DU-145 cell lines, after 48 hours of incubation. However, for the same incubation time, the compound showed moderate toxicity in healthy cells (human dermal fibroblasts HDFa cells, IC₅₀ = 18 μM); this value was similar to that observed in human immortalized T lymphocyte Jurkat cells (IC₅₀ = 20 μM). It was suggested by the researchers that mitochondria do not produce ROS in sufficient quantities to activate these prodrugs. This is probably explained by rapid scavenging by antioxidant systems of the ROS that are produced.[82]

Differently, compound 53, which targets the endoplasmic reticulum (ER) of cancer cells, promotes an increase in ER stress that leads to perturbations in the balance between folding/transport of proteins and degradation of misfolded proteins, culminating in cell death. Despite achieving cancer-cell selectivity, side effects are also expected, which contribute to genome instability and stimulate carcinogenesis. This prodrug, which presented the ER as the primary site of action, exhibited a strong anticancer effect on Burkitt’s lymphoma BL-2 (IC₅₀ = 9 μM), ovarian cancer A2780 (IC₅₀ =5.4 μM), and T-cell leukemia Jurkat (IC₅₀ = 18 μM) cell lines, with a significant increase in total intracellular ROS. Additionally, higher toxicity towards CLL cells (IC₅₀ = 5.6 μM) was observed compared with normal mononuclear cells (IC₅₀ = 18.3 μM). In addition to prodrug 53, the group also designed its fluorogenic analogue 54, which displays similar solubility, lipophilicity, and anticancer activity toward A2780 cells. In addition to being activated in cancer cells (A2780), but not in healthy cells (SBLF9 fibroblasts), 54 was also shown to be efficiently activated in vivo in the NK/Ly murine model.[83]

Platinum-based prodrugs with aminoferrocene units were also developed by this group. The first one prepared was compound 55 (Figure [14]), aimed at improving the potency of carboplatin, a Pt(II)-based anticancer drug.[74] Following their uptake by cancer cells, these types of agents undergo an aquation process to form [L₂Pt(OH₂)₂]²⁺-type complexes, which are active and are able to alkylate the N7-atoms in purine nucleobases of DNA, leading to initiation of cell death.[75] After 24 hours of incubation, the cytotoxicity of the prodrug (IC₅₀ = 35 μM) was superior to that of the parent drug (IC₅₀ > 200 μM) in the human ovarian carcinoma A2780 cell line. This improvement in activity can be explained by the synergistic effect observed between the Pt and ferrocene structures. Furthermore, accumulation in the mitochondria of A2780 cells was 5.7-fold more efficient than that of the known drug.[74]

However, Pt(II) drugs are known for their inadequate cellular uptake, which results in cancer resistance to these drugs. To overcome this problem, Pt(IV) prodrugs have emerged, as these complexes are stable in the extracellular environment and are reduced by ascorbate, GSH, and macromolecular reductants after uptake by cancer cells, to produce Pt(II) drugs.[75] Prodrugs 56 and 57 (Scheme [11]), which are Pt(IV)-based prodrugs with aminoferrocene groups, were also designed by Mokhir’s research group.[75] [76]

Prodrug 56 consists of two N-alkylaminoferrocene moieties covalently linked to a Pt(IV) complex. Upon oxidation by ROS, intermediate 56a is formed; this possesses two N-alkylaminoferrocene fragments that are redox active. Given the proximity of these two residues to the Pt(IV) center, two fast and efficient e^– transfers occur between them to generate cisplatin, a widely used anticancer drug. When tested in ovarian carcinoma A2780 cells, which are sensitive to cisplatin, the prodrug showed similar cytotoxicity to cisplatin (IC_50,56 = 2.5 μM versus IC_50,cisplatin = 2.1 μM), which might indicate that Pt(IV) reduction to Pt(II) occurred in these tumor cells.[75] However, the lower cytotoxic activity observed when compared with the parent molecule suggests that there was an incomplete intramolecular activation of the prodrug.[76] The activity of the prodrug was also evaluated against the cisplatin-resistant A2780cis cell line, whose resistance is due to insufficient drug uptake, and it displayed a twofold increase in activity (IC_50,56 = 6 μM versus IC_50,cisplatin = 13 μM). An important aspect of this compound is that, despite being active in these two cancer cell lines, it was shown to be nontoxic to normal fibroblast HDFa cells (IC₅₀ > 25 μM). Furthermore, it was verified that the prodrug is a strong ROS inducer, which also contributes to the observed antiproliferative properties.[75]

Compound 57 was designed in an attempt to improve the previous prodrug system and to take advantage of its full potential. Thus, a methylene group was removed from the linker that connected the aminoferrocene residues to the Pt(IV) center to facilitate the intramolecular e^– transfer step during prodrug activation, and cisplatin was replaced by oxaliplatin. Oxaliplatin is a clinically used anticancer agent that is safer than cisplatin and which shows a high cytotoxicity in cisplatin-resistant cells. This compound displayed an IC₅₀ value of 0.4 μM in A2780 cells, which is approximately six times lower than that of compound 56 and approximately half the value presented by oxaliplatin (IC₅₀ = 0.9 μM). These results show a more efficient activation of compound 57 in cells when compared with compound 56. Furthermore, similar activity was observed in A2780cis cells (IC_50,57 = 0.7 μM versus IC_{50,oxaliplatin} = 3 μM), and it remained less toxic to nonmalignant HDFa cells (IC_50,57 = 18 μM versus IC_{50,oxaliplatin} = 19 μM).[76]

Despite these promising results, aminoferrocenes display considerable chemical instability under oxidative conditions, with half-lives frequently shorter than one hour, which limits their anticancer activity. To address this limitation, a new class of compounds 58a and 58b was reported in which the amine group is replaced by an aniline (Figure [15]). In contrast to the aminoferrocenes, these compounds are stable for longer than six hours in aqueous buffers under aerobic conditions, while still being able to donate an electron to H₂O₂ to form HO^•. Compound 58a was found to be more cytotoxic than its aminoferrocene analogue 45a in human ovarian cancer A2780 cell line (IC_50,58a = 13.8 μM versus IC_50,45a = 32.2 μM). A similar result was observed for 58b (IC_50,58b = 8.3 μM versus IC_50,50a = 14.6 μM). The ROS-generating ability of these prodrugs was also higher than that of their predecessor analogues.[84]

Copper Chelators

Disulfiram (DSF) is an aldehyde dehydrogenase inhibitor (ALDHi) that, despite being FDA approved for the treatment of alcoholism, has good in vitro anticancer activity. Its metabolite, diethyldithiocarbamate (DTC), can chelate Cu(II) to form the complex Cu(DTC)₂, which induces an increase in intracellular ROS levels. However, because DSF is unstable in the acidic gastric environment and in the bloodstream, it has limited clinical applications.[85]

To address these limitations, Pan et al. developed prodrug 59 (Scheme [12]), which possesses an arylboronic ester as a H₂O₂-sensitive moiety.[85] Upon activation of the prodrug, QM and DTC are released, permitting simultaneous GSH depletion and ROS generation, respectively, to amplify oxidative stress and kill cancer cells through apoptosis (Scheme [12]).[85]

First, the designed prodrug presented better stability in the blood than DSF and, when tested in vitro, was less toxic to nontumorigenic NIH 3T3 cells (IC₅₀ > 100 μM) than DSF (IC₅₀ = 12.5 μM). In the cancer cell line tested (4T1), both compounds alone displayed low cytotoxicity, but the presence of 1 μM of extracellular Cu(II) allowed the prodrug and DSF to achieve IC₅₀ values of 1.4 μM and 0.33 μM, respectively. By adding exogenous H₂O₂ along with extracellular Cu(II), it was possible to enhance the IC₅₀ of the prodrug to 0.80 μM, but the cytotoxicity of DSF was not affected. Similar to DSF (1 μM), the prodrug (2 μM) induced apoptosis in approximately 60% of cells. In contrast, Cu(DTC)₂ showed negligible cell apoptosis, indicating that its anticancer activity was expressed in a nonapoptotic manner.[85]

In 2020, Bao et al. designed the prodrug 60 of naphthazarin (Nap) by masking its two phenolic hydroxy groups with a boronate group (Scheme [13]).[86] Nap is a natural product with anticancer activity, and its toxicity is due to its two keto–enol moieties that can act as a potential Cu(II) ionophore.[86] After activation with H₂O₂, Nap and QM are released. Both these products are electrophilic species capable of depleting intracellular GSH through a rapid alkylation reaction. Furthermore, the Nap–GSH adduct formed can function as a Cu(II) ionophore and will chelate Cu(II) to form compound 60a. In turn, compound 60a is reduced by GSH, leading simultaneously to the release of Cu(I) species and the induction of ROS amplification, causing a redox imbalance in cancer cells.[86]

In vivo studies on human hepatoma HepG2 and human normal liver L02 cells demonstrated a synergistic effect between the prodrug and Cu(II) in killing the HepG2 cells whereas this had no effect on the L02 cells. This can also be concluded from the combination indices, which were 0.003 and 0.474 for HepG2 and L02 cells, respectively, for 10 μM compound 60 and 10 μM Cu(II). When used separately, the prodrug and Cu(II) showed no toxicity in the cell lines studied. Furthermore, the prodrug was more selective for cancer cells than was the parent drug.[86]

Prodrugs Based on Other Types of Anticancer Drugs

Besides the aforementioned examples, BAs and BEs have also found reasonable success in the design of prodrugs of other anticancer agents currently used in the clinic. This strategy permits an improvement in the pharmacokinetic and pharmacodynamic properties of these cytotoxins and addresses their poor selectivity toward tumor cells, which often compromises their therapeutic efficacy. This section discusses the reported boron-based prodrugs of camptothecin derivatives, 5-fluorouracil, gemcitabine, doxorubicin, and crizotinib.

