|
Metals in Medicine (Part 2)
by National Institute of General Medical Sciences
Session 2: Metal-Containing Targets of Drug Action
John Kozarich discussed drug development programs involving two metallopeptidase targets, metallo beta-lactamases (a.k.a., carbapenamases) and peptide deformylase. Metallo beta-lactamases (as opposed to the conventional serine protease type enzymes) have appeared more recently and pose a threat to last line of defense antibiotics, such as imipenem. They are dinuclear Zn enzymes with a bridging water molecule. The Benkovic lab and Lippard lab have contributed studies of the enzyme and active site models, respectively. Merck has used high-throughput screening to identify lead compounds. A group of biphenyl tetrazoles were identified that had originally been synthesized as part of an angiotensin receptor antagonist development program (a non-metal-containing target; however, see Abrams' talk). Another class was developed from a benzyl succinate lead. The most potent candidate presented, a p-phenyl substituted dibenzyl succinate, has an IC50 in the 100 pM range and strongly potentiates imipenem action in resistant cells. High-resolution crystal structures show that in both cases the inhibitor is bound to the Zn sites. In the succinate derivatives, both carboxylate groups are liganded to Zn with one bonded to both Zn atoms, replacing the bridging water molecule.
Peptide deformylase and methionine amino-peptidase are required for sequential removal of the formyl group, then the methionine group, of the translation initiator f-Met in prokaryotic protein biosynthesis. Peptide deformylase is a member of the thermolysin family, containing a single Zn atom coordinated by the HEXXH motif and a cysteine residue. Actinonin is an acyl-hydroxamate antibiotic that was discovered in the 1960s. A crystal structure demonstrates that it binds to the Zn site. Experiments have demonstrated that its mechanism of antibiotic action is indeed inhibition of this enzyme. Unfortunately, resistance can be acquired at very high frequency by mutations which eliminate formylation of the initiator met-tRNAf-met . Therefore, it does not appear that peptide deformylase is in fact a useful antibiotic development target. The methionine amino-peptidase, which is also a metalloenzyme, may be more promising.
This work illustrated:
- The contributions of basic research on metalloenzyme structure/function from several NIH-supported labs to a pharmaceutical company research program
- Compounds developed for other purposes and not targeted toward metal-containing enzyme sites may in fact also interact with metalloenzyme active sites.
- The importance of validating drug targets before investing too much effort in discovering new inhibitors
- The utility of both screening approaches and structure-based rational drug design in the optimization metalloenzyme inhibitors
Tom Poulos presented recent work from his lab and others on the structure and function of nitric oxide synthase. This research seeks to understand the electron transfer process between the two flavins, tetrahydrobiopterin, and heme cofactors per monomer; regulation of the three different isoforms; the reaction mechanism; and to develop isoform selective inhibitors. Crystal structures have been obtained for the heme domains of all three isoforms (eNOS, nNOS, and iNOS). The enzyme forms a dimer stabilized by a tetra-cysteine coordinated zinc atom at the interface. Although spectroscopically and mechanistically related to P450, the NOS fold is different. All three structures are similar (rmsd = 1.06 Å). The active sites are nearly identical (except for an asp/asn substitution at one site). Sequence variations in the pterin site are somewhat greater. Structures have been obtained for a variety of bound arginine and pterin analogs, and other inhibitors. H-bonding and ion-pairing with glutamate and also one of the heme propionate groups hold the arginine substrate in close proximity to the heme iron. The pterin is located near the dimer interface, but also binds to the same heme propionate group. Arginine analogs such as nitro-arginine bind with the nitro-group extending over and blocking oxygen access to the heme. Isothioureas bind in place of the arginine guanidino group, and utilize similar H-bonding and ion-pairing contacts. Some mechanism-based (suicide substrate) inhibitors have been discovered. Several of these cause the heme to oxidize itself, yielding biliverdin, others cause protein modifications. 2-bromo-7-nitro-indazole appears to bind as an electrostatic analog of the guanadino group, but causes changes in position of the propionate group that also result in loss of pterin binding and binding of a second molecule of inhibitor at the pterin site instead (in iNOS, but not in nNOS). Most other inhibitors bind without causing major structural changes.
Future directions:
The active sites are relatively open, and the sequences diverge more, farther away from the active site. It may be possible to exploit secondary contacts outside the site (e.g., neighboring histidine residues) to generate selective inhibitors. Site-directed mutagenesis and rational drug design will be useful to test the utility of targeting the pterin site. Few inhibitors have been found to function by binding directly to the heme iron atom. This may be a way to generate binding affinity, but since the sites are so similar, may not be a useful strategy to generate isoform selective inhibitors.
Ken Merz provided a review of molecular modeling methods applicable to metal-containing systems. Issues include: i) appropriate electrostatic representation of metals, ii) representation of long-range electrostatic effects, iii) inclusion of dynamic behavior, iv) ligand exchange, v) calculation of complex electronic structures, and vi) representing environmental "solvation" effects. Modeling of metal sites in proteins is a different problem than modeling the behavior of the overall metalloprotein; however, both depend on how accurately one needs to represent the metal and its interaction with the protein. Purely molecular mechanical bonded model approaches may be adequate for fixed coordination, structural questions, but are not adequate for questions about energetics, reactivity, and coordination changes. Local environmental effects, solvation, dynamics, etc., significantly alter charge densities so that fixed formal charges are inadequate. For example, calculations for the zinc atom (formal charge = +2.0) in carbonic anhydrase show variations between a charge of +0.6 and 0.0 on a psec time scale. Combined quantum mechanical/molecular mechanical methods that use QM to model the metal site and MM methods to model the surrounding protein provide improved results. The appropriate way to interface these calculational regimes remains an area of research. Problems include how to handle boundary effects, finite size of the QM region, neglect of long-range electrostatic interactions, choice of semi-empirical QM method, limitations of the gas-phase QM calculation (e.g., removal of active site geometric constraints) Fully QM and QM/QM approaches have been developed using density matrix minimization, localized orbitals, divide-and-conquer or onion-skin approaches, and parallel computation. These methods have provided accurate calculations for systems as large as 5,000 atoms, but will not allow MD simulations.
In discussion, the question of accessibility of methods and the utility of commercial software programs for the average chemist was raised. Most off-the-shelf programs are not parameterized for metals. Docking programs for looking at substrate-protein interactions mostly do not include good ways to handle small-molecule ligand binding to metal sites. This would seem to be an opportunity for small businesses to make a real contribution to the field. Additional ways to validate computational results beyond basic structural questions are also needed.
