Despite the great potential of metal compounds, only a few have been finally accepted and marketed by the FDA because of toxicity and resistance induction, poor biodistribution, and short- and mid-term side effects
[32]. Metallodrugs’ toxicity is mainly related to renal damage and neurotoxic effects
[33], whereas resistance is due to targeting individual proteins or enzymes. On the other hand, the biodistribution and side effects could be improved by the further development of drug delivery strategies, such as nanoparticles
[32][34]. The steps for the clinical approval established by the FDA are as follows: (1) discovery and development, (2) preclinical research, (3) clinical research, (4) FDA drug review, and (5) FDA post-market drug safety monitoring (https://www.fda.gov/patients/learn-about-drug-and-device-approvals/drug-development-process, accessed on 12 July 2023). All of these steps aim to determine the safety and efficacy of the drug under development in humans. Specifically, the clinical trial begins with Phase 0, a preliminary study in which some healthy volunteers are studied with an administration dose of less than 1% of the therapeutic dose for a maximum of seven days, followed by Phase I. After these phases, the metal drug can be tested in affected patients, which is Phase II of the clinical trials. The efficacy of the drug is tested in patients, and a placebo is used as a control. The final step is Phase III, which accounts for the majority of the clinical trials’ cost, and aims to confirm the safety and efficacy tested in Phase I and Phase II, respectively
[35].
2. Protein Target Identification through Proteomic Approaches
The study of protein interactions with different classes of biomolecules has ancient roots in the field of proteomics, with many different biochemical strategies developed and advanced in conjunction with mass spectrometry techniques
[39][40]. In recent years, several proteomic strategies coupled with mass spectrometry (MS) have emerged as a powerful and systematic approach for the large-scale identification of drug–protein interactions and the elucidation of their associated mechanisms.
Chemical proteomics, or chemoproteomics, is primarily devoted to the study of protein–small-molecule interactions and is attracting increasing attention in drug target discovery
[44][45]. Chemoproteomics is a leading tool in the field of drug discovery that relies on affinity-based or label-free approaches to identify protein targets.
2.1. Affinity-Based Strategies
The affinity-based approach to chemical proteomics aims to immobilize the drug on a solid support and isolate the target proteins. This strategy is also referred to as drug pulldown
[46][47]. Pulldown strategies and, more generally, affinity purification-mass spectrometry approaches (AP-MS), are proteomic techniques that allow researchers to isolate multiprotein complexes starting from a known molecule to hypothesize the intracellular processes in which it is involved
[48][49][50].
Drug pulldown requires the following four main steps: (1) the immobilization of the pharmacophore on resin beads with affinity chromatography; (2) the isolation of the target molecules; (3) the identification of the target molecules with liquid chromatography–tandem mass spectrometry (LC–MS/MS); and (4) bioinformatics analysis to identify and quantify the proteins (Figure 1).
Figure 1. Drug pulldown consists of four steps: (1) immobilization of the pharmacophore on beads, which can be performed using different approaches (streptavidin beads and biotinylated metallodrugs or human serum albumin (HSA)–antiHSA antibody interaction); (2) target isolation, which can be preceded by a stabilization step involving UV-crosslinking between the proteic target and the drug; (3) mass spectrometry analysis of the target; and (4) bioinformatic analysis for the targets identification. Legend: M: metallodrug; B: biotin.