Camptothecin Derivatives

Camptothecin (CPT) is an alkaloid obtained from the Chinese tree Camptotheca acuminata, isolated for the first time in 1966 by Wall and Wani and their co-workers.[87] [88] [89] This natural product showed high antitumor and antibiotic activities but presented toxicity problems and poor solubility in clinical trials. Later, it was found that the antitumor activity of CPT could be increased by introducing a hydroxy group at position 10, and this gave rise to compounds such as topotecan, 10-hydroxycamptothecin, and 7-ethyl-10-hydroxycamptothecin (SN-38). CPT and its derivatives express their antitumor activity through the inhibition of the enzyme topoisomerase I, which is involved in DNA replication and reassembly. By binding to topoisomerase I, these agents stabilize a covalent DNA–topoisomerase complex, thereby disturbing the catalytic cycle of topoisomerase I.[87,88]

In 2016, Wang et al. reported compound 61, a CPT substituted with a boronic acid at position 10 as a prodrug of SN-38 (Figure [16]), which is the active metabolite of irinotecan.[88] Irinotecan is a drug approved by the FDA in 1994 for the treatment of colorectal cancer that is converted into the active SN-38 by the action of liver enzymes.[87] [88]

In this work, they verified that prodrug 61 can be activated selectively by H₂O₂ over other ROS to form SN-38 with a conversion of about 50% in MCF-7 cells. The prodrug was shown to be equally potent to or more potent than SN-38 in three cell lines tested, and it displayed an improved inhibitory activity. This last observation suggests that, besides being a prodrug of SN-38, compound 61 is also a typical topoisomerase I inhibitor. Finally, prodrug 61 also displayed significant antitumor activity in in vivo assays.[88]

5-Fluorouracil

5-Fluorouracil (5FU) is an FDA-approved anticancer agent that is currently used to treat various solid tumors, such as breast cancer, colorectal cancer, stomach cancer, pancreatic cancer, cervical cancer, and skin cancer.[90] The mechanism of action of 5FU involves not only the inhibition of cellular thymidylate synthase, which prevents DNA replication, but also the inhibition of RNA synthesis by the integration of its metabolites into RNA.[90] [91]

Despite its benefits, 5FU has important side effects and is metabolically unstable. To overcome these limitations, in 2019, Ai et al. designed and evaluated two arylboronate-based prodrugs of 5FU 62a and 62b (Figure [17]).[90] In these prodrugs, a para-boronatobenzyl group was introduced at the N1 position of the 5FU to minimize metabolic degradation. In the presence of H₂O₂, both prodrugs were found to be activated efficiently to release the active drug. Compound 62a demonstrated strong cytotoxic effects at a concentration of 50 μM, with growth inhibition in the range 81–96% in six of the sixty cell lines tested. However, compound 62b was only able to attain 22–65% inhibition in the same cell lines. In in vivo assays, compound 62b was again weaker than compound 62a, which, in turn, showed a lower antiproliferative activity in comparison to 5FU. Furthermore, compared with 5FU, the prodrugs showed no toxicity toward healthy cells, with a cell viability higher than 95% at concentrations of up to 100 μM, and prodrug 62a exhibited once again a more favorable safety profile than the parent drug.[90]

Gemcitabine

Gemcitabine (GEM) is currently the first line of treatment in pancreatic cancer chemotherapy.[92] Its cytotoxic effect is induced after GEM phosphorylation inside cells by deoxycytidine kinase, giving GEM 5′-diphosphate and GEM 5′-triphosphate, which are incorporated into DNA by DNA polymerases.[92] [93] The incorporation of such compounds into DNA prevents new nucleosides from being attached to continue DNA strand synthesis, resulting in apoptosis of cancer cells.[92–94]

However, like other cytotoxic drugs, GEM has several side effects, including nausea, vomiting, fever, and myelosuppression, the last of which is the main dose-limiting toxicity, frequently leading to discontinuation of therapy. Thus, to ensure the selective and specific delivery of this agent into cancer cells, Matsushita et al. developed the H₂O₂-activatable GEM prodrug 63 (Figure [18]).[92]

Among the tested ROS, it was observed that compound 63 reacted selectively with H₂O₂ to generate the active compound. Moreover, the prodrug was more cytotoxic in the human pancreatic cancer cell lines PSN1 and BxPC3 than in the normal pancreatic epithelial cell line, which has a lower concentration of H₂O₂. These results confirm that the cytotoxicity of the prodrug was due to a tumor-specific activation by H₂O₂. Furthermore, in vivo assays in an immunodeficient xenograft mouse model confirmed the prodrug’s ability to induce apoptosis with a similar potency to the parent drug. These assays also established that the myelosuppression caused by compound 63 was less severe than that of GEM. The results obtained from this study show that the modification performed in GEM provided a safer cytotoxic compound with better selectivity toward cancer cells.[92]

Doxorubicin

Doxorubicin (DOX) is produced by cultures of Streptomyces peucetius var. caesius and is a very effective and widely used antitumor drug.[95] [96] Because of its planar structure, this molecule intercalates into DNA, inhibiting the enzyme topoisomerase II. This enzyme is responsible for the cleavage and resealing of double-stranded DNA during its replication, so its inhibition will prevent cancer-cell growth. However, despite being one of the most successful anticancer agents, DOX presents several important side effects that limit its therapeutic utility, including nausea, vomiting, bone-marrow suppression, hair loss, and strong cardiotoxicity.[96] To address these limitations, the three ROS-activatable DOX prodrugs 64a–c (Figure [19]) were developed.[95]

The three compounds incorporate a phenylboronate (64a), its fluorinated benzene homologue (64b) (because electron-withdrawing atoms increase the rate of oxidation), and a furan ring (64c), which is an effective self-immolative spacer. Although the activity of the three compounds was comparable in a variety of cancer cell lines, the prodrug 64a proved to be the most active in the pancreatic cancer MiPaCa-2 cell line, giving an IC₅₀ value similar to that of DOX (IC_50,64a = 0.3 μM versus IC_50,DOX = 0.2 μM). In vivo studies demonstrated that 64a induced tumor regression in a similar way to DOX (~50% regression).[95]

Crizotinib

Crizotinib is a tyrosine kinase inhibitor approved by the FDA for the treatment of anaplastic lymphoma kinase-positive non-small-cell lung cancer and ROS1-positive non-small-cell lung cancer. Tyrosine kinase inhibitors compete with ATP at the active sites of tyrosine kinases, most often reversibly and competitively. Tyrosine kinases are involved in multiple cellular processes, such as differentiation, proliferation, and apoptosis, and are important therapeutic targets due to their abnormal activation in cancer cells. However, because these proteins are also expressed in healthy cells, crizotinib-treated patients often present side effects, including hepatotoxicity and neutropenia, which frequently demand a dose reduction or discontinuation of the treatment. This motivated the development of compound 65 (Figure [20]) in 2020 by Bielec et al. as an ROS-activated prodrug of crizotinib. Since the 2-aminopyridine group is essential to its binding to tyrosine kinase, the introduction of the boronic acid trigger at this position masks its cytotoxic activity.[97]

Docking studies revealed that the designed compound is not able to bind as strongly as crizotinib to anaplastic lymphoma kinase and c-mesenchymal-epithelial transition factor proteins, and prodrug 65 demonstrated a significant decrease in the inhibition of these proteins in comparison with the parent drug. The prodrug was activated in the presence of H₂O₂ to release the active drug, but this activation only occurred in cells with high intracellular ROS levels (non-small-cell lung carcinoma H1993 cells). It was also in these cells that 65 was most active, suggesting that there is a correlation between the activity of this compound and intracellular H₂O₂ levels in cells. Finally, the prodrug was much less toxic than crizotinib in nonmalignant human lung fibroblast HLF cells that expressed the tyrosine kinase c-mesenchymal-epithelial transition factor.[97]

Other ROS-Responsive Moieties

Despite their widespread use in the design of ROS-responsive prodrugs, BAs and their esters display an important number of limitations, including low metabolic stability and promiscuous reactivity with endogenous diols.[98] [99] This nonspecific reactivity often leads to poor pharmacokinetic profiles and the appearance of off-target toxicity. The discovery of new alternatives to BAs in the field of ROS-responsive prodrugs is therefore important. Accordingly, this section briefly discusses some recent examples of alternative ROS-responsive prodrug systems including thiazolidinones, selenium ethers, 1,3-oxathiolanes, sulfones, thioketals, and thioethers.

Thiazolidinones

Thiazolidinones are derivatives of thiazolidine that consist of a five-member ring with a sulfur atom in position 1, a nitrogen atom in position 3, and a carbonyl group in one of the remaining three positions of the heterocyclic ring (Figure [21]). These various derivatives are biologically important, as they are associated with several pharmacological properties. For example, 1,3-thiazolidin-2-ones have been shown to inhibit the BRD4 bromodomain, and the 1,3-thiazolidin-4-one unit is known to exhibit antitubercular, antimicrobial, antiinflammatory, antiviral, antioxidant, and antidiabetic activities.[100] [101]

Moreover, these molecules, particularly 1,3-thiazolidin-2-ones, have been reported to be sensitive to ROS and can, therefore, be used as warheads to elaborate H₂O₂-activated prodrugs.[102] [103] [104] Thiazolidinone-protecting groups are useful because they enable the masking of carboxylic acids in drugs that exhibit toxicity in healthy cells and lack the desired druglike properties. This is an important advantage of this system over BAs as, until now, BAs have never been reported to mask a drug featuring a carboxylic acid. The drug release occurs after hydrolysis of the promoiety in the presence of a high concentration of H₂O₂ (Scheme [14]).[102]

This strategy was employed by Perez et al. to construct a prodrug based on matrix metalloproteinase (MMP) inhibitors (MMPi).[102] MMPs are a group of matrix Zn(II)-dependent endopeptidases that are involved in the cleavage of proteins in the extracellular matrix, and are overexpressed and misregulated in several pathologic disorders, including arthritis, cancer, and cardiovascular disease. Inhibitors of this enzyme family exist, but they are not selective, being unable to distinguish MMPs involved in disease progression from those that have a normal physiological function. As a result, MMPi frequently displays off-target side effects and would benefit from a targeted prodrug strategy.[105] A designed prodrug 66 (Figure [22]) did not show significant activity against MMPs without external activation. Its inhibitory activity was, however, restored in the presence of H₂O₂. The prodrug with the thiazolidinone group was stable as it was degraded by less than 5% in buffer (Tris-Cl; pH 7.4) or in the presence of biologically relevant nucleophiles (serine, lysine, and glutathione).[102] However, further in vitro and in vivo studies are necessary to support the suitability of this prodrug.

In 2018, a similar strategy was used by Andersen et al. to develop 67, a prodrug of methotrexate (Figure [22]).[103] Methotrexate is a drug used in the treatment of rheumatoid arthritis, but has shown some important adverse effects. The prodrug was designed to localize and accumulate methotrexate in inflammatory tissues, to improve its safety profile and efficacy. In vivo studies in a murine collagen-induced arthritis model revealed that compound 67 displays a similar efficacy to the parent drug and effectively reduces the severity of arthritis with a safer toxicity profile.[103]

Note that although these two strategies were applied to antiinflammatory agents and not to anticancer agents, which are the focus of this work: they are mentioned here to demonstrate that thiazolidines are plausible scaffolds for the design of tumor-targeting ROS-activated prodrugs, particularly for drugs featuring carboxylic acid groups.

