9 minutes reading time (1859 words)

    Application of Basic Chemical Properties in Drug Design: The Development of Hypoxia-Selective Anticancer Drugs

    William A. Denny
    University of Auckland

    The first useful drugs were complex mixtures of compounds of unknown structure, extracted from natural sources (plants or animals), and resulted from chance observations of useful effects, probably backed up by trial and error experiments on humans over many generations. A therapeutic drug is now defined as a pure compound of known molecular structure, and while some are still extracted from natural sources, most are obtained by synthesis in the laboratory. The modern pharmaceutical industry is a huge global business, worth about A$ 300 billion annually, and the availability of safe and effective drugs for many diseases is one of the significant contributions which chemistry has made towards our well-being.

    At first sight the process of drug design and development seems difficult to relate to the chemistry one learns (or teaches) in secondary school. Certainly, it is no longer a random process, but a highly multidisciplinary effort, most effectively carried out by teams of scientists who are experts in different areas, and backed up by complex instrumentation. However, chemistry remains the discipline at the heart of drug design, and interesting new drugs can be developed by applying basic chemical knowledge to exploit new biological information. This can be illustrated by some of the work currently in progress in our laboratory on the design of new classes of anticancer drugs which take advantage of the fact that cells in many tumours are oxygen-deficient.

    Tumour hypoxia

    A major reason why the available anticancer drugs have only limited effectiveness is because cancer cells do not differ very much from normal cells. Therefore it is difficult to find any structural targets which exist only in tumour cells, and most drugs are not very selective for tumour cells, killing large numbers of normal ones as well. However, recent research in cell biology has shown that some cells in most tumours do differ from all normal cells in the body in an important environmental sense, in that they are deficient in oxygen (hypoxic). This is due to the limited and relatively inefficient blood vessel network inside tumours, which means that some cells are a long way from the nearest blood vessel. Oxygen cannot diffuse more than about 150 μm in tissue because it is metabolised, so that cells further away than this from a blood vessel are hypoxic, but still viable (Figure 1). Even further away cells die, generating the necrotic areas seen in many tumours. Tumour hypoxia therefore provides an opportunity for the design of new types of drugs, which would become active only under hypoxic conditions, generating toxic species exclusively in tumours. What basic chemical principles can assist us in the design of compounds with these properties? 

    Hypoxia-selective drugs

    To be successful, hypoxia-selective drugs must have a number of basic properties. They must be normally non-toxic, and be able to move (diffuse) into the interior of solid tumours where the hypoxic cells are. Once there, they must be able to sense the lack of oxygen and use it to convert themselves into potent cell-killing agents. One general design which we have been studying contains two chemically distinct units; a trigger which will be converted to some other form only in the hypoxic target cells, linked to an effector which will be released as a very toxic species only when the trigger is activated (Figure 2).

    Trigger units

    The most reliable way to achieve the required triggering is to use reductive processes. Many types of molecules are reduced in cells by enzymes which can sequentially add electrons to them. This happens in all cells, both oxygenated normal cells.and hypoxic tumour cells. However, in many cases these electron adducts are rapidly re-oxidised by molecular oxygen, a process which can only occur in oxygenated cells. Therefore, for these compounds, the initially formed one-electron adduct can selectively accumulate in hypoxic cells, because of the lack of oxygen to re-oxidise it back to the parent drug. Because such one-electron adducts are usually unstable, in the absence of oxygen they eventually convert to some other form. Several types of molecule are known to undergo this behaviour at the appropriate rates in cells (Figure 3), and these are suitable candidates to use as triggers.

    Effector units.

    These must be non-toxic until released by the trigger, and then be able to kill cells very effectively. The active form must have a lifetime of a few minutes; long enough to be able to diffuse back out of the hypoxic cell where it originates and kill surrounding tumour cells, but not so long that it can move out of the tumour altogether and kill normal cells. There are only a few types of compounds which fill these requirements; one is the nitrogen mustards. These were originally developed as war gases and later as conventional anticancer drugs. In aqueous solution they have half-lives of a few minutes, undergoing spontaneous cyclisation to the aziridium cation. This is a powerful alkylating agent, and mustards can sequentially form two such intermediates, which can kill cells by joining the two strands of DNA in the double helix together (cross-linking it), preventing the strand separation necessary for cell division (Figure 4). Nitrogen mustards are reactive molecules because of the high electron density on the nitrogen. Therefore, they can be deactivated (made non-toxic) by drawing electrons away from this nitrogen atom (through the N-R bond; Figure 4), and must be linked to a suitable trigger unit through this bond.

    ​Cobalt complexes of mustards

    Three basic chemical principles lay behind our study of these compounds as potential hypoxia-selective drugs. One was the well-known fact that coordination complexes with transition metals such as cobalt show huge differences (up to 1012-fold) in the stability of the coordinate bonds with nitrogen ligands, depending on the oxidation state of the metal. Thus Co(III) complexes have halflives of many months, but when the metal is reduced (by addition of one electron), the corresponding Co(II) complexes have halflives of only milliseconds (Figure 5). The second principle was that Co(III) complexes can be efficiently reduced by enzymes in cells to the corresponding Co(II) complexes. The third principle was that in a Co-N coordinate bond, electrons are withdrawn from the nitrogen, thus deactivating a nitrogen mustard if this is used as a ligand.

