Metallodrugs are unique: opportunities and challenges of discovery and development

Metals play vital roles in nutrients and medicines and provide chemical functionalities that are not accessible to purely organic compounds. At least 10 metals are essential for human life and about 46 other non-essential metals (including radionuclides) are also used in drug therapies and diagnostic agents. These include platinum drugs (in 50% of cancer chemotherapies), lithium (bipolar disorders), silver (antimicrobials), and bismuth (broad-spectrum antibiotics). While the quest for novel and better drugs is now as urgent as ever, drug discovery and development pipelines established for organic drugs and based on target identification and high-throughput screening of compound libraries are less effective when applied to metallodrugs. Metallodrugs are often prodrugs which undergo activation by ligand substitution or redox reactions, and are multi-targeting, all of which need to be considered when establishing structure–activity relationships. We focus on early-stage in vitro drug discovery, highlighting the challenges of evaluating anticancer, antimicrobial and antiviral metallo-pharmacophores in cultured cells, and identifying their targets. We highlight advances in the application of metal-specific techniques that can assist the preclinical development, including synchrotron X-ray spectro(micro)scopy, luminescence, and mass spectrometry-based methods, combined with proteomic and genomic (metallomic) approaches. A deeper understanding of the behavior of metals and metallodrugs in biological systems is not only key to the design of novel agents with unique mechanisms of action, but also to new understanding of clinically-established drugs.


Tables cited in main text
. Medicines containing a metal in their active pharmaceutical ingredient Table S2. Composition of RPMI-1640 cancer cells growth medium Table S3. Composition of foetal bovine serum (FBS) Table S4. Common media used for in vitro screening of antimicrobial activity Table S5. Some analytical methods for metallodrug speciation Table S6. Cisplatin cytotoxicity assays Table S7. Examples of target classes for identification of off-target effects Table S8. Cell death pathway inhibitors Table S9. Different stages of cell cycle arrest for ruthenium anticancer complexes Table S10. RNAi studies for cisplatin Table S11. Characteristic electron-binding energies and X-ray emissions of elements Table S12 NMR parameters for commonly used nuclei in metallodrug speciation Table S13 EPR parameters of some common metallodrug paramagnetic nuclei Table S14 Summary of cell and tissue preservation methods

S2. Analytical techniques for metallodrug speciation
Several common analytical techniques have metal-specific features, which can be used for characterising metallodrugs and their metabolites and detecting them in biological media. The discussion here is not exhaustive, and certain important techniques (e.g. Mössbauer spectroscopy) will not be discussed. The combined application of a range of them is usually necessary.

S20
NMR. As a semi-quantitative, non-destructive technique probing both covalent and long-range non-covalent interactions, NMR spectroscopy is widely used for metallodrug speciation, often for solutions of isolated metallodrugs or complexes with specific biomolecules. It can provide information about multiple species simultaneously without need for prior separation, allowing analysis of cell lysates or other biological materials. 24 However, it is relatively insensitive for physiologically-relevant drug concentrations, requiring tens of micromolar for even the most sensitive nucleus 1 H ( 3 H is more sensitive, but radioactive).
Advanced EPR methods can probe metal-ligand bonds through hyperfine interactions, including electron-nucleus double resonance (ENDOR), electron spin echo envelope modulation (ESEEM), and 2D hyperfine sublevel correlation (HYSCORE). 30 ENDOR has been used to deconvolute hyperfine interactions between 1 H, 14 N, and Ru(III) in NAMI-A, which confirmed coordination of NAMI-A to HSA through His residues. 31 EPR measurements are possible for biological materials; EPR of fractionated lysates of Saccharomyces cerevisiae cells showed that KP1019 was more localised in cell walls compared to NAMI-A, which localises mainly in the cytoplasm and mitochondria. 30 Vibrational spectroscopy. Vibrational spectroscopy involves relatively easy, label-free sample preparation with short data collection times. However, many complexes are studied in isolation as a powder, pellet, or evaporated solution, which limits the applicability.
Computational studies are often needed to assign fully the observed vibrations.
The exclusion rule allows distinction between cis and trans isomers of metallodrugs. Trans The metal itself may be a chiral centre, or chirality may come from ligands. Induced chirality in metal complexes can arise through interactions with chiral biomolecules, a powerful way of using metal absorption bands to study such binding.
Although CD cannot provide detailed structural information, it is fast and requires very little sample. CD measurements are largely confined to in vitro solution samples, as unfiltered S23 samples of cell lysates or biological fluids often contain too many chiral biomolecules to study any one specific molecule or interaction.
The CD spectrum provides a fingerprint for a particular enantiomer. 35 Such information can be extremely important for a drug since enantiomers may have differing biological activity if they involve reactions with chiral biomolecules. For example, the R,R isomer of oxaliplatin is more active than its S,S counterpart, 36 and anticancer Ru and Ir complexes have enantioselective binding affinities for target proteins such as glycogen synthase kinase 3 (GSK-3), carbonic anhydrase II, histone deacetylases, and trypsin, affecting their activity as inhibitors. 37,38 Mass spectrometry. The necessary ions are usually produced using electrospray ionisation, which allows retention of metal-bound ligands in the gas phase. 39 Figure S1). Due to different nuclear binding energies, these ions have a slightly different mass (ca. 4 mDa mass difference for OCS), and this provides strong evidence for a postulated molecular formula. 45 All the benefits outlined here also apply in tandem MS experiments, in which a precursor ion is isolated and fragmented in the gas phase, e.g. determination of oxaliplatin-treated ubiquitin Ion mobility MS. Ions in IM spectrometry are separated by their electrophoretic mobility, dependent on their collision cross-section (CCS) as they move through a collision gas (e.g.

Ru17)
. 48 From reactions of [( 6 -bip)Ru(en)Cl] + with d(CACGTG), IMS can separate several conformations of mono-and di-ruthenated species on the basis of the change in size and shape. 49 Conformational changes in platinated proteins have also be detected by IMS, e.g. for ubiquitin adducts with cisplatin. 46,50

S3. Methods for cell preservation
Chemical fixation. One of the most common methods of cell and tissue preservation for biological research is chemical fixation: using chemicals to prevent enzymatic autolysis.
Chemical fixatives can be broadly categorized as crosslinking, precipitating, or oxidizing fixatives. Crosslinking fixatives include formaldehyde and glutaraldehyde, which bind covalently to proteins -preserving both chemical and structural integrity. 51 Precipitating fixatives work by reducing protein solubility (cell dehydration), causing precipitation of proteins. 52 In particular, ethanol and methanol are precipitating fixatives which are commonly employed in flow cytometry and confocal microscopy. Finally, oxidizing chemical fixatives (e.g. osmium tetroxide) react with biomolecules and proteins to stabilize cellular or tissue structure. 53 The efficiency of chemical fixation depends on a variety of factors: pH, time, temperature, osmolarity, penetration rate, and volume ratio. 54 The advantages of preserving biological samples with chemical fixatives include rapid preservation, good preservation of cellular morphology, suitable tissue penetration, and low expense. However, chemical fixation can be disadvantageous as it can alter endogenous factors (i.e. cell volume, ATP content), 55 as they typically bind to proteins (which may alter the observed drug distribution), have detrimental effects on DNA and can be toxic (Table S14). 55  Co(II) 3/2 (high-spin) 59 Co, I = 7/2 (100%) 10 11 -10 12 Co(II) 1/2 (low-spin) 59 Co, I = 7/2 (100%)