1,3-Oxathiolanes

The 1,3-oxathiolane group is commonly used in organic chemistry as an aldehyde-protecting group that can be deprotected with hypochlorous acid (HOCl). Additionally, the aldehyde can also be deprotected in the presence of H₂O₂, which means that 1,3-oxathiolanes can be used in the design of ROS-responsive prodrugs.[106]

Won et al. proposed the aldehyde dehydrogenase (ALDH)-targeting prodrug 68 (Scheme [15]), which is active against both tumor cells and cancer stem cells (CSCs). CSCs have an important role in tumor formation, metastasis, and recurrence, and even a small population of CSCs can lead to the onset of cancer. However, CSCs possess a multidrug resistance property, which allows them to survive even after the treatment with an anticancer agent. It has already been reported that the high activity of ALDH is related to the resistance of CSCs to chemotherapeutic drugs, so a simultaneous treatment with an ALDH inhibitor (ALDHi) is recommended. For this reason, besides having a masked 4-(diethylamino)benzaldehyde (DEAB) as an ALDHi, compound 68 also incorporates a camptothecin moiety. In this dual-action prodrug, the 1,3-oxathiolane group acts both as an ROS-responsive self-immolative unit and as an aldehyde-protecting group of DEAB.[106]

The researchers found that the prodrug activation can be triggered by both H₂O₂ and HOCl, with subsequent release of DEAB and CPT. Furthermore, a comparison of the cell viability in the presence of 68 with that in the presence of CPT revealed that the designed compound has a higher selectivity toward cancer cells. For example, in the healthy cell line (human breast epithelial MCF10A cells) tested, a concentration of 30 μM of the prodrug provided a cell viability of approximately 80%, whereas in the case of CPT, this value was less than 40%. In the breast cancer cell lines studied (BT474 and MDA-MB-231 cells), the cell viability in the presence of each compound was similar. Finally, when compared with CPT, this prodrug features an improved induction of apoptosis of cancer cells through the simultaneous release of DEAB and CPT.[106]

Selenium Ethers

Selenium is an essential and unique element that plays an important role in both health and disease.[107] It is mostly present in plant-based foods but can also be found in some meats and seafood. When supplied at low nutritional levels, selenium can function as an antioxidant, but it becomes a prooxidant when present at higher doses, exhibiting high cytotoxic activities.[107] [108] As a result, several selenium-containing compounds have shown promise as cancer chemopreventive and chemotherapeutic agents.[108] Moreover, this atom is also able to react with ROS, which allows an extension of its utility for the construction of prodrugs.[109]

In 2018, Pan et al. reported the selenium-containing prodrug 69 (Scheme [16]), which releases carbon monoxide (CO) into cells displaying high levels of oxidative stress upon activation with H₂O₂.[109] This is of great significance, given that CO has been successfully identified as a potential therapeutic agent against a variety of human diseases. In cancer treatment, this molecule is capable of sensitizing cancer cells to anticancer agents while sparing healthy cells. In the designed prodrug, the phenylselenyl group is susceptible to oxidation by ROS. When this occurs, a selenoxide analogue is formed and rapidly undergoes a syn-elimination that generates a double bond between the C5 and C6 positions. Finally, a cheletropic reaction occurs that allows the liberation of CO under very mild conditions (Scheme [16]).[109]

In vitro, the prodrug proved to be stable in aqueous solution and it was activated by hypochlorite (ClO^–), singlet oxygen (¹O₂), or O₂ ^•–, among which ClO^– was the most reactive trigger. This compound also selectively delivered CO to cervical carcinoma HeLa and macrophage-like Raw264.7 cell lines, which are characterized by elevated ROS levels. The prodrug also sensitized cancer cells to DOX. When used alone, the anticancer agent presented a cell viability of around 80%, but this value decreased to around 20% when DOX was used in combination with the prodrug. In addition, compound 69 showed no toxicity in healthy cells.[109]

Four years later, Yang and co-workers described a new family of H₂O₂-inducible theranostic prodrugs 70a–g (Figure [23]) that contain a CPT moiety, a self-cleavable linker, and an allyl phenyl selenide moiety as an H₂O₂-responsive trigger. In these theranostic prodrugs, CPT exhibits a strong fluorescence which permits its use as both a drug and an optical reporter.[98]

Treatment with H₂O₂ induced activation of the compounds by the mechanism shown in Scheme [17]. Taking compound 70a as an example, the Se atom is oxidized by H₂O₂ to give transition state a. The allylic selenoxide b undergoes a Mislow–Evans rearrangement to form the selenenate c, which, in turn, gives the hemiacetal d through nucleophilic cleavage of the Se–O bond. Subsequently, hydrolysis of hemiacetal d releases acrolein and phenoxide e. Finally, a 1,6-elimination reaction of phenoxide e releases active strongly fluorescent CPT. The fluorescence generated after ROS-mediated oxidation permits the detection/quantification of H₂O₂ in solution and monitoring of CPT release.[98]

The designed compounds were less toxic than CPT for normal cells. Among them, 70a stood out as the most potent in the cancer cell lines used (cervical carcinoma HeLa, human alveolar cell carcinoma A549, and human colon cancer HCT116 cells), with IC₅₀ values lower than 0.31 μM, compared with 0.11 μM observed for CPT. Furthermore, compound 70a was able to induce a selective apoptosis in cells with high levels of ROS. The differences observed in cytotoxic activity between the prodrugs might have been due to differences in their membrane permeability. There is also a synergistic effect between the released acrolein and CPT, which is responsible for the antiproliferative activity of the prodrugs. This was concluded after it was observed that acrolein exhibited an IC₅₀ value of 5 mM for the HCT116 cancer cell line. The cytotoxicity of compound 70a is similar to the arylboronate ester-based prodrug of CPT, but the former exhibits a greater stability in human plasma and a more-complete release of CPT. These results validate the use of selenium ethers as a plausible strategy for the development of theranostic prodrugs activated by ROS.[98]

Sulfur-Containing ROS-Responsive Moieties

Sulfur is a nontoxic, biocompatible, essential element in biological systems and is a macronutrient necessary for the growth and development of living organisms. Consequently, it is found in several important metabolites for the maintenance of cell structure and biological activities, and is also part of various important proteins.[110] [111]

The sulfur atom is able to share two electrons with other atoms and can display oxidation states of +6, +4, +2, +1 and –2. For this reason, together with its low toxicity, ROS-responsive prodrugs containing sulfur groups have already been developed, including sulfonate esters, and thioketals, which are discussed below.[20]

Sulfonate Esters

Sulfonate esters are commonly prepared through the reaction of alcohols with sulfonyl halides and they have already been reported as protecting groups for alcohols present in fluorescent probes that, when exposed to certain ROS, release a sulfonic acid and a fluorescent dye (Scheme [18]A).[33] ^, [112] [113] [114]

In 2011, Daniel et al. used sulfonate esters in the design of the prodrugs 71a and 71b (Scheme [18]B) based on 1-hydroxypyridin-2-one (1,2-HOPO-2).[105] 1,2-HOPO-2 is an MMPi with IC₅₀ values under 100 nM and a inhibition percentage of MMP-12 of about 55% at 50 nM. To improve the selectivity of this inhibitor, an ROS-responsive warhead was attached to the hydroxylamine group, which is important in the chelation to the catalytic Zn(II) ion of MMP. When tested in the presence of H₂O₂, prodrugs 71a and 71b were activated to the parent drug, leading to an increase in the inhibition of MMP-12 and, at a concentration of 50 nM, the percentage inhibitions were approximately 30 and 50%, respectively. However, the researchers noticed that both prodrugs were completely hydrolyzed after 24 hours, which proved to be a significant drawback of this methodology. One way to overcome this problem and to produce more-stable prodrugs might be through the combination of sulfonate esters with self-immolative spacers.[105]

Thioketals

Thioketals (TKs) are the sulfur analogues of ketals, and can be synthesized by the condensation of thiols with ketones. The thioketal functional group is advantageous because it is stable in the presence of enzymes and under both acidic and basic conditions.[115] However, when exposed to pathological levels of ROS, it is quickly cleaved to form a thiol and a ketone, and these products are nontoxic.[115] Although the mechanism of oxidation is not yet well understood, El-Mohtadi et al. reported that it is commonly accepted that these are the final products formed after ROS attack, with the generation of hemithioketal intermediates (Scheme [19]A).[116] Nevertheless, in 2020, Liu and Thayumanavan proposed a new mechanism for the oxidative cleavage of TKs.[117] In this proposed mechanism, one of the thioether groups is oxidized to the corresponding sulfoxide, and this reaction promotes the cleavage of one C–S bond to form a sulfenic acid and a thiocarbenium intermediate. Subsequently, the hydrolysis of the thiocarbenium intermediate affords the ketone and the free thiol. Disulfide products might also be obtained, because thiols react rapidly with sulfenic acids (Scheme [19]B).

In 2016, Liu et al. engineered the multifunctional prodrug 72 (Scheme [20]), which is activated by red light and enables an image-guided treatment of cancer using photodynamic therapy (PDT) and cascaded chemotherapy.[118]

PDT is a noninvasive method for cancer therapy that uses the ROS formed after light irradiation of a photosensitizer to irreversibly destroy cancer cells.[115] [118] Liu and co-workers conjugated GEM with a fluorescent photosensitizer (meso-tetraphenylporphyrin; TPP) through an ROS-oxidizable TK linker. TPP is an efficient ¹O₂ generator, so its irradiation with red light permits the induction of cell damage through PDT. Subsequently, the generated ¹O₂ cleaves the TK linkage in the prodrug, leading to cascaded release of gemcitabine, which produces additional damage to cells.[118]

Compound 72 performed well in both in vivo and in vitro studies. After light irradiation, cervical carcinoma HeLa cells incubated with 2 μM of the prodrug showed an apoptotic and necrotic percentage of about 73.8%, a value much higher than that obtained with GEM (<7%). Also, significant suppression of tumor growth without apparent bodyweight loss in mice treated with this prodrug under light irradiation was reported, suggesting its outstanding capacity for tumor inhibition. These and other results have shown that this prodrug permits noninvasive in situ drug tracking and release after red-light irradiation.[118]

By taking advantage of their reactivity, TKs groups have also been used as responsive linkers for the development of selective and upgraded drug-delivery systems, most often through their incorporation in polymer chains. Although those supramolecular ROS-responsive structures are effective in selectively delivering drugs to tumor cells, they cannot be considered actual prodrugs with defined structures and, therefore are beyond the scope of this Account, and have been reviewed elsewhere.[115] However, we decided to include one example of a TK-containing delivery system, to showcase the potentialities of these linkers.

In 2021, Ma et al. built an ROS-responsive dimeric prodrug-based nanosystem 73 for the treatment of gastric cancer (Scheme [21]). In this prodrug, two ursolic acid (UA) molecules were connected through a thioketal linker attached to the OH group at position 3 of each UA molecule. This product (TK-UA₂) can self-assemble into stable nanoparticles in the presence of a surfactant [1,2-distearoyl-sn-glycero-3-phosphoethanolamine–poly(ethylene glycol); DSPE-PEG] and subsequently undergo rapid and selective conversion into the active drug in the presence of ROS (Scheme [21]).[119]

UA is a natural product possessing several pharmacological properties, including antitumor activity, and can effectively prevent the progression of various cancers. However, its clinical application is limited due to its low aqueous solubility, short blood-circulation times, and low bioavailability. Hence it is important to develop new UA-based drug-delivery methodologies.[119]

This delivery system, in addition to the dimeric prodrug inner core, comprises a poly(ethylene glycol) (PEG) shell to improve colloid stability and prolong blood-circulation time, and a surface-modified internalizing RGD peptide to improve tumor targeting. RGD is a peptide that specifically recognizes and binds to α_vβ₃ integrins. These transmembrane proteins are overexpressed in various tumor cells but are expressed at low levels in normal cells, which allows a specific and selective internalization of nanoparticles (NPs) by cancer cells.[119]

The study showed that this nanosystem improved the solubility and extended the blood-circulation time of UA. Furthermore, the sensitivity of the thioketal present in the structure of prodrug 76 was also confirmed; this releases the UA molecules in the presence of ROS through the mechanism shown in Scheme [21]. The nanoparticles (NPs) showed a higher IC₅₀ value than UA in normal NIH-3T3 cells (IC_50,NPs > 50 μg/mL versus IC_50,UA = 18.5 μg/mL), and were more toxic than UA to human gastric SGC 7901 cancer cells (IC_50,NPs = 7.2 μg/mL versus IC_50,UA = 10.3 μg/mL). In mice, whereas UA achieves a tumor suppression ratio of 45.0%, the NPs were able to reach 86.5%. Furthermore, mice treated with the NPs showed no significant decrease in their body weight, which confirmed the safety of this methodology.[119]

Thioethers

Thioethers are sulfur-containing analogues of ethers that can undergo conversion into sulfoxides and sulfones under mild oxidative conditions and at low temperatures. This ROS-triggered hydrophobic-to-hydrophilic transition has already been exploited in the design of targeted drug-delivery systems, given the sensitivity of this group to low concentrations of ROS (100 μM of H₂O₂).[20] [110] [120]

In 2018, Yang et al. synthesized a paclitaxel (PTX) prodrug 74 (Scheme [22]A) with a thioether oxidative-sensitive linker.[121] This compound contains a maleimide group that rapidly conjugates to albumin in vivo. Human serum albumin (HSA) is an abundant serum protein that can work as a transport vehicle to improve the prodrug’s accumulation in a tumor. Accordingly, after intravenous administration, the prodrug binds to the thiol group of Cys-34 in HSA and concentrates in solid tumors because of the passive targeting and the enhanced permeation and retention effect (Scheme [22]B).