The advantage of pulldown techniques is that the physiological state of the proteins, the concentration levels (i.e., abundance), the post-translational modifications, and the natural binding partners are preserved by simply using non-denaturing lysis buffers. In particular, for Step 1 of the drug-pulldown workflow (
Figure 1), different methods can be used to immobilize the active pharmacophore, which also depend on the stability of the organic- and inorganic-metal complexes. For this reason, Babak et al. used a biotin/streptavidin approach to immobilize a Ru(II) compound that is classified as a member of the RAPTA family
[51]. The arene ligand of the complex was functionalized with a primary amine that can be biotinylated via an aminocaproic acid linker. The research results led to the identification of 15 cancer-related proteins that can explain the activity of RAPTA molecules, including
MK,
PTN, and
FGF3 as metastasis-related effectors, but also proteins related to cell cycle regulation, i.e.,
GNL3,
CGBP1,
FAM32A, and
VIR. Drug biotinylation can also be performed by using the click reaction (Cu(I)-catalyzed alkyne–azide cycloaddition (CuAAC)). CuAAC is a reaction between an alkyne and an azide group, catalyzed by Cu(I) ions
[52][53]. This reaction was used by Neuditschko and colleagues to derivatize the drug with biotin and fish the targets with streptavidin beads. Another clever method for the immobilization of metal drugs was reported by the same
[54] research group. A complex was formed between the Ru(III) compound (i.e., BOLD-100) and human serum albumin (HSA). Then, the authors used the anti-HSA beads utilized for the human serum depletion to form the BOLD-100-HSA-beads adduct used to isolate drug targets
[55]. Because high false positive rates occur in pulldown experiments, the authors also performed a competitive assay. The cell lysate was pre-treated with free BOLD-100 before exposure to the immobilized drug on the beads. This latter strategy allowed the saturation of the selective binding sites, as shown in
Figure 2, so that the resulting target profile contains only non-selective binding partners. The subtraction of these from the target profile obtained by normal pulldown deletes the non-selective interactors and provides a list of selective binding partners. The competitive pulldown experiment is an alternative way to validate and remove the possible false-positive targets from a canonical pulldown experiment. Using this approach, the authors identified ribosomal proteins and the transcription factor
GTF2I as BOLD-100 targets. In particular, further transcriptome analysis validated
RPL10 and/or
RPL24 among the ribosomal protein BOLD-100 targets.
Figure 2. Competitive pulldown workflow. (A) The total protein extract was incubated with the free metallodrug and then subjected to the chemoproteomic strategy. Only the non-specific binders were eluted. Instead, the specific target was not retained on the beads. (B) The total protein extract was incubated with the metallodrug-derivatized beads for the chemoproteomic strategy. The non-specific and specific proteins were eluted. The comparison between the eluted proteins in the competitive and non-competitive will highlight the false-positive content.
Although the simple pulldown strategy is a straightforward approach, it suffers from the classic drawbacks that are generally associated with affinity purification strategies, such as the lack of transient interactions and the problem of false positives, as mentioned above (see
[38] for specific further insights on this topic). For these reasons, the drug immobilization step is the most critical, as the final yield of target purification depends on it. In classical pulldown experiments with small organic molecules, the drug is usually covalently functionalized on an N-hydroxysuccinimidyl (NHS) sepharose carrier, then the non-functionalized portion of the carrier is blocked using several washes with ethanolamine
[45]. For metal-based drugs, these types of washes are not compatible, because ethanolamine can compete with the metal complex ligands. Therefore, alternative methods have been developed, for example, using the biotin–streptavidin interaction (as described previously). The metal complex site carrying the label must be carefully selected in order to avoid interference with the biological activity of the pharmacophore. The CuAAC strategy is one of the best known for metal complex biotinylation and offers the advantage of high reaction efficiency and a favorable reaction time under mild conditions. However, a key drawback is the toxicity of copper ions to proteins. They can damage the structure and function of the target proteins and promote the formation of ROS
[56]. In addition to the clickable part, i.e., the alkyne, a photo affine group can often also be produced
[57]. Indeed, this strategy can be used to stabilize transient interactions, and it is also possible to stabilize the interactions in living cells, i.e., photoaffinity labeling (PAL). The photoactivatable group, usually a diazirine or benzophenone group, has the function of facilitating covalent adducts with the metal/metal-drug-interacting proteins upon light irradiation, as reported by Liu et al.
[58]. The authors synthesized a photoactivatable molecule that enables the probe to form covalent adducts with the proteins interacting with the metal drug upon irradiation with UV light
[59]. The authors incubated HeLa cells with a Au(III) meso-tetraphenylporphyrin (gold-1 a) compound and irradiated them at 365 nm to activate the benzophenone reaction with proteins via a radical mechanism. After cell lysis, an alkyne group on the probe was used to functionalize the protein–drug complexes with biotin in order to enrich the adducts on streptavidin beads. In their study, they identified heat shock protein 60 (Hsp60) as a target of gold-1a. These approaches can be used to identify interactions characterized by non-covalent and coordinative bonds. The covalent detection of drug–target interactions in cells can be used to identify physiologically occurring and moderately strong binding events by affinity-based chemoproteomics in vivo
[60][61].