Benjamin Burke provide a discussion of structure-based drug design in the development of matrix metalloprotease inhibitors. These zinc-containing enzymes have recently been among the hottest pharmaceutical targets (e.g., relevant to basement membrane invasion and metastasis, tumor angiogenesis, and connective tissue breakdown in arthritis). Approximately 23 MMPs are known, including gelatinases, collagenases, thermolysin, stromalysins, and membrane-bound proteases. They are regulated by a self-inhibitory pro-domain, activated in cascades by other MMPs and other proteases, transcriptionally regulated by cytokines and growth hormones, and subject to inhibition by tissue inhibitors (TIMPs). Although differing in substrate specificity, the MMPs are structurally homologous, containing: an N-terminal self-inhibitory domain, a C-terminal specificity domain, and a catalytic Zn (and Ca) domain. The alpha/beta sandwich catalytic domain is fully active. Two calcium atoms and a structural zinc site contribute to the stability of the structure. The active site zinc is coordinated by three histidines with a nearby glutamate residue essential for activity. Backbone carbonyl groups are also involved in both the catalytic mechanism and inhibitor binding. Inhibitors have targeted the selectivity specifying P1' binding sites, while also providing zinc ligation by oxygen, nitrogen, and/or sulfur atoms. Hydroxamate, reverse hydroxymate, carboxylate, phosphinate, amino carboxylate, and sulfodiimine compounds have been developed by several groups. Both tetrahedral and bipyramidal (TS mimic) coordination have been observed. Electrostatic and H-bonding interactions at other sites, shifts in pKa, and strain, influence the crystallographically observed distortion from ideal geometries and binding affinities. Calorimetry, NMR, and calculational methods were used to examine the energetics and suggest the next compounds to make. MMP inhibitors were shown to inhibit tumor growth and neovascularization. Side effects in cancer therapy (e.g., joint pain) were reduced by minimizing activity against MMP1 through P1' site moiety design.
Future directions:
The number of MMPs is not fully known. Regulation and activation cascades are not fully understood. Knockout mouse models may not be possible in all cases; therefore, specific inhibitors will be useful tools to sort this out. Ligand coordination modeling in metalloproteins is still difficult. Energetic calculations need to be improved to account for changes in H-bonding, protonation, pKa, torsional strain, and charge on zinc. It is difficult to make comparisons across compound classes. Earlier fundamental studies on the mechanism of thermolysin were useful. Additional studies on the basic chemistry of small chelators focused on pH effects would be useful.
Session 3: Radiology, Imaging, and Photodynamic Therapy
Shuang Liu provided a discussion of radioimaging agents. 99mTechnetium (t1/2 = 6.02 hr, 140 keV g-emission, readily available, and low cost), is the radioisotope of choice. Two strategies were discussed: i) the integrated approach, where the metal complex, itself, contains targeting information to provide localization of the radiolabel, and ii) the bifunctional chelator (BFC) approach, where the chelator provides a binding site for the metal and a means of attaching the radiolabel to some targeting moiety (e.g., antibodies, peptides, peptidomimetics, non-peptide small molecules). Illustrative targeting moieties used by Dupont, Diatide, various other companies, and academic researchers include: somatostatin analogs, RGD peptides, LTB4 antagonists, folate antagonists, fMLFK chemotactic peptide, and vitronectin receptor antagonists. Generally, antagonists are preferred because free ligand causes fewer side effects than with agonists. Tc-cores have been based of the chemistry of Tc(III), Tc(IV), [TcO]+3, [TcO2]+, [TcºN], [Tc(CO)3]+, and [TcN=N-R]. A wide range of linker and chelation chemistry has been employed.
Important issues in radiopharmaceutical design include: selection of Tc-core and BFC strategy, choice of targeting molecules, ease of synthesis, stability to target linkage and radiolabeling steps, immunogenicity, pharmacokinetics, and characterization methods, including attention to possible isomerism. Pharmacokinetic preferences include: fast receptor binding, slow dissociation, fast blood clearance, short blood residence time, and renal clearance. Selection from a variety of cationic, anionic, neutral, and enzymatically cleavable linker units can tailor the pharmacokinetic properties. Metal chelates must be thermodynamically stable and kinetically inert, should minimize the formation of isomers, possess high labeling efficiency, and have high hydrophilicity, which facilitates renal versus hepatobiliary excretion. Development depends on a better understanding of both biology and coordination chemistry (particularly at tracer levels of material), and the application of appropriate analytical methods.
Carolyn Anderson provided a discussion of radiation therapy using targeted metal complexes. Advantages include: i) administration of higher radiation doses to achieve greater efficacy; ii) matching the type and energy of the ionizing radiation delivered (e.g., b, a, or Auger electron) to the size of the tumor, and iii) use of the same or analogous agents for imaging and therapy to allow prediction of tumor response and toxic side effects. Some of the metals used in targeted radiotherapy include: Re, Y, In, Lu, and Cu. Examples were presented based on somatostatin analogs linked to DTPA, TETA, or DOTA chelators. Several molecules from various academic and corporate programs are currently in Phase I/II clinical trials. Promising results have been achieved with minimal overall toxicity. Dose-limiting renal toxicity is due to radiation exposure, not to metal exposure per se. Cu-64 decay modes are useful for both PET imaging and therapy, and the isotope can be economically produced using local facilities. Progress in the Washington University program on Cu-64 somatostatin analogs includes: synthesis, receptor binding assays, intracellular localization, biodistribution and targeted therapy in tumor-bearing rats, in vivo metabolism, toxicity studies in rats, dosimetry in rodents and primates, and a clinical PET imaging trial. Useful diagnostic imaging results in humans and promising therapeutic effects in animal models have been demonstrated. The mechanism of action appears to involve receptor-mediated endocytosis, dissociation of the metal from the chelator, and delivery to the nucleus, where isotopes such as Cu-64 that emit Auger electrons as well as other radiation were shown to be particularly effective.
Considerations in targeted radiotherapy include:
- Tumor type--which receptors/antigens are present and which molecules will target the tumor?
- Tumor location and effect on normal tissue uptake and radiation exposure
- Tumor size and decay scheme of radionuclide
- Matching of radiopharmaceutical and radionuclide half-life
- Cost and availability of isotope
Challenges include the development of thermodynamically stable and kinetically inert complexes for a wide variety of radiometals, retention of biological targeting activity during synthesis, optimization of tumor uptake and minimization of non-target tissue effects, understanding how small changes in radiometal or BFC affect biodistribution, understanding how intracellular metabolism and trafficking affect delivery of metals to the nucleus, and understanding better the mechanism of cell killing.