Once at the tumor site, the thioether is oxidized to a hydrophilic sulfone, and hydrolysis of the ester function is subsequently observed to release an active PTX molecule.[110] [122]

At an H₂O₂ concentration of 10 mM (the ROS concentration in cancer cells), about 40% of PTX was released from the prodrug, whereas only 10% of the active drug was released at an H₂O₂ concentration of 10 μM (the ROS concentration in healthy cells) or in phosphate-buffered saline. The thioether linker used enhanced the accumulation of PTX in tumors by facilitating the hydrolysis that occurs after oxidation. Despite the decreased in vitro cytotoxicity of 74 compared with PTX, the conjugation of the maleimide group with albumin improved the biodistribution and tumor accumulation of PTX. These prodrugs exhibited potent antitumor activity in 4T1 breast cancer-bearing BALB/c mice, with no relevant off-target toxicities to major organs or tissues.[121]

Summary and Future Perspectives

In this Account, we have summarized the recent advances in the design of tumor-targeted ROS-responsive prodrugs. The design of prodrugs is a very popular strategy in the medicinal chemistry field, as it allows the modulation of pharmacokinetic and photodynamic properties of active drugs through simple and straightforward modifications. This route avoids the customary approaches of drug discovery, which are time-consuming and have high associated costs. Moreover, the possibility of attaching stimuli-responsive groups permits selective delivery of drug entities to their intended targets, which, in turn, minimizes their side effects, broadens their therapeutic window, and improves their efficacy and safety.

Despite their ubiquitous presence in most cancers, ROS are still considerably underexplored as stimuli for the development of prodrugs, particularly when compared with other common stimuli such as pH or GSH. This is probably explained by an important lack of functionalities that combine good physiological stability and selectivity for ROS. Nonetheless, various efficient ROS-responsive groups have already been developed, and their good in vitro and in vivo performances validate ROS as an effective trigger to develop tumor-targeting prodrugs. Of those, BAs and their esters are still, by far, the most widely utilized in the design of prodrugs.

When attached to active drugs, BAs can mask their activity, ensuring that their release is selective and ROS-induced. Whether directly or through a self-immolative linker, BAs are able to uncage a wide range of drugs and they have attained widespread success in the field of prodrugs. Furthermore, the introduction of these groups also permits modulation of the lipophilicity and cell permeability of the final compounds.

Nonetheless, BAs display important stability shortcomings that must be addressed to further solidify the position of BAs as valid and robust ROS-responsive triggers. In this sense, new boron-containing structures have been recently reported that are able to increase the stability of the BA group, either through the formation of an ester with salicyl hydroxamic acid[120] or through its insertion in a diazaborine heterocycle.[96] Despite not being used directly in the preparation of prodrugs, both examples displayed improved stability and H₂O₂ responsiveness, and present themselves as significant candidates for the design of a new generation of improved boron-based ROS-responsive prodrugs.

Following the same trend, a recent work by Liu et al, reported the photoactivated cleavage of the C–B bond of a phenylboronic acid derivative to generate a carbon-centered radical.[123] This highly reactive phenyl radical is able to capture O₂ which, after hydrolysis to the phenol, promotes the release of caged drugs. The researchers demonstrated that this photoactivation can be achieved with both an external (methylene blue) or internal [covalently linked iridium(II) complex] photosensitizer. While this work does not involve a direct oxidation of a boron-based prodrug by ROS per se, photoactivated cleavage of the C–B bond represents an innovative mechanism for releasing boronic acid-caged payloads, and might pave the way for the discovery of novel mechanisms in this field.

The existence of other moieties capable of undergoing oxidation by ROS extends the applicability of this strategy to other types of compounds and drug-delivery systems, while overcoming the promiscuous reactivity and pharmacokinetic limitations associated with BAs and their esters. For example, the incorporation of triggers such as thiazolidinones or 1,3-oxathiolanes permits the masking of carboxylic acid and aldehyde groups, respectively, which, until now, was not possible with boron-containing triggers. The TK group is a trigger with a high reactivity towards ¹O₂ and, therefore, has the advantage that it can be employed as a linker in the design of prodrugs that combine PDT with chemotherapy, leading to a photocontrolled drug release. On the other hand, thioethers, which undergo a hydrophobic-to-hydrophilic phase transition upon oxidation, offer the possibility of encapsulating prodrugs in self-assembled nanoparticles that are disrupted upon interaction with ROS.

Despite the progress in this area and the promising results achieved, none of the prodrugs mentioned in this work has entered clinical trials or has been approved for clinical use. In fact, for the vast majority of compounds, further in vivo testing is still required. These tests should include studies on metabolism and pharmacokinetics to confirm the prodrug’s stability and site-selective delivery.

Moreover, despite its validated responsiveness to ROS, the oxidative sensitivity of the C–B bond and its modulation are still not well documented. As such, a thorough, comprehensive study on the effects of the substituents and the Lewis acidity of the boron center on the kinetics and selectivity of the oxidation process would certainly be of high importance to future endeavors in this field.

It is also crucial to have a better understanding of the ROS-responsive mechanism and to characterize the drug-release kinetics, particularly for TK and thioethers, which is still currently not well established. In future studies, it is important to keep in mind that expressed levels of ROS may vary depending on the disease and patient, making it essential to discover new and more-sensitive moieties that respond well to low concentrations of ROS. Moreover, due to the broad range of different cellular ROS, it is important to understand the mechanisms and selectivity for each one because, in most reports, only H₂O₂ was evaluated as an ROS model. This knowledge, combined with advances in the identification and quantification of ROS levels in different diseases, might permit the design of specific prodrugs that will contribute to a new generation of precision therapies.

Despite the current limitations and the challenging future ahead, ROS are gradually cementing their position as key players in the field of tumor-targeted prodrugs, and will certainly play an important role in the future of precision chemotherapy.

Conflict of Interest

The authors declare no conflict of interest.