    We therefore prepared various cobalt complexes of mustards (eg., SN 24771), designed to have appropriate reduction potentials and stability (Figure 5). The other (acetonylacetone) ligands on the cobalt were chosen to provide an overall single positive charge for the complex (to allow rapid uptake into cells) and an appropriate reduction potential for the metal (to allow efficient reduction by enzymes in cells). While stable (and thus not toxic) in the Co(III) form, SN 24771 is radidly reduced in cells to the Co(II) form. While this happens in all cells, in oxygenated cells the Co(II) form is rapidly re-oxidized by molecular oxygen back to the stable Co(III) complex. However in hypoxic cells, where such back-oxidation cannot occur, the labile Co(II) complex breaks down to the hexaaquo form, releasing the ligands, including the toxic nitrogen mustard (Figure 5). Thus, SN 24771 is about 35-fold more toxic to human tumour cells (grown in the laboratory as single cell suspensions in tissue culture) in the absence of oxygen (hypoxic conditions) than it is under normal oxygenated conditions (Figure 6). Some cell lines can also grow in small clumps called spheroids, which are a better model for solid tumours because only the cells in the centre of the spheroid become hypoxic. When tested against such tumour spheroids, SN 24771 shows effects which suggest that it behaves as designed, diffusing to the hypoxic core of the spheroid where it is activated (by reduction), releasing the activated mustard which then diffuses back out to kill tumour cells (which may not be hypoxic) surrounding the hypoxic core.

    Nitrobenzyl quartenary mustards

    In the design of these compounds we took advantage of two well-known chemical principles. One was the knowledge that the most effective way of deactivating a nitrogen mustard is to generate a permanent cationic charge on the nitrogen by quaternising it. The second was that conjugated systems possessing extremes of electron density in different parts of the molecule are liable to fragment. We therefore prepared and studied the properties of nitrobenzyl quaternary mustards such as SN 25246 (Figure 7). One-electron reduction of this molecule gives a radical anion, the lifetime of which can be controlled by the structure of the compound. It can be reoxidised by molecular oxygen to regenerate the parent compound, but in the absence of oxygen it spontaneously fragments, generating a benzyl radical and releasing the activated free mustard (Figure 7). In this case the released mustard is the clinical anticancer drug mechlorethamine, which is known to have a halflife of several minutes, sufficient to allow back-diffusion of the released toxin to surrounding tumour cells (which may not be hypoxic). 

    The quartenary mustard SN 25246 shows an extraordinarily high selectivity for hypoxic human tumour cells in tissue culture (>9000-fold under some conditions). This is by far the largest ratio known for any compound to date. In the spheroid assay mentioned above, SN 25246 again shows effects suggesting that it is able to diffuse to the hypoxic core of the spheroid, there being activated to release the toxic free mustard which then back-diffuses to kill oxygenated tumour cells surrounding the hypoxic core.

    Conclusions
    These two examples show that modern drug design is no longer a hit and miss affair, but proceeds from a knowledge of basic chemical principles. In the present case, a simple concept of linking suitable trigger and effector units has resulted in two completely new classes of hypoxia-selective anticancer drugs. The cobalt complex SN 24771 is the first example of a metal-based hypoxia-selective drug, while the quaternary mustard SN 25246 is the most hypoxia-selective compound known to date. In both of the described cases, the work is at an early stage; much more will be required before it is known whether drugs of either of these two classes would be suitable for human trial. However, the basic concept has been amply justified. Furthermore, because a number of different types of units might be used, many other classes may also be discovered; we are limited only by our imagination (and by our knowledge of chemistry).

    About the author
    Bill Denny completed a PhD at the University of Auckland in 1970, and following three years post-doctoral work at Oxford University joined the Cancer Research Laboratory (CRL), in the School of Medicine, University of Auckland, where he is now Co-Director and Professor in the Department of Chemistry.

    During 20 years of work in the CRL he has sought innovation in the development of new drugs for cancer treatment; work reported in nearly 300 publications and resulting in four drugs (amsacrine, CI-921, DACA and DMXAA) currently in clinical use or in clinical trial. Present projects include the development of hypoxia-selective cytotoxins, inhibitors of tyrosine kinase enzymes involved in signal transduction in cancer cells, and prodrugs for use in gene therapy. He is currently President of the New Zealand Institute of Chemistry, and in April/May 1995 will be the Wilsmore Fellow in the School of Chemistry, University of Melbourne. 

    A New Look at the Old Drug - Aspirin
    Artificial Sweeteners
     

    Comments

    No comments made yet. Be the first to submit a comment
    Already Registered? Login Here
    Guest
    Sunday, 18 November 2018

    My Account



    Get Social