Randall Lauffer provided a discussion of MRI contrast agents. The current market is about $600 million and has almost doubled since 1994. MRI contrast agents offer several advantages over other imaging methods. They are non-radioactive, and administered at concentrations of 0.1 to 1 mM, whereas iodine-based x-ray contrast agents require 1-10 mM and radiation exposure. Functional components of an imaging agent [e.g., Schering AG's Magnevist (Gd-DTPA)] are a Gd ion (large S value) held tightly by a chelating agent to prevent Gd toxicity, and may also include a targeting moiety. Signal is derived from the Gd ion and from proton signals of spin-coupled water molecules. The biophysical properties of the complexes (electron spin relaxation, rotational correlation, and exchangeable water residence times) can be tuned to match medical imaging needs. Development issues include achieving sufficient relaxation rates changes in the target tissue, and safety (non-toxic, readily excreted, and stable complexes). Other practical concerns include cost, administration, viscosity, and timing. EPIX has been developing non-covalently targeted MRI agents. Potential advantages are: i) low retention of gadolinium in the body, ii) low immunogenicity, iii) tunable binding affinity to allow a convenient imaging time window, and iv) higher magnetic efficiency (protein binding was shown to increase relaxation changes 10x). MS-325 is a diphenylcyclohexyl-phosphate derivative of Gd-DTPA designed to target serum albumin as a blood pool marker. It is being tested in MRI cardiac angiography and as a monitor of tissue ablation in focused ultrasound surgery (protein denaturation results in dissociation of the marker and thus changed relaxation). Another agent has been developed as a pro-drug with a phosphate group that blocks protein binding until acted upon by a phosphatase. This agent may be useful for imaging in the lymphatic system A third agent was presented that utilizes a fibrin binding peptide to image clot formation.
Future research opportunities in the area of MRI contrast agents include:
- Improved understanding of lanthanide coordination chemistry
- Development of new chelates
- Analysis of structures and spin dynamics
- Understanding of structure activity relationships for water exchange and T1 relaxation rates.
Regarding practical problems in developing MRI agents, the following points were made. Academic research is the cradle of creativity, but translating ideas to the marketplace can only be done by industry, and requires lots of money. However, finding translatable licensing opportunities is difficult. Advantages of the industrial research setting are highly motivated employees, focused multidisciplinary teams, careful treatment of proprietary information, and regulatory expertise. All this takes a lot of money. Suggestions included increased funding for basic and applied research, and promotion of responsible treatment of proprietary information. The present limited level of SBIR funding is not sufficient to accomplish the job and the turnaround time is too slow. Increasing the level of funding for companies to at least $4 million over two years and decreasing review time to three months with minimal disclosure of proprietary material was suggested.
Richard Miller presented work from Pharmacyclics that derives from the research of John Sessler on expanded porphyrins (Texaphryins). These include gadolinium motexafin (Xcytrin, a radiation enhancer), and lutetium motexafin (Lutrin, a photosensitizer). The actions of these complexes are dependent on their redox and photochemical properties, in the first case through depletion of cellular reductants, in the other case through both depletion of cellular reductants and generation of singlet oxygen. Both compounds are selectively concentrated in tumors and localize in the mitochondria of cells, where they may have effects on cellular bioenergetics. Radiation therapy is thought to act through generation of hydroxyl radical and superoxide-mediated DNA damage. Cellular reductants protect the cell by quenching initially formed DNA radicals. Xcytrin has been shown to catalyze ascorbate and NAD(P)H oxidation in vitro and to lower reductant levels in cells. Enhanced apoptosis has been demonstrated after radiation treatment in a tumor bearing mouse model. Clinical trials have demonstrated enhanced benefits of radiation therapy with Xcytrin co-therapy, particularly in treatment of brain metastases derived from lung and breast cancers. Localization of the drug can be monitored by MRI due to relaxation effects of the Gd atom, and the sensitivity of tumor imaging is enhanced. Normal tissue uptake was limited to liver and kidney. Dose-limiting renal toxicity was reversible. Blood clearance was rapid, but the drug accumulated in target tissue such that efficacy and also imageability of the tumor increased with multiple doses. Lutrin has shown efficacy in cutaneous breast cancer. Limiting toxicity was due to skin necrosis and pain in the treatment area. The ability to deliver light internally through catheters, plus the ability of long-wavelength light to penetrate tissue means that photodynamic therapy (PDT) is not limited to the skin. NCI-sponsored clinical trials are in progress for PDT of prostate cancer, cervical neoplasia, and esophageal cancer. Other applications of Texaphrins are being developed for atherosclerosis and ophthalmologic applications.
Research objectives in this area included:
- Exploitation of the unique redox and photochemical properties of metals
- Modification of the porphyrin nucleus to bind metals with a larger radius than that of iron
- Optimization of drug delivery and tumor localization
- Red shift of absorption wavelength to increase tissue penetration in PDT
- Pharmacokinetic optimization of dose and radiation (or light irradiation) timing protocols.
Bruce Averill gave a brief presentation of recent results, including crystallographic studies, on proteolytic activation of purple acid phosphatase, a diiron enzyme. In the unactivated state, a surface-loop asparate residue is postulated to interact with either a histidine or asparagine near the Fe+2 site, and reduces the nucleophilicity of hydroxide coordinated at the Fe+3 site. Proteolytic cleavage eliminates these interactions.
Tom Meade presented a short video on the use of MRI contrast agents to visualize gene expression and cell signaling activity. These experiments are based on agents that are capped to prevent water exchange (and hence water relaxation effects) at the Gd site until modified by enzymatic cleavage or ligand binding. They have been used to map the anatomical expression of genes that are important in development.
Day 1--Afternoon Discussion
Discussion focused on the reasons that more inorganic agents have not been developed as pharmaceuticals. The reality is that most chemists working for companies are organic chemists by training and naturally look in that direction for leads. There is a perception, right or wrong, that metal-containing agents are more toxic. Metal-containing agents have generally been excluded from the high-throughput library screening programs that have become the major paradigm for drug lead discovery. Metal-based drug discovery may be more amenable to a rational design approach. Industry has been involved in this area in the past and remains very active in some specialty areas. If sufficient potential benefits can be identified, there would be a willingness to explore metal-containing leads.
Discussion recapitulated the question about the quality of metal site modeling and the need for better programs and more easily accessible properly parameterized programs. Discussion also emphasized the continued need for basic research into the coordination chemistry of metals and SAR for ligand control of metal properties. Such work is applicable in all areas of metallobiochemistry and medicinal inorganic chemistry.
Day 2--Morning Session
Session 4: Metal Metabolism
Tom O'Halloran discussed the emerging view that cells are not bags of buffer in which metals freely diffuse between thermodynamically controlled binding sites. Rather, a series of membrane metal transporters, metal chaperones, and assembly complexes have recently been identified that regulate the uptake and delivery of metal ions to specific sites. Despite reasonably high total concentrations (0.02-0.75 mM for Cu, Zn, Fe), the concentration of free metal is vanishingly small (less than one free atom per cell). In this context, thermodynamics of metal binding (which can be very tight, e.g., Kd up to 10-40 for mercury binding to vicinyl thiol groups), gives way to kinetics where dissociation rates can, nonetheless, be rapid. Specific examples were presented based on work in collaboration with the labs of Culotta, Penner-Hahn, and Rosenzweig, on the exchange of copper between the Ctr1 uptake pump, Atx1 chaperone, and ccc2 vesicular pump of yeast. Structural studies suggest an exchange mechanism in docked complexes involving a sequence of di- and tri-thiol coordinated intermediates. A similar pathway has been shown for delivery of copper to Cu,Zn-SOD via another chaperone (CCS) that, itself, includes an Atx1-like domain. Exchange in this case is not thermodynamically driven (Keq » 1.5), and both are tight-binding. However, the exchange is rapid, kf >> 30
M-1sec-1. The hypothesis is that protein/protein interaction lowers the activation barrier for dissociation by distortion of the metal binding site. Regulation of zinc in cells also appears to result in low free concentrations. In this case, zinc-containing vesicles are important. Fluorescent sensor compounds (Zinquin and derivatives) have been used to image zincergic neurons in the hippocampal region of the brain and the appearance of vesicles in other cell types as well.