• References and Notes

• 1 Cancer . World Health Organization; Geneva: 2023. https://www.who.int/news-room/fact-sheets/detail/cancer (accessed Aug 23, 2023)
  Search in Google Scholar
• 2 Overview: Cancer . NHS; London: 2023. https://www.nhs.uk/conditions/cancer/ (accessed Aug 23, 2023)
  Search in Google Scholar
• 3 Patrick GL. An Introduction to Medicinal Chemistry, 5th ed. Oxford University Press; Oxford: 2013
  Search in Google Scholar
• 4 Hagen H, Marzenell P, Jentzsch E, Wenz F, Veldwijk MR, Mokhir A. J. Med. Chem. 2012; 55: 924
  CrossrefPubMedSearch in Google Scholar
• 5 Chemotherapy to Treat Cancer . National Cancer Institute; Bethesda: 2022. https://www.cancer.gov/about-cancer/treatment/types/chemotherapy (accessed Aug 23, 2023)
  Search in Google Scholar
• 6 Zhang X, Li X, You Q, Zhang X. Eur. J. Med. Chem. 2017; 139: 542
  CrossrefPubMedSearch in Google Scholar
• 7 Souza C, Pellosi DS, Tedesco AC. Expert Rev. Anticancer Ther. 2019; 19: 483
  CrossrefPubMedSearch in Google Scholar
• 8 Rautio J, Kumpulainen H, Heimbach T, Oliyai R, Oh D, Järvinen T, Savolainen J. Nat. Rev. Drug Discovery 2008; 7: 255
  CrossrefPubMedSearch in Google Scholar
• 9 Rautio J, Meanwell NA, Di L, Hageman MJ. Nat. Rev. Drug Discovery 2018; 17: 559
  CrossrefPubMedSearch in Google Scholar
• 10 Maslah H, Skarbek C, Pethe S, Labruère R. Eur. J. Med. Chem. 2020; 207: 112670
  CrossrefPubMedSearch in Google Scholar
• 11 Peng X, Gandhi V. Ther. Delivery 2012; 3: 823
  CrossrefPubMedSearch in Google Scholar
• 12 Wang P, Gong Q, Hu J, Li X, Zhang X. J. Med. Chem. 2021; 64: 298
  CrossrefPubMedSearch in Google Scholar
• 13 Ye H, Zhou Y, Liu X, Chen Y, Duan S, Zhu R, Liu Y, Yin L. Biomacromolecules 2019; 20: 2441
  CrossrefPubMedSearch in Google Scholar
• 14 Trachootham D, Alexandre J, Huang P. Nat. Rev. Drug Discovery 2009; 8: 579
  CrossrefPubMedSearch in Google Scholar
• 15 Saravanakumar G, Kim J, Kim WJ. Adv. Sci. (Weinheim, Ger.) 2017; 4: 1600124
  Search in Google Scholar
• 16 Papot S, Tranoy I, Tillequin F, Florent J.-C, Gesson J.-P. Curr. Med. Chem. 2002; 2: 155
  Search in Google Scholar
• 17 Al-Omari MK, Elaarag M, Al-Zoubi RM, Al-Qudimat AR, Zarour AA, Al-Hurani EA, Fares ZE, Alkharraz LM, Shkoor M, Bani-Yaseen AD, Aboumarzouk OM, Yassin A, Al-Ansari AA. J. Enzyme Inhib. Med. Chem. 2023; 38: 2220084
  CrossrefPubMedSearch in Google Scholar
• 18 Liang J, Liu B. Bioeng. Transl. Med. 2016; 1: 239
  CrossrefPubMedSearch in Google Scholar
• 19 Ballance WC, Qin EC, Chung HJ, Gillette MU, Kong H. Biomaterials 2019; 217: 119292
  CrossrefPubMedSearch in Google Scholar
• 20 Gao F, Xiong Z. Front. Chem. 2021; 9: 649048
  CrossrefPubMedSearch in Google Scholar
• 21 Liu J, Jia B, Li Z, Li W. Front. Bioeng. Biotechnol. 2023; 11: 1115603
  CrossrefPubMedSearch in Google Scholar
• 22 Messner K, Vuong B, Tranmer GK. Pharmaceuticals 2022; 15: 264
  CrossrefPubMedSearch in Google Scholar
• 23 Silva MP, Saraiva L, Pinto M, Sousa ME. Molecules 2020; 25: 4323
  CrossrefPubMedSearch in Google Scholar
• 24 Saxon E, Peng X. ChemBioChem 2022; 23: e202100366
  CrossrefPubMedSearch in Google Scholar
• 25 Boronic Acids: Preparation and Applications in Organic Synthesis, Medicine and Materials. Hall DG. Wiley-VCH; Weinheim: 2011
  Search in Google Scholar
• 26 Jiménez-Aligaga K, Bermejo-Bescós P, Martín-Aragón S, Csákÿ AG. Bioorg. Med. Chem. Lett. 2013; 23: 426
  CrossrefPubMedSearch in Google Scholar
• 27 Hiller N. deJ, do Amaral e Silva NA, Tavares TA, Faria RX, Nogueira Eberlin M, de Lun Martins D. Eur. J. Org. Chem. 2020; 4841
  Crossref
• 28 Frankland E, Duppa B. Proc. R. Soc. London 1860; 10: 568
  Search in Google Scholar
• 29 Lennox AJ. J, Lloyd-Jones GC. Chem. Soc. Rev. 2013; 43: 412
  CrossrefPubMedSearch in Google Scholar
• 30 Gosecki M, Gosecka M. Polymers (Basel, Switz.) 2022; 14: 842
  CrossrefPubMedSearch in Google Scholar
• 31 Miyaura N, Suzuki A. Chem. Rev. 1995; 95: 2457
  Search in Google Scholar
• 32 António JP. M, Russo R, Carvalho CP, Cal PM. S. D, Gois PM. P. Chem. Soc. Rev. 2019; 48: 3513
  CrossrefPubMedSearch in Google Scholar
• 33 Peiró Cadahía J, Previtali V, Troelsen NS, Clausen MH. MedChemComm 2019; 10: 1531
  CrossrefPubMedSearch in Google Scholar
• 34 Chen J.-W, Wu T.-C, Liang W, Ciou J.-J, Lai C.-H. Drug Delivery Transl. Res. 2022; 13: 1305
  CrossrefPubMedSearch in Google Scholar
• 35 Kennedy L, Sandhu JK, Harper M.-E, Cuperlovic-Culf M. Biomolecules 2020; 10: 1429
  CrossrefPubMedSearch in Google Scholar
• 36 Cao S, Wang Y, Peng X. Chem. Eur. J. 2012; 18: 3850
  CrossrefPubMedSearch in Google Scholar
• 37 Cao S, Christiansen R, Peng X. Chem. Eur. J. 2013; 19: 9050
  CrossrefPubMedSearch in Google Scholar
• 38 Cao S, Wang Y, Peng X. J. Org. Chem. 2014; 79: 501
  CrossrefPubMedSearch in Google Scholar
• 39 Wang Y, Fan H, Balakrishnan K, Lin Z, Cao S, Chen W, Fan Y, Guthrie QA, Sun H, Teske KA, Gandhi V, Arnold LA, Peng X. Eur. J. Med. Chem. 2017; 133: 197
  CrossrefPubMedSearch in Google Scholar
• 40 Uversky VN. In Brenner’s Encyclopedia of Genetics, 2nd ed. Maloy S, Hughes K. Elsevier; Amsterdam: 2013: 425
  CrossrefSearch in Google Scholar
• 41 Vaissière T, Sawan C, Herceg Z. Mutat. Res. 2008; 659: 40
  CrossrefPubMedSearch in Google Scholar
• 42 Ni X, Li L, Pan G. Oncol. Lett. 2015; 9: 515
  CrossrefPubMedSearch in Google Scholar
• 43 Daniel KB, Sullivan ED, Chen Y, Chan JC, Jennings PA, Fierke CA, Cohen SM. J. Med. Chem. 2015; 58: 4812
  CrossrefPubMedSearch in Google Scholar
• 44 Hull EE, Montgomery MR, Leyva KJ. BioMed Res. Int. 2016; 8797206
  CrossrefPubMed
• 45 Kavianpour P, Gemmell MC. M, Kahlert JU, Rendina LM. ChemBioChem 2020; 21: 2786
  CrossrefPubMedSearch in Google Scholar
• 46 Bhagat SD, Singh U, Mishra RK, Srivastava A. ChemMedChem 2018; 13: 2073
  CrossrefPubMedSearch in Google Scholar
• 47 Li Y, Seto E. Cold Spring Harbor Perspect. Med. 2016; 6: a026831
  CrossrefPubMedSearch in Google Scholar
• 48 Zheng S, Guo S, Zhong Q, Zhang C, Liu J, Yang L, Zhang Q, Wang G. ACS Med. Chem. Lett. 2018; 9: 149
  CrossrefPubMedSearch in Google Scholar
• 49 Fed. Regist. FR Doc. 2022-06182, 2022
  PubMedSearch in Google Scholar
• 50 Liao Y, Xu L, Ou S, Edwards H, Luedtke D, Ge Y, Qin Z. ACS Med. Chem. Lett. 2018; 9: 635
  CrossrefPubMedSearch in Google Scholar
• 51 Sreerama L. In Encyclopedia of Cancer, 3rd ed. Schwab M. Springer; Berlin: 2011: 132
  CrossrefSearch in Google Scholar
• 52 Lehmann F, Wennerberg J. Processes 2021; 9: 377
  CrossrefSearch in Google Scholar
• 53 Finch GL, Burns-Naas LA. In Encyclopedia of Toxicology, 3rd ed. Wexler P. Academic Press; London: 2014: 630
  CrossrefSearch in Google Scholar
• 54 Colvin OM. In Encyclopedia of Cancer, 2nd ed. Schwab M. Springer; Berlin: 2009: 35
  Search in Google Scholar
• 55 Chen W, Balakrishnan K, Kuang Y, Han Y, Fu M, Gandhi V, Peng X. J. Med. Chem. 2014; 57: 4498
  CrossrefPubMedSearch in Google Scholar
• 56 Singh RK, Kumar S, Prasad DN, Bhardwaj TR. Eur. J. Med. Chem. 2018; 151: 401
  CrossrefPubMedSearch in Google Scholar
• 57 Colvin M. In Holland-Frei Cancer Medicine, 6th ed., Chap. 51. Kufe D, Pollock RE, Weichselbaum RR, Bast RC, Gansler TS, Holland JF, Frei E. B. C. Decker; Hamilton: 2003
  Search in Google Scholar
• 58 Kuang Y, Balakrishnan K, Gandhi V, Peng X. J. Am. Chem. Soc. 2011; 133: 19278
  CrossrefPubMedSearch in Google Scholar
• 59 Chen W, Han Y, Peng X. Chem. Eur. J. 2014; 20: 7410
  CrossrefPubMedSearch in Google Scholar
• 60 Chen W, Fan H, Balakrishnan K, Wang Y, Sun H, Fan Y, Gandhi V, Arnold LA, Peng X. J. Med. Chem. 2018; 61: 9132
  CrossrefPubMedSearch in Google Scholar
• 61 Di Antonio M, Doria F, Richter SN, Bertipaglia C, Mella M, Sissi C, Palumbo M, Freccero M. J. Am. Chem. Soc. 2009; 131: 13132
  CrossrefPubMedSearch in Google Scholar
• 62 Pande P, Shearer J, Yang J, Greenberg WA, Rokita SE. J. Am. Chem. Soc. 1999; 121: 6773
  CrossrefSearch in Google Scholar
• 63 Patel HK, Bihani T. Pharmacol. Ther. 2018; 186: 1
  CrossrefPubMedSearch in Google Scholar
• 64 Hormone Therapy for Breast Cancer Fact Sheet§#/ITL#§. National Cancer Institute; Bethesda: 2022. (accessed Aug 23, 2023) https://www.cancer.gov/types/breast/breast-hormone-therapy-fact-sheet
  Search in Google Scholar
• 65 Liu J, Zheng S, Akerstrom VL, Yuan C, Ma Y, Zhong Q, Zhang C, Zhang Q, Guo S, Ma P, Skripnikova EV, Bratton MR, Pannuti A, Miele L, Wiese TE, Wang G. J. Med. Chem. 2016; 59: 8134
  CrossrefPubMedSearch in Google Scholar