2.2. Label-Free Approaches
On the other hand, label-free approaches exploit protein stability due to the drug interaction.
For example, thermal proteome profiling (TPP)
[62] has been used to study metallodrug targets
[63]. TPP measures the extent of the drug–target interaction by monitoring the effects of pharmacological treatment on protein denaturation/solubility as a function of a progressive increase in temperature.
The extent of protein stabilization by the small molecule is proportional to the strength of the interaction and can be assessed by multiplexed mass spectrometry analysis (see
[64] for a comprehensive description of all TPP applications). As shown in
Figure 3, TPP can be performed by varying the temperature range (TR-TPP) (
Figure 3A) or drug concentration (CR-TPP) (
Figure 3B). After protein digestion and mass spectrometry analysis using LC–MS/MS, proteins are relatively quantified by the comparison between the drug-treated and drug-free conditions. The soluble fraction for each protein is reported as a function of the temperature or drug concentration.
Figure 3. (A) TR-TPP experiment. The protein extract is incubated with both the vehicle and the metallodrug. Ten aliquots are prepared for each condition. Each aliquot is heated at a specific temperature, and then the soluble fractions are digested with trypsin and labeled with a TMT isotope tag. The samples are analyzed with LC–MS/MS and the protein is identified. Melting curves are fitted and melting temperatures are calculated for all proteins. (B) CR-TPP experiment. Untreated protein extract, as a control, and those treated with nine different concentrations of metallodrug, are heated to the same temperature. The soluble fractions are then digested with trypsin and labeled with a TMT isotope tag. The samples are analyzed using LC–MS/MS and proteins are identified and quantified. Dose–response curves are fitted, and thermal stability parameters are calculated. Legend: M: metallodrug.
TR-TPP was recently applied to discover the antitumor Bis(
N-Heterocyclic Carbene)Pt(II) complex targets in intact cells, leading to the identification of asparagine synthetase (
ASNS)
[65]. Another label-free method used in the study of metallodrug targets is the functional identification of target by expression proteomic (FITExP) approach
[66]. The basic principle is that protein targets and major mechanistically related proteins are modulated upon prolonged drug exposure; in particular, they are overexpressed when a lethal concentration (LC
50) is administered to the cells. Lee et al.
[67] applied the FITExP method to evaluate the mechanism of action of Ru(II) complexes, RAPTA-T, and RAPTA-EA. Although the approach suggested many protein targets, the best results were obtained in terms of revealing the mechanism of action of the molecules. For example, RAPTA-EA and RAPTA-T were found to induce the overexpression of several oxidative-stress-related and tumor-suppressing proteins, respectively
[67]. FITExP analysis of proteins extracted from cells treated with RAPTA-T revealed, among the proteins identified, PLD3 protein, which belongs to the phospholipase D enzyme (PLD) family, with the lowest
p-value. Instead, cells treated with RAPTA-EA had heat shock protein 1A/1B (
HSPA 1A/1B) as the protein with the top-ranked
p-value. This approach allowed the authors to hypothesize the following two different mechanisms of action for the two drugs: RAPTA-T targets are involved in multiple processes, suggesting a broader mechanism of action of RAPTA-T. In contrast, the targets of RAPTA-EA are exclusively involved in the regulation of the oxidative stress response. The combination of TPP and FITExP was performed in the work of Saei and colleagues to reduce the number of false positives and false negatives
[63]. In principle, the other label-free strategies used in the study of protein–small-molecule interactions could also be applied to the field of metallodrugs. These methods encompass the protein stability based on the oxidation rate (SPROX),
[68] pulse proteolysis (PP)
[69], drug affinity responsive target stability (DARTS)
[70], and limited proteolysis-coupled mass spectrometry (LiP-MS)
[71]. However, no work has been published.
Label-free techniques are capable of detecting the transient interactions of drug candidates with potential protein targets and globally assessing drug-altered temperature and the proteolytic or chemical stability of proteins, and otherwise do not require synthetic modification of the drug
[72][73].