Opportunities and obstacles:
- Metal trafficking is emerging as a new area of cell biology and a potential site for therapeutic interventions.
- Improved metal ion selective sensors (particularly ratiometric fluorescent probes) are needed in varying affinity ranges to study metal concentrations and fluxes in cells.
- Improved understanding of metal metabolism may be relevant to treatment of diseases such as Menkes and Wilson's diseases, and ALS.
Liz Theil provided an overview of iron metabolism, then focused on the regulation of iron-responsive protein synthesis by iron response element (IRE) recognition by IRE-binding proteins (IRPs). IREs have been identified in iron storage, transport, and utilizing enzymes (e.g., ferritin, transferrin receptor, DMT1, ferroportin, d-ALA synthase, m-aconitase). Multiple IRPs have been identified. These are themselves iron-binding proteins and may have other cellular functions (e.g., cytoplasmic aconitase). Many are regulated by cellular kinases as well as iron levels. Various IRPs have differential effects on various IRE-containing proteins such that a given iron level produces multiple partial activations. Understanding iron regulation is relevant to diseases of iron accumulation (e.g., sickle cell anemia, thalassemia, hemochromatosis) and also to diseases of iron deficiency, that together may affect 20-25% of the population. An objective is to increase translation of ferritin from the large untranslated mRNA pool to enhance iron storage capacity. IREs are hairpin loop structures that may occur at either 3'- or 5'-ends of mRNA. They may affect ribosome binding, message stability, or both. Details of IRE structure/function, dynamics, and stability have been studied by NMR spectroscopy and metal-based cleavage agents (e.g., Cu(phen)2, Ru(tipy)(bipy), Rh(phen)(phe)), which result in selective modification of only certain IREs based on shape complementarity. Studies on the structure and function of the iron storage protein ferritin were described with the aim of facilitating iron chelator removal of trapped iron. Ferritin is an oligomeric protein encapsulating a nanoscale volume in which mineral iron-oxide is reversibly deposited. Mutational studies of the proposed iron-entrance/exit channel suggest that large scale opening/closing motions may be important in function (e.g., perhaps for docking of iron delivery molecules).
Conclusions:
- Messenger RNA is a good drug target.
- Three-dimensional structures are stable enough for tight recognition and binding.
- Cell specificity of mRNA expression, in contrast to DNA, enhances biological selectivity.
- Backbone cleavage can provide high efficacy and information on the binding site.
- Targets are present in many EST databases.
- mRNA translational regulation is emerging as relatively common mechanism of control.
- IREs are ideal for selectivity studies to model other RNAs.
Ken Raymond provided a discussion of iron chelation therapy. The daily iron requirement for humans is low, in part, because there is no excretion mechanism. Iron levels are regulated by duodenal absorption. However, several diseases result in iron overload. Hemochromatosis is the most common hereditary disorder in the U.S. and results in excess dietary iron absorption. Accumulation of the excess iron in the liver leads to cirrhosis and hepatocarcinoma. Thalassemia(s) result from defects in hemoglobin chain synthesis, requiring frequent transfusions of normal hemoglobin carrying red blood cells. Excessive iron accumulates in the reticuloendothelial cells of the spleen. Thalassemia is widespread in the tropics. Dietary management and phlebotomy were the principal ways to reduce iron loads until introduction of desferrioximine B chelation therapy in the last decade. Unfortunately, this drug is slow acting, rapidly excreted, and is not orally active. Bacterial siderophores (iron transport molecules) have provided leads for chelator development, including the discovery of desferrioximine. Others include enterobactin, cepabactin, and deferipron, which utilize catechol, hydroxypyrimidate, and hydroxypyridinone chelating groups, respectively. Dr. Raymond has created a series of hydroxypyridinone derivatives that improve efficacy and lower toxicity of this class by combining multiple chelating units in a single molecule [e.g., TREN-(Me-3,2-HOPO)]. This creates a hexadentate ligand that binds iron much more tightly by forming a 1:1 complex, as opposed to the 3:1 complexes formed by bidentate ligands, such as desferroximine and deferipron. The new compounds have been successfully tested in mouse and rat models of iron overload.
Conclusions from these studies include:
- Thermodynamically predicted greater effectiveness of higher dentate chelators was confirmed by animal testing results.
- The multidentate chelators showed low toxicity and no significant removal of iron from normal animals (i.e., they did not remove iron from essential sites).
- Iron removal was equal to or better that for desferrioximine, the currently used therapeutic, when both were injected.
- TREN-(Me-3,2-HOPO) was active when administered either by injection or orally.
Design criteria for new iron chelators include:
- selectivity for Fe+3, not Ca, Mg, Cu, or Zn
- High affinity and thus ability to compete with in vivo complexation sites
- Rapid excretion as the iron complex
Richard Anderson provided a discussion of chromium intake levels and potential beneficial effects of chromium supplementation. Many effects related to glucose and lipid metabolism, and insulin action, have been noted. It appears from a few cases involving complete parenteral nutrition that some level of chromium is essential. However, the required daily amount is not known. The current Estimated Safe and Adequate Daily Dietary Intake (ESADDI) amount is 50-200 micrograms. The EPA reference dose for Cr+3 toxicity is 70,000 micrograms. However, Cr+6 is toxic (see discussion by Costa below) and the difference may be due to differences in bioavailability and cell uptake, rather than the innocuousness of Cr+3, per se. Actual average daily intakes in the U.S. population are 33 microgram in males and 25 micrograms in females. In a controlled diet study providing intakes less than 20 micrograms per day for 5 weeks, little effect was found in most subjects. However, glucose tolerance was decreased in patients with initial borderline hypoglycemia. Chromium requirements may be increased in diabetes. Several clinical trials have suggested the efficacy of chromium supplementation in patients with Type I, Type II, gestationally induced, and steroid-induced, diabetes. Various markers (e.g., glucose tolerance, fasting blood glucose, urinary glucose levels, hemoglobin AIC, and circulating insulin) have been monitored. Potential benefits have also been observed for lipid level risk factors for cardiovascular disease. However, as noted in the discussion, an approximately equal number of trials have demonstrated no beneficial effect. An important consideration is the poor bioavailability of chromium supplements (0.5-2% of dose for the most available chemical forms and completely unavailable in other forms). Measurement of urinary Cr loss is a valuable measure of bioavailability. Another complication is the lack of a simple relationship between hypoglycemia and blood or urinary chromium levels. In fact, blood Cr levels and urinary losses may be elevated in diabetes. Urinary losses are also influenced by physical stress or corticosteroid treatments. In vitro and animal studies have demonstrated: i) increases in insulin binding, ii) increases in insulin receptor number, iii) increased insulin internalization, iv) increases in receptor phosphorylation, and v) inhibition of receptor dephosphorylation. All of these effects increase insulin sensitivity. A low molecular weight chromium (LMWCr) factor has been isolated by John Vincent, which stimulates insulin receptor tyrosine kinase and may also inhibit phosphatase activity. Whether this observation adequately explains the clinical data and whether this is the only role of chromium in the body remains to be seen.