• 66 Jiang Q, Zhong Q, Zhang Q, Zheng S, Wang G. ACS Med. Chem. Lett. 2012; 3: 392
  CrossrefPubMedSearch in Google Scholar
• 67 Zhong Q, Zhang C, Zhang Q, Miele L, Zheng S, Wang G. BMC Cancer 2015; 15: 625
  CrossrefPubMedSearch in Google Scholar
• 68 Zhang C, Zhong Q, Zhang Q, Zheng S, Miele L, Wang G. Breast Cancer Res. 2015; 152: 283
  CrossrefPubMedSearch in Google Scholar
• 69 Zhang C, Guo S, Yang L, Liu J, Zheng S, Zhong Q, Zhang Q, Wang G. Oncotarget 2017; 8: 103874
  CrossrefPubMedSearch in Google Scholar
• 70 Liu J, Zheng S, Guo S, Zhang C, Zhong Q, Zhang Q, Ma P, Skripnikova EV, Bratton MR, Wiese TE, Wang G. ACS Med. Chem. Lett. 2017; 8: 102
  CrossrefPubMedSearch in Google Scholar
• 71 Noh J, Kwon B, Han E, Park M, Yang W, Cho W, Yoo W, Khang G, Lee D. Nat. Commun. 2015; 6: 6907
  CrossrefPubMedSearch in Google Scholar
• 72 Liu Y, Liu H, Wang L, Wang Y, Zhang C, Wang C, Yan Y, Fan J, Xu G, Zhang Q. Nanoscale 2020; 2: 3872
  CrossrefPubMedSearch in Google Scholar
• 73 Fang J, Nakamura H, Iyer AK. J. Drug Targeting 2007; 15: 475
  CrossrefPubMedSearch in Google Scholar
• 74 Reshetnikov V, Daum S, Janko C, Karawacka W, Tietze R, Alexiou C, Paryzhak S, Dumych T, Bilyy R, Tripal P, Schmid B, Palmisano R, Mokhir A. Angew. Chem. Int. Ed. 2018; 57: 11943
  CrossrefPubMedSearch in Google Scholar
• 75 Reshetnikov V, Daum S, Mokhir A. Chem. Eur. J. 2017; 23: 5678
  CrossrefPubMedSearch in Google Scholar
• 76 Reshetnikov V, Arkhypov A, Julakanti PR, Mokhir A. Dalton Trans. 2018; 47: 6679
  CrossrefPubMedSearch in Google Scholar
• 77 Marzenell P, Hagen H, Sellner L, Zenz T, Grinyte R, Pavlov V, Daum S, Mokhir A. J. Med. Chem. 2013; 56: 6935
  CrossrefPubMedSearch in Google Scholar
• 78 Schikora M, Reznikov A, Chaykovskaya L, Sachinska O, Polyakova L, Mokhir A. Bioorg. Med. Chem. Lett. 2015; 25: 3447
  CrossrefPubMedSearch in Google Scholar
• 79 Daum S, Chekhun VF, Todor IN, Lukianova NY, Shvets YV, Sellner L, Putzker K, Lewis J, Zenz T, de Graaf IA. M, Groothuis GM. M, Casini A, Zozulia O, Hampel F, Mokhir A. J. Med. Chem. 2015; 58: 2015
  CrossrefPubMedSearch in Google Scholar
• 80 Daum S, Reshetnikov MS. V, Sisa M, Dumych T, Lootsik MD, Bilyy R, Bila E, Janko C, Alexiou C, Herrmann M, Sellner L, Mokhir A. Angew. Chem. Int. Ed. 2017; 56: 15545
  CrossrefPubMedSearch in Google Scholar
• 81 Daum S, Babiy S, Konovalova H, Hofer W, Shtemenko A, Shtemenko N, Janko C, Alexiou C, Mokhir A. J. Inorg. Biochem. 2018; 178: 9
  CrossrefPubMedSearch in Google Scholar
• 82 Reshetnikov V, Özkan HG, Daum S, Janko C, Alexiou C, Sauer C, Heinrich MR, Mokhir A. Molecules 2020; 25: 2545
  CrossrefPubMedSearch in Google Scholar
• 83 Xu H.-G, Schikora M, Sisa M, Daum S, Klemt I, Janko C, Alexiou C, Bila G, Bilyy R, Gong W, Schmitt M, Sellner L, Mokhir A. Angew. Chem. Int. Ed. 2021; 60: 11158
  CrossrefPubMedSearch in Google Scholar
• 84 Xu H.-G, Annamadov S, Mokhir A. J. Organomet. Chem. 2022; 964: 122305
  CrossrefSearch in Google Scholar
• 85 Pan Q, Zhang B, Peng X, Wan S, Luo K, Gao W, Pu Y, He B. Chem. Commun. 2019; 55: 13896
  CrossrefPubMedSearch in Google Scholar
• 86 Bao X.-Z, Wang Q, Ren X.-R, Dai F, Zhou B. Free Radical Biol. Med. 2020; 597
  CrossrefPubMed
• 87 Dewick PM. Medicinal Natural Products: A Biosynthetic Approach, 3rd ed. Wiley; Chichester: 2009: 380
  Search in Google Scholar
• 88 Wang L, Xie S, Ma L, Chen Y, Lu W. Eur. J. Med. Chem. 2016; 116: 84
  CrossrefPubMedSearch in Google Scholar
• 89 Wall ME, Wani MC, Cook CE, Palmer KH, McPhail AT, Sim GA. J. Am. Chem. Soc. 1966; 88: 3888
  CrossrefSearch in Google Scholar
• 90 Ai Y, Obianom ON, Kuser M, Li Y, Shu Y, Xue F. ACS Med. Chem. Lett. 2019; 10: 127
  CrossrefPubMedSearch in Google Scholar
• 91 Ghafouri-Fard S, Abak A, Anamag FT, Shoorei H, Fattahi F, Javadinia SA, Basiri A, Taheri M. Front. Oncol. 2021; 11: 658636
  CrossrefPubMedSearch in Google Scholar
• 92 Matsushita K, Okuda T, Mori S, Konno M, Eguchi H, Asai A, Koseki J, Iwagami Y, Yamada D, Akita H, Asaoka T, Noda T, Kawamoto K, Gotoh K, Kobayashi S, Kasahara Y, Morihiro K, Satoh T, Doki Y, Mori M, Ishii H, Obika S. ChemMedChem 2019; 14: 1384
  CrossrefPubMedSearch in Google Scholar
• 93 Martino R, Gilard V, Malet-Martino M. In NMR Spectroscopy in Pharmaceutical Analysis . Holzgrabe U, Wawer I, Diehl B. Elsevier; Amsterdam: 2008: 369
  CrossrefSearch in Google Scholar
• 94 Waks AG, Winer EP. In The Breast: Comprehensive Management of Benign and Malignant Diseases, 5th ed. Bland KI, Copeland EM. III. Saunders Elsevier; Philadelphia: 2009: 885
  Search in Google Scholar
• 95 Skarbek C, Serra S, Maslah H, Rascol E, Labruère R. Bioorg. Chem. 2019; 91: 103158
  CrossrefPubMedSearch in Google Scholar
• 96 Dewick PM. Medicinal Natural Products: A Biosynthetic Approach, 3rd ed. Wiley; Chichester: 2009: 129
  CrossrefSearch in Google Scholar
• 97 Bielec B, Poetsch I, Ahmed E, Heffeter P, Keppler BK, Kowol CR. Molecules 2020; 25: 1149
  CrossrefPubMedSearch in Google Scholar
• 98 Yang X.-Y, Yuan B, Xiong H, Zhao Y, Wang L, Zhang S.-Q, Mao S. Bioorg. Chem. 2022; 129: 106154
  CrossrefPubMedSearch in Google Scholar
• 99 António JP. M, Carvalho JI, André AS, Dias JN. R, Aguiar SI, Faustino H, Lopes RM. R. M, Veiros LF, Bernardes GJ. L, da Silva FA, Gois PM. P. Angew. Chem. Int. Ed. 2021; 60: 25914
  CrossrefPubMedSearch in Google Scholar
• 100 Nirwan S, Chahal V, Kakkar R. J. Heterocycl. Chem. 2019; 56: 1239
  CrossrefSearch in Google Scholar
• 101 Liaras K, Fesatidou M, Geronikaki A. Molecules 2018; 23: 685
  CrossrefPubMedSearch in Google Scholar
• 102 Perez C, Monserrat J.-P, Chen Y, Cohen SM. Chem. Commun. 2015; 51: 7116
  CrossrefPubMedSearch in Google Scholar
• 103 Andersen NS, Peiró Cadahía J, Previtali V, Bondebjerg J, Hansen CA, Hansen AE, Andresen TL, Clausen MH. Eur. J. Med. Chem. 2018; 156: 738
  CrossrefPubMedSearch in Google Scholar
• 104 Lu M, Zhang X, Zhao J, You Q, Jiang Z. Redox Biol. 2020; 34: 101565
  CrossrefPubMedSearch in Google Scholar
• 105 Daniel KB, Major Jourden JL, Negoescu KE, Cohen SM. JBIC, J. Biol. Inorg. Chem. 2011; 16: 313
  CrossrefPubMedSearch in Google Scholar
• 106 Won M, Kim JH, Ji MS, Kim JS. Chem. Commun. 2022; 58: 72
  CrossrefPubMedSearch in Google Scholar
• 107 Fernandes AP, Gandin V. Biochim. Biophys. Acta, Gen. Subj. 2015; 1850: 1642
  CrossrefPubMedSearch in Google Scholar
• 108 Sharma AK, Amin S. Future Med. Chem. 2013; 5: 163
  CrossrefPubMedSearch in Google Scholar
• 109 Pan Z, Zhang J, Ji K, Chittavong V, Ji X, Wang B. Org. Lett. 2018; 20: 8
  CrossrefPubMedSearch in Google Scholar
• 110 Criado-Gonzalez M, Mecerreyes D. J. Mater. Chem. B 2022; 10: 7206
  CrossrefPubMedSearch in Google Scholar
• 111 Bohrer A.-S, Takahashi H. Int. Rev. Cell Mol. Biol. 2016; 326: 1
  CrossrefPubMedSearch in Google Scholar
• 112 Maeda H, Fukuyasu Y, Yoshida S, Fukuda M, Saeki K, Matsuno H, Yamauchi Y, Yoshida K, Hirata K, Miyamoto K. Angew. Chem. Int. Ed. 2004; 43: 2389
  CrossrefPubMedSearch in Google Scholar
• 113 Maeda H, Yamamoto K, Nomura Y, Kohno I, Hafsi L, Ueda N, Yoshida S, Fukuda M, Fukuyasu Y, Yamauchi Y, Itoh N. J. Am. Chem. Soc. 2005; 127: 68
  CrossrefPubMedSearch in Google Scholar
• 114 Xu K, Tang B, Huang H, Yang G, Chen Z, Li P, An L. Chem. Commun. 2005; 5974
  CrossrefPubMed
• 115 Rinaldi A, Caraffi R, Grazioli MV, Oddone N, Giardino L, Tosi G, Vandelli MA, Calzà L, Ruozi B, Duskey JT. Polymers (Basel, Switz.) 2022; 14: 687
  CrossrefPubMedSearch in Google Scholar
• 116 El-Mohtadi F, d’Arcy R, Tirelli N. Macromol. Rapid Commun. 2019; 40: 1800699
  CrossrefPubMedSearch in Google Scholar
• 117 Liu B, Thayumanavan S. Cell Reports Phys. Sci. 2020; 1: 100271
  CrossrefSearch in Google Scholar
• 118 Liu L.-H, Qiu W.-X, Li B, Zhang C, Sun L.-F, Wan S.-S, Rong L, Zhang X.-Z. Adv. Funct. Mater. 2016; 26: 6257
  CrossrefSearch in Google Scholar
• 119 Ma J, Chen Y, Liang W, Li L, Du J, Pan C, Zhang C. Drug Delivery 2021; 28: 1204
  CrossrefPubMedSearch in Google Scholar
• 120 Kim Y.-S, Kim S, Kang HC, Shim MS. J. Ind. Eng. Chem. 2019; 75: 238
  CrossrefSearch in Google Scholar
• 121 Yang J, Lv Q, Wei W, Yang Z, Dong J, Zhang R, Kan Q, He Z, Xu Y. Drug Delivery 2018; 25: 807
  CrossrefPubMedSearch in Google Scholar
• 122 Luo C, Sun J, Liu D, Sun B, Miao L, Musetti S, Li J, Han X, Du Y, Li L, Huang L, He Z. Nano Lett. 2016; 16: 5401
  CrossrefPubMedSearch in Google Scholar
• 123 Liu M, Luo Y, Yan J, Xiong X, Xing X, Kim JS, Zou T. J. Am. Chem. Soc. 2023; 145: 10082
  CrossrefPubMedSearch in Google Scholar

Corresponding Authors

Publication History

Received: 14 June 2023

Accepted after revision: 18 July 2023

Accepted Manuscript online:
18 July 2023

Article published online:
05 September 2023

© 2023. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany

• References and Notes

• 1 Cancer . World Health Organization; Geneva: 2023. https://www.who.int/news-room/fact-sheets/detail/cancer (accessed Aug 23, 2023)
  Search in Google Scholar
• 2 Overview: Cancer . NHS; London: 2023. https://www.nhs.uk/conditions/cancer/ (accessed Aug 23, 2023)
  Search in Google Scholar
• 3 Patrick GL. An Introduction to Medicinal Chemistry, 5th ed. Oxford University Press; Oxford: 2013
  Search in Google Scholar
• 4 Hagen H, Marzenell P, Jentzsch E, Wenz F, Veldwijk MR, Mokhir A. J. Med. Chem. 2012; 55: 924
  CrossrefPubMedSearch in Google Scholar
• 5 Chemotherapy to Treat Cancer . National Cancer Institute; Bethesda: 2022. https://www.cancer.gov/about-cancer/treatment/types/chemotherapy (accessed Aug 23, 2023)
  Search in Google Scholar
• 6 Zhang X, Li X, You Q, Zhang X. Eur. J. Med. Chem. 2017; 139: 542
  CrossrefPubMedSearch in Google Scholar
• 7 Souza C, Pellosi DS, Tedesco AC. Expert Rev. Anticancer Ther. 2019; 19: 483
  CrossrefPubMedSearch in Google Scholar
• 8 Rautio J, Kumpulainen H, Heimbach T, Oliyai R, Oh D, Järvinen T, Savolainen J. Nat. Rev. Drug Discovery 2008; 7: 255
  CrossrefPubMedSearch in Google Scholar
• 9 Rautio J, Meanwell NA, Di L, Hageman MJ. Nat. Rev. Drug Discovery 2018; 17: 559
  CrossrefPubMedSearch in Google Scholar
• 10 Maslah H, Skarbek C, Pethe S, Labruère R. Eur. J. Med. Chem. 2020; 207: 112670
  CrossrefPubMedSearch in Google Scholar
• 11 Peng X, Gandhi V. Ther. Delivery 2012; 3: 823
  CrossrefPubMedSearch in Google Scholar
• 12 Wang P, Gong Q, Hu J, Li X, Zhang X. J. Med. Chem. 2021; 64: 298
  CrossrefPubMedSearch in Google Scholar
• 13 Ye H, Zhou Y, Liu X, Chen Y, Duan S, Zhu R, Liu Y, Yin L. Biomacromolecules 2019; 20: 2441
  CrossrefPubMedSearch in Google Scholar
• 14 Trachootham D, Alexandre J, Huang P. Nat. Rev. Drug Discovery 2009; 8: 579
  CrossrefPubMedSearch in Google Scholar
• 15 Saravanakumar G, Kim J, Kim WJ. Adv. Sci. (Weinheim, Ger.) 2017; 4: 1600124
  Search in Google Scholar
• 16 Papot S, Tranoy I, Tillequin F, Florent J.-C, Gesson J.-P. Curr. Med. Chem. 2002; 2: 155
  Search in Google Scholar
• 17 Al-Omari MK, Elaarag M, Al-Zoubi RM, Al-Qudimat AR, Zarour AA, Al-Hurani EA, Fares ZE, Alkharraz LM, Shkoor M, Bani-Yaseen AD, Aboumarzouk OM, Yassin A, Al-Ansari AA. J. Enzyme Inhib. Med. Chem. 2023; 38: 2220084
  CrossrefPubMedSearch in Google Scholar
• 18 Liang J, Liu B. Bioeng. Transl. Med. 2016; 1: 239
  CrossrefPubMedSearch in Google Scholar
• 19 Ballance WC, Qin EC, Chung HJ, Gillette MU, Kong H. Biomaterials 2019; 217: 119292
  CrossrefPubMedSearch in Google Scholar
• 20 Gao F, Xiong Z. Front. Chem. 2021; 9: 649048
  CrossrefPubMedSearch in Google Scholar
• 21 Liu J, Jia B, Li Z, Li W. Front. Bioeng. Biotechnol. 2023; 11: 1115603
  CrossrefPubMedSearch in Google Scholar
• 22 Messner K, Vuong B, Tranmer GK. Pharmaceuticals 2022; 15: 264
  CrossrefPubMedSearch in Google Scholar
• 23 Silva MP, Saraiva L, Pinto M, Sousa ME. Molecules 2020; 25: 4323
  CrossrefPubMedSearch in Google Scholar
• 24 Saxon E, Peng X. ChemBioChem 2022; 23: e202100366
  CrossrefPubMedSearch in Google Scholar
• 25 Boronic Acids: Preparation and Applications in Organic Synthesis, Medicine and Materials. Hall DG. Wiley-VCH; Weinheim: 2011
  Search in Google Scholar
• 26 Jiménez-Aligaga K, Bermejo-Bescós P, Martín-Aragón S, Csákÿ AG. Bioorg. Med. Chem. Lett. 2013; 23: 426
  CrossrefPubMedSearch in Google Scholar
• 27 Hiller N. deJ, do Amaral e Silva NA, Tavares TA, Faria RX, Nogueira Eberlin M, de Lun Martins D. Eur. J. Org. Chem. 2020; 4841
  Crossref
• 28 Frankland E, Duppa B. Proc. R. Soc. London 1860; 10: 568
  Search in Google Scholar
• 29 Lennox AJ. J, Lloyd-Jones GC. Chem. Soc. Rev. 2013; 43: 412
  CrossrefPubMedSearch in Google Scholar
• 30 Gosecki M, Gosecka M. Polymers (Basel, Switz.) 2022; 14: 842
  CrossrefPubMedSearch in Google Scholar
• 31 Miyaura N, Suzuki A. Chem. Rev. 1995; 95: 2457
  Search in Google Scholar
• 32 António JP. M, Russo R, Carvalho CP, Cal PM. S. D, Gois PM. P. Chem. Soc. Rev. 2019; 48: 3513
  CrossrefPubMedSearch in Google Scholar
• 33 Peiró Cadahía J, Previtali V, Troelsen NS, Clausen MH. MedChemComm 2019; 10: 1531
  CrossrefPubMedSearch in Google Scholar
• 34 Chen J.-W, Wu T.-C, Liang W, Ciou J.-J, Lai C.-H. Drug Delivery Transl. Res. 2022; 13: 1305
  CrossrefPubMedSearch in Google Scholar
• 35 Kennedy L, Sandhu JK, Harper M.-E, Cuperlovic-Culf M. Biomolecules 2020; 10: 1429
  CrossrefPubMedSearch in Google Scholar
• 36 Cao S, Wang Y, Peng X. Chem. Eur. J. 2012; 18: 3850
  CrossrefPubMedSearch in Google Scholar
• 37 Cao S, Christiansen R, Peng X. Chem. Eur. J. 2013; 19: 9050
  CrossrefPubMedSearch in Google Scholar
• 38 Cao S, Wang Y, Peng X. J. Org. Chem. 2014; 79: 501
  CrossrefPubMedSearch in Google Scholar
• 39 Wang Y, Fan H, Balakrishnan K, Lin Z, Cao S, Chen W, Fan Y, Guthrie QA, Sun H, Teske KA, Gandhi V, Arnold LA, Peng X. Eur. J. Med. Chem. 2017; 133: 197
  CrossrefPubMedSearch in Google Scholar
• 40 Uversky VN. In Brenner’s Encyclopedia of Genetics, 2nd ed. Maloy S, Hughes K. Elsevier; Amsterdam: 2013: 425
  CrossrefSearch in Google Scholar
• 41 Vaissière T, Sawan C, Herceg Z. Mutat. Res. 2008; 659: 40
  CrossrefPubMedSearch in Google Scholar
• 42 Ni X, Li L, Pan G. Oncol. Lett. 2015; 9: 515
  CrossrefPubMedSearch in Google Scholar
• 43 Daniel KB, Sullivan ED, Chen Y, Chan JC, Jennings PA, Fierke CA, Cohen SM. J. Med. Chem. 2015; 58: 4812
  CrossrefPubMedSearch in Google Scholar
• 44 Hull EE, Montgomery MR, Leyva KJ. BioMed Res. Int. 2016; 8797206
  CrossrefPubMed
• 45 Kavianpour P, Gemmell MC. M, Kahlert JU, Rendina LM. ChemBioChem 2020; 21: 2786
  CrossrefPubMedSearch in Google Scholar
• 46 Bhagat SD, Singh U, Mishra RK, Srivastava A. ChemMedChem 2018; 13: 2073
  CrossrefPubMedSearch in Google Scholar
• 47 Li Y, Seto E. Cold Spring Harbor Perspect. Med. 2016; 6: a026831
  CrossrefPubMedSearch in Google Scholar
• 48 Zheng S, Guo S, Zhong Q, Zhang C, Liu J, Yang L, Zhang Q, Wang G. ACS Med. Chem. Lett. 2018; 9: 149
  CrossrefPubMedSearch in Google Scholar
• 49 Fed. Regist. FR Doc. 2022-06182, 2022
  PubMedSearch in Google Scholar
• 50 Liao Y, Xu L, Ou S, Edwards H, Luedtke D, Ge Y, Qin Z. ACS Med. Chem. Lett. 2018; 9: 635
  CrossrefPubMedSearch in Google Scholar
• 51 Sreerama L. In Encyclopedia of Cancer, 3rd ed. Schwab M. Springer; Berlin: 2011: 132
  CrossrefSearch in Google Scholar
• 52 Lehmann F, Wennerberg J. Processes 2021; 9: 377
  CrossrefSearch in Google Scholar
• 53 Finch GL, Burns-Naas LA. In Encyclopedia of Toxicology, 3rd ed. Wexler P. Academic Press; London: 2014: 630
  CrossrefSearch in Google Scholar
• 54 Colvin OM. In Encyclopedia of Cancer, 2nd ed. Schwab M. Springer; Berlin: 2009: 35
  Search in Google Scholar
• 55 Chen W, Balakrishnan K, Kuang Y, Han Y, Fu M, Gandhi V, Peng X. J. Med. Chem. 2014; 57: 4498
  CrossrefPubMedSearch in Google Scholar
• 56 Singh RK, Kumar S, Prasad DN, Bhardwaj TR. Eur. J. Med. Chem. 2018; 151: 401
  CrossrefPubMedSearch in Google Scholar
• 57 Colvin M. In Holland-Frei Cancer Medicine, 6th ed., Chap. 51. Kufe D, Pollock RE, Weichselbaum RR, Bast RC, Gansler TS, Holland JF, Frei E. B. C. Decker; Hamilton: 2003
  Search in Google Scholar
• 58 Kuang Y, Balakrishnan K, Gandhi V, Peng X. J. Am. Chem. Soc. 2011; 133: 19278
  CrossrefPubMedSearch in Google Scholar
• 59 Chen W, Han Y, Peng X. Chem. Eur. J. 2014; 20: 7410
  CrossrefPubMedSearch in Google Scholar
• 60 Chen W, Fan H, Balakrishnan K, Wang Y, Sun H, Fan Y, Gandhi V, Arnold LA, Peng X. J. Med. Chem. 2018; 61: 9132
  CrossrefPubMedSearch in Google Scholar
• 61 Di Antonio M, Doria F, Richter SN, Bertipaglia C, Mella M, Sissi C, Palumbo M, Freccero M. J. Am. Chem. Soc. 2009; 131: 13132
  CrossrefPubMedSearch in Google Scholar
• 62 Pande P, Shearer J, Yang J, Greenberg WA, Rokita SE. J. Am. Chem. Soc. 1999; 121: 6773
  CrossrefSearch in Google Scholar
• 63 Patel HK, Bihani T. Pharmacol. Ther. 2018; 186: 1
  CrossrefPubMedSearch in Google Scholar
• 64 Hormone Therapy for Breast Cancer Fact Sheet§#/ITL#§. National Cancer Institute; Bethesda: 2022. (accessed Aug 23, 2023) https://www.cancer.gov/types/breast/breast-hormone-therapy-fact-sheet
  Search in Google Scholar
• 65 Liu J, Zheng S, Akerstrom VL, Yuan C, Ma Y, Zhong Q, Zhang C, Zhang Q, Guo S, Ma P, Skripnikova EV, Bratton MR, Pannuti A, Miele L, Wiese TE, Wang G. J. Med. Chem. 2016; 59: 8134