A lively discussion following Dr. Anderson's talk suggests the following conclusions:
- The potential benefits of chromium supplementation remain controversial.
- A problem is the lack of a validated measure of chromium status.
- Analytical methods are needed that reflect the biologically important chromium pools.
- Molecular mechanisms of chromium action in the body are poorly understood.
- Availability and widespread use of chromium dietary supplements by the population underscores the need for additional research.
- The low bioavailability of chromium contributes to confusion in the results of clinical trials, but may also be a major factor in safety.
- The extremely slow ligand exchange kinetics of Cr+3 dominates its chemistry, but perhaps should not be allowed to completely dominate thinking about its biology. There are things a metal (or metal complex absorbed from the diet) could conceivably do in the body that are not limited by slow ligand exchange.
Max Costa provided balance to the program by covering two examples of known metal toxicity, due to particulate nickel sulfides and high valence chromium. Both are carcinogenic, but by different mechanisms. Crystalline NiS and Ni2S2 particles with negative surface charges are phagocytosed by cells and degraded in low pH vacuoles to release soluble Ni ions into the cell at very high (>0.25 M) concentrations. Soluble nickel compounds, which cannot achieve such high concentrations, are not very carcinogenic. The carcinogenic effect of nickel appears to be due to the silencing of tumor suppressor genes by inducing DNA methylation and inhibiting histone acetylation. Methylation of DNA, once altered, is propogated upon cell replication, resulting in effectively permanent mutagenesis. Acetylation of histones is necessary for unwinding the contacts between adjacent nucleosome particles to allow transcription. Nickel binds to a peptide sequence in histones adjacent to the site of acetylation, which may normally bind copper or zinc ions. Failure to unwind may affect a large number of genes. Other effects may also be involved, and their interrelationships are not yet clear.
Biomarkers of nickel exposure include:
- Decreased degradation of key transcription factors, that are regulated by ubiquitin-mediated pathways (e.g., HIF-alpha, p53, ATF), resulting in downstream gene activation
- Activation of very specific signaling cascades Ca//NFAT, NFkB, but not RARE, GRE CRE, P53, AP1
- Cap43 induction by nickel through both Ca and HIF-alpha
- DNA hypermethylation and inactivation of genes
High valent chromium compounds are man-made oxidants (e.g., CrO3-2), and are toxic environmental pollutants. CrO3-2 is an oxyanion, and is efficiently transported into cells and subcellular compartments by carriers that normally transport HPO4-2 and SO4-2. Once internalized, it can react with cellular reductants to generate radical species and then undergoes reduction to Cr+3.. Cr+3 forms tight adducts with proteins, amino acids, glutathione, and DNA. A UvrABC DNA repair assay for DNA adducts and PCR mapping methods revealed hot spots for DNA damage, e.g., in the p53 tumor suppressor gene. Sensitivity of the adducts to EDTA suggests that they are due to metal binding.
Controversy over the toxicity of chromates may relate to variable susceptibility in the population. Improved biomarkers of exposure, effect, and of individual susceptibility are needed.
Biomarkers of chromate (VI) exposure and effect include:
- Cr in red blood cells (exposure); Chromate is actively taken up by RBC dianion transporters.
- DNA-protein cross-links and EDTA removal tests (exposure and effect)
- UvrABC/mapping of Cr adducts with ligation-mediated PCR (exposure and effect)
4) Amino acid/GSH, Cr(III) DNA adducts (exposure and effect)
Biomarkers of Genetic Susceptibility to Cr(VI):
- 5-methyl cytosine content in DNA increases adduct formation in opposite G residues.
2) Cellular uptake of Cr(VI) is variable (WBC uptake in vitro test may be useful).
- Differences in reduction kinetics of Cr(VI) to Cr(III), which may be mediated in part by variations in cytochrome P450s
Christopher Frederickson provided a brief presentation on zincergic neuron signaling. This process is analogous to other neurotransmitter mechanisms. Zinc is stored in pre-synaptic vesicles, released in brief pulses into the synaptic cleft, and acts on recognition sites in the post-synaptic membrane. Zinc-containing nerves are anatomically widely distributed in the brain. In normal brain tissue, most zinc is localized to vesicles in the nerve terminals as visualized using EM histology or the TSQ fluorescent indicator. In certain pathologies, e.g., a rat convulsive model, zinc is depleted from the pre-synaptic terminals, enters the post-synaptic cells via Zn-channels, and is associated with death of the post-synaptic cells. The hypothesis is that zinc excitotoxic reactions, analogous to glutamate excitotoxic pathways may be important, and might be amenable to prevention by appropriate zinc chelators.
This field stands where the field of calcium signal transduction stood a decade ago. There is a need for additional tools to study localized zinc concentrations and fluxes and to intervene in zinc signaling pathways. The macromolecular players involved in zinc signal transduction pathways remain to be determined. The role of such pathways in tissues other than the brain remains to be determined. The potential involvement of abnormal zinc signaling and accumulations in various pathologies remains to be fully clarified.
Jie Liu provided an example of the application of microarray technology to studying the problem of arsenite toxicity. In this work, RNA extracted from control and arsenite transformed/tolerant cells was converted to cDNAs which were used to probe a gene array. Changes in expression levels were confirmed by RT-PCR and Western blots. Expression levels of more than 70 genes were observed to have changed. The presentation highlighted only the few changes most relevant to arsinite tolerance. Upregulated genes included GSTs and GSH-arsenite adduct transporters (MRP1,2 and MDR).
Microarray technologies have yet to be widely applied to problems in metal metabolism and metal toxicity. They have much potential to help sort out the multi-factorial mechanisms of metals in biological systems.
Day 2--Morning Discussion
Questions were raised about differences in the results obtained by O'Halloran showing the occurrence of zinc vesicles in non-neural cells and the results of Frederickson, who has only observed them in neural tissue. Differences in reagents and methods (Zinquin versus TSQ fluorescent agents and the silverstain method used in EM studies), hence differences in sensitivity, were noted. A concern in using chelators is whether they reflect the localization of Zn in situ without perturbing it (e.g., by artificially localizing it to vesicles that sequester the chelator). Another issue is the integrity of the system studied. Do cells in culture reflect the normal state of cells in tissues? Do sample preparation methods affect the results (e.g., Frederickson sees Zn in endosomes in cells that have been damaged)? Permeabilization of cells that may be required for some in vitro studies may alter cellular zinc compartmentation. In situ electron beam x-ray microanalysis might be a useful tool for some vesicle sizes.
The idea that proteins may catalyze copper transfer rates brought up a question about small molecule effectors of the process. Thus far, there is no evidence of this. It was noted that catalysis of metal transfer rates is not unique to protein bound metals, rather it is the norm in ligand exchange chemistry.
A discussion of chelation therapy developed around the tradeoff between the thermodynamic advantage of larger multi-dentate molecules and the problem of oral drug bioavailability. Preliminary results in rodent models may not be sufficiently predictive of oral bioavailability in humans. The average molecular weight of drugs currently on the market is about 260 daltons. The sizes of leads in the pipeline is increasing to an average of about 400 daltons. The potential of chelator co-therapy to minimize toxic side effects was discussed. This is already done in the case of some radiation therapies and in some cases with cisplatin to protect non-target tissues and scavenge residual metals after the planned treatment.
The problem of identifying the true metal associated with the in vivo functional form of enzymes versus metal exchanged into enzymes upon purification was discussed. This may be a particular problem in overexpression systems. How proteins insert the correct metal (when and where in the cell) is incompletly understood. In some cases, this can clearly be post-translational, in others it may be co-translational. In the case of heme proteins, metal insertion into protoporphyrins occurs in the mitochondria, and insertion of the heme unit into protein occurs post-translationally.
A continued discussion on the potential benefits of chromium supplementation suggests that the basic science is not yet available to make clear determinations. There have not yet been definitive studies establishing dose/response and mechanisms of action. A large single study was suggested for consideration.
Day 2--Afternoon Session
Session 5: Metallotherapeutics and Disease
Chris Orvig discussed the therapeutic potential of vanadium compounds for diabetes. Such compounds may be useful adjuvants in Type I and useful treatments in Type II diabetes. They are orally active, but not effective in the complete absence of insulin. A vast literature of in vitro insulin-like effects has been reported for vanadate ion (VO4-3). Initial observations on the activity of sodium vanadate in vivo were published in 1985. Other oxidation states (e.g., peroxyvanadates and vanadyl (VO+2) compounds), have also shown activity. The Orvig lab has been pursuing bis(maltolato)-oxo-vanadium (BMOV), which is a chelated form of the vanadyl ion. It is neutral, water soluble, and small (mw = 317), all of which favor bioavailability. It is easy to prepare and is based on maltol, which is an FDA-approved food additive. It was shown to provide rapid and sustained correction of hyperglycemia in the STZ rat model of diabetes. This effect was due to enhanced insulin responsiveness, rather than altered insulin levels. It was 2-3 times more potent that vanadyl sulfate, a difference that correlates with the demonstrated higher bioavailability of BMOV using 48V tracers. Although oxidative metabolites are formed, activity resides with the parent compound. BMOV appears to act mainly as a vanadyl delivery vehicle since dissociation occurs in the plasma. EPR studies demonstrate transfer of VO+2 to serum transferrin can occur. The ethyl-maltol derived analog of BMOV is currently in Phase I clinical trials. BMOV has been marketed as a dietary supplement under the trade names "Supervanadyl Fuel" and "Gluconase".
The mechanism of vanadium compound action in diabetes is not yet clear. Vanadate is known to be a phosphate mimic. Vanadyl is able to bind to various divalent metal binding sites. Vanadium compounds may alter the redox poise responsiveness of the insulin receptor kinase and may inhibit the receptor phosphatase. Vanadium is a trace element in the diet and as in the case of many trace elements, it is difficult to establish the range of optimum physiological response. Important questions remain about speciation, oxidation state, ligation, bioavailability, and how these are altered by ligand designs. The existence of the dietary supplement industry has resulted in marketing of vanadyl compounds (including BMOV) in the absence of the studies needed to establish them as drugs.
Orvig pointed out that the actions of metals in most cases means the action metal complexes and the real research has to do with ligand design. This requires good synthetic organic chemistry as well as an understanding of coordination chemistry. Design issues include: stability (thermodynamic and kinetic, and particularly, hydrolytic), molecular weight, charge, lipophilicity, water solubility, functionalization for targeting, and metabolism. Naturally occurring ligands and previously approved materials such as maltol can simplify the regulatory process somewhat.
Dennis Riley discussed the design chemistry involved in development of the Metaphore superoxide dismutase mimics. Superoxide anion and hydrogen superoxide (pKa » 5) are implicated in a very large number of pathological processes. Three types of superoxide dismutases are known, containing Cu, Fe, or Mn as the redox-active center. These same metals are active in model complexes. Mn compounds were selected for development because the toxicity of Mn aquo complexes is lower than that of Fe or Cu aquo complexes. Mn is an essential trace element, and the body has mechanisms for its regulation and elimination. A penta-aza-crown compound, identified as an early lead, showed activity 20% of the enzymatic rates. Design objectives were to increase stability of the complexes, yet also increase catalytic activity. This was accomplished through understanding the mechanism of reaction and optimizing the geometry of the complexes. The rate limiting step in the Mn catalyzed superoxide dismutase reaction is oxidation of Mn+2 to Mn+3. The reductive half-reaction is facile and need not be specific to superoxide to be useful. Complex stability was increased by increasing the rigidity and pre-organization of the free ligand toward its metal-bound geometry. Catalytic activity was optimized by minimizing the structural reorganization upon changes in Mn oxidation state and enhancing the rate of proton-coupled outer sphere electron transfer from Mn bound H2O to HO2· to yield H2O2. The present generation of mimics are stable and active in vivo. Efficacy has been demonstrated in a rat ischemia/reperfusion model. Data on additional indications for the Metaphore SOD mimics was presented in a poster by Daniela Salvemini.
This work illustrated the application of molecular mechanics to ligand design and the application of basic principles in inorganic chemistry to generating stable, yet catalytically active metallotherapeutics. Other factors considered in the discussion included reactivity with species other than superoxide, and effects of design on overall mid-point potentials.
Susan Doctrow made a short presentation (actually later in the program) on Mn Salen complexes that are being developed by Eukarion, Inc, as SOD and catalase mimics. These have shown activity in cellular and rodent models for stroke, myocardial infarction, inflammatory diseases, and neurodegenerative diseases. Experimental results were reported for studies in Mn-SOD knock-out and the human FALS mutant Cu,Zn-SOD mouse models. In both cases, improvements in lifespan, markers of oxidative stress, and measures of neurodegeneration were observed.
Craig Hill presented data on the potential of polyoxometallate complexes as broad spectrum antiviral agents. These are nanoscale assemblies of early transition metals (e.g., vanadium, tungsten, molybdenum) with oxygen to form a variety of cage-like structures. Single or multiple sites of heterometal atom substitution may be entrapped or exposed on selected surfaces and vertices. Potential advantages are low cost of synthesis, self-assembly, and highly versatile combinatorial approaches, including variations in elemental composition, redox potentials, acidities, solubilities, shape, size, charge, and catalytic activity. Approximately 300 have been screened and about 80% are active in cell culture models with some fairly potent examples against HIV and CMV. Unfortunately, the only clinical trial to date was conducted with a material that, in hind-sight, is one of the most toxic members of the class. This may have thrown an early damper on the entire area. Studies of HIV protease inhibition have revealed compounds with Ki » 1 nM and EC50 in cells of about 1 mM. Furthermore, they appear to be active by different mechanisms than the current HIV protease inhibitors and thus might be active against resistant strains. Molecular modeling for a K+7(P2NbW17O62)-7 salt suggests binding of the oxyanion to a cationic pocket on the outer surface of the protein in the hinge region of the active site flaps, rather than binding within the active site.