  CrossrefPubMedSearch in Google Scholar
• 66 Jiang Q, Zhong Q, Zhang Q, Zheng S, Wang G. ACS Med. Chem. Lett. 2012; 3: 392
  CrossrefPubMedSearch in Google Scholar
• 67 Zhong Q, Zhang C, Zhang Q, Miele L, Zheng S, Wang G. BMC Cancer 2015; 15: 625
  CrossrefPubMedSearch in Google Scholar
• 68 Zhang C, Zhong Q, Zhang Q, Zheng S, Miele L, Wang G. Breast Cancer Res. 2015; 152: 283
  CrossrefPubMedSearch in Google Scholar
• 69 Zhang C, Guo S, Yang L, Liu J, Zheng S, Zhong Q, Zhang Q, Wang G. Oncotarget 2017; 8: 103874
  CrossrefPubMedSearch in Google Scholar
• 70 Liu J, Zheng S, Guo S, Zhang C, Zhong Q, Zhang Q, Ma P, Skripnikova EV, Bratton MR, Wiese TE, Wang G. ACS Med. Chem. Lett. 2017; 8: 102
  CrossrefPubMedSearch in Google Scholar
• 71 Noh J, Kwon B, Han E, Park M, Yang W, Cho W, Yoo W, Khang G, Lee D. Nat. Commun. 2015; 6: 6907
  CrossrefPubMedSearch in Google Scholar
• 72 Liu Y, Liu H, Wang L, Wang Y, Zhang C, Wang C, Yan Y, Fan J, Xu G, Zhang Q. Nanoscale 2020; 2: 3872
  CrossrefPubMedSearch in Google Scholar
• 73 Fang J, Nakamura H, Iyer AK. J. Drug Targeting 2007; 15: 475
  CrossrefPubMedSearch in Google Scholar
• 74 Reshetnikov V, Daum S, Janko C, Karawacka W, Tietze R, Alexiou C, Paryzhak S, Dumych T, Bilyy R, Tripal P, Schmid B, Palmisano R, Mokhir A. Angew. Chem. Int. Ed. 2018; 57: 11943
  CrossrefPubMedSearch in Google Scholar
• 75 Reshetnikov V, Daum S, Mokhir A. Chem. Eur. J. 2017; 23: 5678
  CrossrefPubMedSearch in Google Scholar
• 76 Reshetnikov V, Arkhypov A, Julakanti PR, Mokhir A. Dalton Trans. 2018; 47: 6679
  CrossrefPubMedSearch in Google Scholar
• 77 Marzenell P, Hagen H, Sellner L, Zenz T, Grinyte R, Pavlov V, Daum S, Mokhir A. J. Med. Chem. 2013; 56: 6935
  CrossrefPubMedSearch in Google Scholar
• 78 Schikora M, Reznikov A, Chaykovskaya L, Sachinska O, Polyakova L, Mokhir A. Bioorg. Med. Chem. Lett. 2015; 25: 3447
  CrossrefPubMedSearch in Google Scholar
• 79 Daum S, Chekhun VF, Todor IN, Lukianova NY, Shvets YV, Sellner L, Putzker K, Lewis J, Zenz T, de Graaf IA. M, Groothuis GM. M, Casini A, Zozulia O, Hampel F, Mokhir A. J. Med. Chem. 2015; 58: 2015
  CrossrefPubMedSearch in Google Scholar
• 80 Daum S, Reshetnikov MS. V, Sisa M, Dumych T, Lootsik MD, Bilyy R, Bila E, Janko C, Alexiou C, Herrmann M, Sellner L, Mokhir A. Angew. Chem. Int. Ed. 2017; 56: 15545
  CrossrefPubMedSearch in Google Scholar
• 81 Daum S, Babiy S, Konovalova H, Hofer W, Shtemenko A, Shtemenko N, Janko C, Alexiou C, Mokhir A. J. Inorg. Biochem. 2018; 178: 9
  CrossrefPubMedSearch in Google Scholar
• 82 Reshetnikov V, Özkan HG, Daum S, Janko C, Alexiou C, Sauer C, Heinrich MR, Mokhir A. Molecules 2020; 25: 2545
  CrossrefPubMedSearch in Google Scholar
• 83 Xu H.-G, Schikora M, Sisa M, Daum S, Klemt I, Janko C, Alexiou C, Bila G, Bilyy R, Gong W, Schmitt M, Sellner L, Mokhir A. Angew. Chem. Int. Ed. 2021; 60: 11158
  CrossrefPubMedSearch in Google Scholar
• 84 Xu H.-G, Annamadov S, Mokhir A. J. Organomet. Chem. 2022; 964: 122305
  CrossrefSearch in Google Scholar
• 85 Pan Q, Zhang B, Peng X, Wan S, Luo K, Gao W, Pu Y, He B. Chem. Commun. 2019; 55: 13896
  CrossrefPubMedSearch in Google Scholar
• 86 Bao X.-Z, Wang Q, Ren X.-R, Dai F, Zhou B. Free Radical Biol. Med. 2020; 597
  CrossrefPubMed
• 87 Dewick PM. Medicinal Natural Products: A Biosynthetic Approach, 3rd ed. Wiley; Chichester: 2009: 380
  Search in Google Scholar
• 88 Wang L, Xie S, Ma L, Chen Y, Lu W. Eur. J. Med. Chem. 2016; 116: 84
  CrossrefPubMedSearch in Google Scholar
• 89 Wall ME, Wani MC, Cook CE, Palmer KH, McPhail AT, Sim GA. J. Am. Chem. Soc. 1966; 88: 3888
  CrossrefSearch in Google Scholar
• 90 Ai Y, Obianom ON, Kuser M, Li Y, Shu Y, Xue F. ACS Med. Chem. Lett. 2019; 10: 127
  CrossrefPubMedSearch in Google Scholar
• 91 Ghafouri-Fard S, Abak A, Anamag FT, Shoorei H, Fattahi F, Javadinia SA, Basiri A, Taheri M. Front. Oncol. 2021; 11: 658636
  CrossrefPubMedSearch in Google Scholar
• 92 Matsushita K, Okuda T, Mori S, Konno M, Eguchi H, Asai A, Koseki J, Iwagami Y, Yamada D, Akita H, Asaoka T, Noda T, Kawamoto K, Gotoh K, Kobayashi S, Kasahara Y, Morihiro K, Satoh T, Doki Y, Mori M, Ishii H, Obika S. ChemMedChem 2019; 14: 1384
  CrossrefPubMedSearch in Google Scholar
• 93 Martino R, Gilard V, Malet-Martino M. In NMR Spectroscopy in Pharmaceutical Analysis . Holzgrabe U, Wawer I, Diehl B. Elsevier; Amsterdam: 2008: 369
  CrossrefSearch in Google Scholar
• 94 Waks AG, Winer EP. In The Breast: Comprehensive Management of Benign and Malignant Diseases, 5th ed. Bland KI, Copeland EM. III. Saunders Elsevier; Philadelphia: 2009: 885
  Search in Google Scholar
• 95 Skarbek C, Serra S, Maslah H, Rascol E, Labruère R. Bioorg. Chem. 2019; 91: 103158
  CrossrefPubMedSearch in Google Scholar
• 96 Dewick PM. Medicinal Natural Products: A Biosynthetic Approach, 3rd ed. Wiley; Chichester: 2009: 129
  CrossrefSearch in Google Scholar
• 97 Bielec B, Poetsch I, Ahmed E, Heffeter P, Keppler BK, Kowol CR. Molecules 2020; 25: 1149
  CrossrefPubMedSearch in Google Scholar
• 98 Yang X.-Y, Yuan B, Xiong H, Zhao Y, Wang L, Zhang S.-Q, Mao S. Bioorg. Chem. 2022; 129: 106154
  CrossrefPubMedSearch in Google Scholar
• 99 António JP. M, Carvalho JI, André AS, Dias JN. R, Aguiar SI, Faustino H, Lopes RM. R. M, Veiros LF, Bernardes GJ. L, da Silva FA, Gois PM. P. Angew. Chem. Int. Ed. 2021; 60: 25914
  CrossrefPubMedSearch in Google Scholar
• 100 Nirwan S, Chahal V, Kakkar R. J. Heterocycl. Chem. 2019; 56: 1239
  CrossrefSearch in Google Scholar
• 101 Liaras K, Fesatidou M, Geronikaki A. Molecules 2018; 23: 685
  CrossrefPubMedSearch in Google Scholar
• 102 Perez C, Monserrat J.-P, Chen Y, Cohen SM. Chem. Commun. 2015; 51: 7116
  CrossrefPubMedSearch in Google Scholar
• 103 Andersen NS, Peiró Cadahía J, Previtali V, Bondebjerg J, Hansen CA, Hansen AE, Andresen TL, Clausen MH. Eur. J. Med. Chem. 2018; 156: 738
  CrossrefPubMedSearch in Google Scholar
• 104 Lu M, Zhang X, Zhao J, You Q, Jiang Z. Redox Biol. 2020; 34: 101565
  CrossrefPubMedSearch in Google Scholar
• 105 Daniel KB, Major Jourden JL, Negoescu KE, Cohen SM. JBIC, J. Biol. Inorg. Chem. 2011; 16: 313
  CrossrefPubMedSearch in Google Scholar
• 106 Won M, Kim JH, Ji MS, Kim JS. Chem. Commun. 2022; 58: 72
  CrossrefPubMedSearch in Google Scholar
• 107 Fernandes AP, Gandin V. Biochim. Biophys. Acta, Gen. Subj. 2015; 1850: 1642
  CrossrefPubMedSearch in Google Scholar
• 108 Sharma AK, Amin S. Future Med. Chem. 2013; 5: 163
  CrossrefPubMedSearch in Google Scholar
• 109 Pan Z, Zhang J, Ji K, Chittavong V, Ji X, Wang B. Org. Lett. 2018; 20: 8
  CrossrefPubMedSearch in Google Scholar
• 110 Criado-Gonzalez M, Mecerreyes D. J. Mater. Chem. B 2022; 10: 7206
  CrossrefPubMedSearch in Google Scholar
• 111 Bohrer A.-S, Takahashi H. Int. Rev. Cell Mol. Biol. 2016; 326: 1
  CrossrefPubMedSearch in Google Scholar
• 112 Maeda H, Fukuyasu Y, Yoshida S, Fukuda M, Saeki K, Matsuno H, Yamauchi Y, Yoshida K, Hirata K, Miyamoto K. Angew. Chem. Int. Ed. 2004; 43: 2389
  CrossrefPubMedSearch in Google Scholar
• 113 Maeda H, Yamamoto K, Nomura Y, Kohno I, Hafsi L, Ueda N, Yoshida S, Fukuda M, Fukuyasu Y, Yamauchi Y, Itoh N. J. Am. Chem. Soc. 2005; 127: 68
  CrossrefPubMedSearch in Google Scholar
• 114 Xu K, Tang B, Huang H, Yang G, Chen Z, Li P, An L. Chem. Commun. 2005; 5974
  CrossrefPubMed
• 115 Rinaldi A, Caraffi R, Grazioli MV, Oddone N, Giardino L, Tosi G, Vandelli MA, Calzà L, Ruozi B, Duskey JT. Polymers (Basel, Switz.) 2022; 14: 687
  CrossrefPubMedSearch in Google Scholar
• 116 El-Mohtadi F, d’Arcy R, Tirelli N. Macromol. Rapid Commun. 2019; 40: 1800699
  CrossrefPubMedSearch in Google Scholar
• 117 Liu B, Thayumanavan S. Cell Reports Phys. Sci. 2020; 1: 100271
  CrossrefSearch in Google Scholar
• 118 Liu L.-H, Qiu W.-X, Li B, Zhang C, Sun L.-F, Wan S.-S, Rong L, Zhang X.-Z. Adv. Funct. Mater. 2016; 26: 6257
  CrossrefSearch in Google Scholar
• 119 Ma J, Chen Y, Liang W, Li L, Du J, Pan C, Zhang C. Drug Delivery 2021; 28: 1204
  CrossrefPubMedSearch in Google Scholar
• 120 Kim Y.-S, Kim S, Kang HC, Shim MS. J. Ind. Eng. Chem. 2019; 75: 238
  CrossrefSearch in Google Scholar
• 121 Yang J, Lv Q, Wei W, Yang Z, Dong J, Zhang R, Kan Q, He Z, Xu Y. Drug Delivery 2018; 25: 807
  CrossrefPubMedSearch in Google Scholar
• 122 Luo C, Sun J, Liu D, Sun B, Miao L, Musetti S, Li J, Han X, Du Y, Li L, Huang L, He Z. Nano Lett. 2016; 16: 5401
  CrossrefPubMedSearch in Google Scholar
• 123 Liu M, Luo Y, Yan J, Xiong X, Xing X, Kim JS, Zou T. J. Am. Chem. Soc. 2023; 145: 10082
  CrossrefPubMedSearch in Google Scholar