Opportunities and obstacles for the future:
- Synthesis, purification, and characterization of these highly varied species present a chemical challenge.
- Clinical tests of an early lead may have been premature.
- The role of ion-pairing with these polyanions in vivo is not understood.
- Additional pharmacokinetic data is needed. How stable are they? How are they handled in the body? Does the presence of metals pose an insurmountable public relations obstacle?
- So many of the compounds are active in cell cultures that such results provide limited guidance for further development. Additional studies in animal models are needed.
Jill Johnson provided an overview of the NCI Developmental Therapeutics Program and their experience screening metal-containing agents for both cancer and AIDS therapy. The program, which began in 1955, has gone through several evolutions. Until recently an "empirical drug discovery" paradigm has been used based on screening for anti-proliferative effects in animal models (1975-1989) or a 60 cell-line screen (1990-present) with implanted hollow-fiber (1995-present) and human xenograft mouse models (1990-present) as the secondary screens. During this time 14,900 metal-containing compounds have been tested (including a wide range of elements). Of those, 1,242 were selected as active in the mouse in vivo screen, 191 were selected in the 60 cell-line screen. Nine were investigated as clinical candidates. Five resulted in INDs. Two resulted in NDAs and are currently used mainstream therapies (cisplatin and carboplatin). Reasons for dropping clinical candidates included: i) stability, formulation, or other pharmaceutical development issues; ii) renal toxicity or neurotoxicity; and iii) insufficient advantage over current drugs. For comparison, of the 550,000 total compounds tested by the DTP, 14,475 were active in mouse screen models and 7,741 (out of 77,000) were active in the 60 cell-line assay; 59 INDs and 11 NDAs resulted. Since 1997, the NCI has also operated a screen for compounds active against the cytotoxic effects of HIV in CEM cells. Of 80,000 compounds tested, 4,050 (or about 5% were active). Of the compounds tested, 2,291 have included metals. Of those, 136 (about 6%) were active, two became clinical candidates. Both were dropped due to toxicity problems. The one was a polyoxometallate, and about 80 others were tested. These were found to be strongly active in vitro, but too toxic in animal models in vivo. If a way around the toxicity problem can be found, interest in these compounds would be very high. However, these obstacles in drug development are not unique to metal-containing agents. The frequencies of success in screens and also of failure in clinical development are not different than the experience with development of non-metal containing agents.
The model for drug development within the NCI program has changed to focus on mechanism-oriented rational drug discovery based on known molecular targets. Comparisons of compound activity against the 60 cell-line screen have developed correlations between cellular sensitivity and molecular target expression (e.g., using the COMPARE algorithm). Microarray expression and specific target activity assay programs to characterize the levels of 300-400 targets in the 60 cell-line screen are underway. A yeast assay screen against a large number of mutants has been developed to further help identify targets of compound action. Interestingly, metals as a whole, and even the group of platinum compounds tested, do not cluster well in these studies. This implies that different metals act through different mechanisms of action. In the new drug development model, it is not sufficient to find anti-proliferative activity and test additional chemically related compounds. It is necessary to demonstrate and optimize activity against a specific target. Furthermore, DNA targeting, per se, is no longer considered a sufficiently good rationale. A more specific cellular effect of DNA modifying agents must be demonstrated. These same constraints, however, operate equally in the selection of non-metal containing candidates.
Take home messages:
- A metal is not a metal is not a metal.
- Metals fared about as well as non-metals.
- Metals remain viable candidates for testing and development; however, the rationale for considering all compounds has become more stringent.
- Future developments must be based on rational targeted design.
- The greatest need is for improved ability to predict human activity from cell culture and animal studies.
Nicholas Farrell provided a summary of some general principles (see below) and specific examples from his own work on platinum anticancer drugs. Most platinum compounds that have been tested are derivatives of cisplatin. Although they may show altered pharmacokinetics, they are likely to act in the same way. With the intent to develop compounds with new activities, the Farrell lab has developed compounds designed to form different types of cross-links (e.g., long range inter- and intra-strand cross links) by combining two mono-chloro Pt-based DNA binding groups with a diamine linker. Variations have examined leaving group, linker chemistry, chain length, steric effects, H-bonding, and charge. The most successful compound (BBR3464) contains an non-reactive tetra-amino-platinum unit in the linker, but this can be replaced with other positively charged units. This compound is active at nM concentrations in vitro, positive in animal and human tumor xenograft screens (against 15 types of tumor), and showed an acceptable pre-clinical toxicology profile. Phase I human trials were completed and Phase II trials are underway. Conclusions from Phase I included an acceptable maximum tolerated dose limit, an altered dose-limiting toxicity in comparison with cisplatin (diarrhea and myelosuppression as opposed to renal toxicity), linear pharmacokinetics with a long t1/2, and preliminary indications of efficacy in GI malignancy (colon cancer, pancreatic cancer, and associated hepatic metastases). Studies on the mechanism of action showed formation of the predicted long-range cross-links, which are not recognized by HMG1 (which recognizes cisplatin cross-links). Tissue selectivity and downstream signaling effects are also different than for cisplatin.
Another effort has been directed toward examining trans-platinum derivatives. Although earlier studies had shown these are less active than cis-platinium derivatives, this is partly because they are too reactive. By using planar amine ligands to slow down the reactions, potency was increased. The types of cross-links formed and the mechanism of action differ. In these compounds, protein-associated strand breaks were found that are similar to those produced by DNA topoisomerase inhibitors such as camptothecin.
Application of the COMPARE program to the NCI data on platinum complexes reveals several well separated clusters with different spectra of activity against the 60 cell lines. These results clearly show that a monolithic model of platinum compound action is inadequate, and that multiple mechanisms of action are more likely.
Although the collection of presentations at the Metals in Medicine meeting might be considered eclectic, they actually represent an emerging and cohesive field. The talks span a continuum of interests from essential elements, in both normal and disease processes, to non-essential elements and the potential beneficial as well as detrimental effects thereof. They are joined by the necessity of bringing both a deep understanding of inorganic chemistry and state-of-the-art research in biology to bear on the problems. They are further characterized by a focus on basic research, but with a clear vision of potential future medical application.
Medicinal Inorganic Chemistry is a multidisciplinary field combining elements of
chemistry (synthesis, reactivity); pharmacology (pharmacokinetics, toxicology); biochemistry (targets, structure, conformational changes); and medicinal chemistry (therapeutics, pharmacodynamics, SAR).
Medicinal inorganics have an appreciable current market impact and significant growth potential. Current sales, including imaging, diagnostics, and therapeutics are on the order of $2 billion per year. Platinum-based chemotherapeutics are among the most successful of all anti-cancer drugs with combined sales of $700 million per year, second only to Taxol. New compounds are likely to gain a unique market niche by acting through different mechanisms and/or avoiding common clearance mechanisms (e.g., MDR). Successful or at least promising results have been achieved for various indications involving a significant number of inorganic elements from the periodic table. Additional exploration should continue to be fruitful. The success rate of metallopharmaceutical advanced clinical leads is about the going rate (e.g., 1:10 survive Phase I trials and 1:4 survive in Phase II). However, a limiting factor may be the relatively limited expertise of most pharmaceutical companies in the area of inorganic chemistry.
Drug development is a long and non-linear process. It would be incorrect to say major pharmaceutical companies have no interest. AstraZeneca, Warner-Lambert, Boehringer Manheim, Hoffman LaRoche, NovusPharma, Johson Matthey, Bristol-Myers Squibb, Dupont, Merck, and others were mentioned during the meeting as having collaborated or contributed support at various points along the way. However, a promising research lead, alone, does not a drug make. Issues of pharmacokinetics, efficacy, toxicity, and economics, along with patents, politics, and personalities, all factor into decisions about whether to continue or to drop a project. There is at least a perception that the bar is set higher for metal-containing drugs. The government-based development programs may need to reach further to off-set industrial caution in this area.
David Place of the FDA provided a discussion of regulatory issues for metallopharmaceuticals. The key point, however, is that there is really no difference in the process or in the expectations for demonstration of safety and efficacy that must be met. Phase 1 must establish safety. Phase 2 must establish safety and efficacy in a defined population. Phase 3 should provide definitive evidence of efficacy and safety. Phase 4 studies may be needed to resolve some post-approval questions. Several Centers and Divisions within FDA have responsibility for regulation of metallopharmaceuticals. The largest number have come to the Division of Medical Imaging and Radiopharmaceutical Drug Products within the Center for Drug Evaluation and Research.
Important issues with respect to metallopharmaceuticals are chemical characterization (e.g., complexes versus mixtures) and stability. New Molecular Entities must be fully characterized. This designation affects the number of years of exclusivity that may be received. Applicable analytical and spectroscopic methods must be available and appropriate animal models must be used for the pre-clinical pharm/tox studies. The mode of physiological/pharmacological action is important. Which part is the active unit (i.e., intact complex, part of ligand, the metal ion, itself) should be established. Knowledge of contaminants and their activity profiles is also important. Early planning is valuable, and interactions with FDA staff is encouraged at multiple stages in the development process. Inorganic chemistry training of FDA staff and reviewers is variable and it may be useful to provide additional material to help educate them about issues, such as "exotic" analytical methods which may be unfamiliar in a system where the majority of materials under consideration are organics.
Mike Abrams presented results on the anti-HIV activity of bicyclams and metallo-bicyclams. These materials were identified by a program of screening metals and metal chelators. Bicyclam was identified as the active 1% contaminant present in samples of cyclam and was shown to have nanomolar activity against HIV with low cytotoxicity. The current AnorMED lead compound AMD3100 (a para-substituted phenyl bridged bicyclam) was shown to have a novel mechanism of action. It inhibits HIV replication by binding to the chemokine receptor CXCR4, which is used as a co-receptor by the virus for membrane fusion and cell entry in its T-tropic stage. Binding of AMD3100 inhibits binding of an anti-CXCR4 mAB and of the endogenous ligand SDF-1a. Zinc ions were shown to enhance the binding of AMD3100 to the receptor and several other metal ion complexes bind with affinity similar to AMD3100. Modeling and mutagenesis suggest binding to the 7TM receptor at two sites involving asparate residue stabilization of positive charges. Phase I trials were completed in healthy volunteers in 1999. Phase II trials are in progress in HIV patients.
Discussion brought out the problem of knowing whether a metal chelator, when tested in a metal-free form actually acts by picking up a metal in vitro and in vivo. Is the activity of AMD3100, alone, due to that species, or a metal chelate that may well form by picking up metals in the buffer? It was pointed out that perhaps 40 percent of the drugs on the market are potential metal chelators. Whether their pharmacokinetics and/or pharmacodynamics may be affected by metal chelation in the body is largely unknown. The availability of Zn, Fe, Cu in the cell may be exploited in drug design by engineering specific chelating scaffolds. Zn is known to inhibit some enzymes, such as serine proteases, which contain suitable coordination sites. It may be possible to design ligands that in conjunction with target protein residues create a zinc binding inhibitor site.
Dr. Abrams offered the following general comments:
Reasons why major pharmaceutical companies have shown only modest interest in the field include:
- A high degree of serendipity versus target-directed design in the past
- Concerns about metal accumulation versus long-term therapeutic requirements
- Concern about inadequate specificity
- Concern about reversibility (or lack thereof)
- Unfamiliarity--few coordination chemists are employed in industry
Ways to make inorganic medicines more "drug-like" and of greater interest include:
- Take advantage of the enormous growth in knowledge of drug targets to find unique applications.
- Make use of catalytic mechanisms that cannot be duplicated by other agents.
- Look for reversible competitive inhibitors of enzymes and ligands for receptors that take advantage of the unique properties of metal ions.
- Take advantage of opportunities to recruit endogenous metals to avoid having to administer metal-containing agents.
Day 2--Afternoon Discussion and Overall Meeting Discussion
Various speakers identified ways that basic research in metallobiochemistry and bioinorganic chemistry can contribute to the development of future pharmaceuticals and the health benefits therein. Studies of metalloenzyme structure and function, and of the roles of metalloenzymes in biological systems, are a vigorous area of ongoing activity. Emerging areas of interest include the roles of metals in cell regulation and the pathways by which cells regulate metal ion concentrations and delivery to selected metal sites. Opportunities exist to target these processes to alter metalloenzyme synthesis and to alter metal ion accumulations. Opportunities exist to exploit the unique properties of metal containing compounds as pharmaceuticals. These include unique physical properties, unique types of biomolecule interactions, redox, and other reactivities.
A limiting factor in the development of metallopharmaceuticals is the perception that metals are toxic. This is certainly true of some metals. Even essential metals are toxic under some circumstances. However, metals are not necessarily toxic under all circumstances. All metals are clearly not heavy metals. To rule out a large fraction of the periodic table, a priori, eliminates any potential benefit that may be discovered among those elements. Understanding and learning to control the actions of metals in vivo is the key to potential beneficial applications of metals, and the key to avoiding metal toxicities.
It was pointed out that work in this area requires review by both chemists and biologists, and that both types of expertise were not often present on the same panel. CSR staff pointed out that NIH review activities are primarily proposal-driven. If increased numbers of applications are received in this area, potential changes in the distribution of applications to various study sections and the memberships of the study sections will be adjusted to accommodate the workload.
--------------------------------------------------------------------------------
|
 |
