Antiviral Metal Complexes

The initial events (virus adsorption and fusion with the cells) in the replicative cycle of human immunodeficiency virus (HIV) can serve as targets for the antiviral action of metal-binding compounds such as polyanionic compounds (polysulfates, polysulfonates, polycarboxylates, polyoxometalates, and sulfonated or carboxylated metalloporphyrins), bicyclams and G-octet-forming oligonucleotides. The adsorption and fusion of HIV with its target cells depends on the interaction of the viral envelope glycoproteins (gp 120) with the receptors (CD4, CXCR4) at the outer cell membrane. We are currently investigating how the aforementioned compounds interfere with these viral glycoproteins and/or cell receptor.


Introduction
There are ten steps in the replicative cycle of human immunodeficiency virus (HIV) that could be considered as targets for chemotherapeutic interventions (Table 1) (1). The early events in HIV infection, i.e. virus adsorption to the cells and virus cell fusion, have been shown to be the points of attack for some metal complexes or organic compounds containing metals. Here I will discuss those compounds among the metal complexes that interfere with the virus adsorption and/or virus cell fusion (i.e. polyanionic substances, bicyclam derivatives and G-octet-forming oligonucleotides). The key molecule in the viral adsorption/fusion process is the viral envelope glycoprotein gp120 (Fig. 1), which has a highly convoluted structure containing several regions referred to as variable regions such as V3 and V4, which are assumed to interact with the corresponding receptors at the host cell surface.
An interesting feature of the polyanionic substances is that their antiviral activity is not limited to HIV but also extends to various other enveloped viruses such as herpesviruses [herpes simplex virus (HSV), cytomegalovirus (CMV)], influenza A virus, respiratory syncytial virus (RSV), arenaviruses (Junin virus, Tacaribe virus) and rhabdoviruses [such as vesicular stomatitis virus (VSV)] (Fig. 6). This broad spectrum antiviral action considerably enhances the therapeutic potential of these compounds for the treatment of viral diseases. We have recently succeeded in obtaining mutants resistant to polyanionic substances after passaging HIV in the presence of dextran sulfate (5). It thus appears possible for the virus to develop resistance to these polyanionic substances (Fig. 7). The resistance mutations appear to be located predominantly in the V3 loop of the gpl20 glycoprotein ( Fig. 8) and render the overall charge of the V3 loop less positive, thus resulting in a diminished electrostatic interaction with the polyanionic compounds.
Polyanionic substances such as polyoxometalates (i.e. polyoxosilicotungstates) inhibit the replication of HIV, HSV and CMV at the virus adsorption step. They inhibit influenza A virus and RSV at the virus-cell fusion step (6).
Little is known about the therapeutic/prophylactic potential of the polyanionic substances in the clinic. Protective effects with these compounds against HSV and influenza A virus infections in vivo (mice) have been described with polyoxotungstates and polysulfonates when administered systemically (intraperitoneally) or topically (intranasally), respectively (7,8). In particular, topical administration of the polyanionic compounds, for instance as a vaginal formulation, seems to be an attractive modality for the prevention of sexually transmitted HIV and HSV infections (9).

Bicyclams
The bicyclams (10) represent a new class of highly potent and selective HIV inhibitors.
They originated from the serendipitous discovery of anti-HIV activity in a monocyclam preparation that contained bicyclam as contaminant. Starting from this lead compound, several bicyclam derivatives were prepared that showed increased anti-HIV activity ( Fig. 9) (11). The most active of this series is the bicyclam JM3100 ( Fig. 10) which has been found to inhibit HIVinduced cytopathicity at a concentration of a few nanograms per ml, while not being toxic to the host cells at concentrations up to 500 tg/ml, thus achieving a selectivity index of 100,000 or higher (Table 5) (12). so-so-soso-sosososososo; so-so-so-so-sosososososo-   11). From these time-of-addition experiments we must conclude that the bicyclams interact with the fusion/uncoating process (13). This process involves the removal of the envelope as well as capsid proteins from the viral RNA genome so that the latter can be transcribed by the reverse transcriptase. Theoretically, any of the viral envelope glycoproteins (gpl20, gp41) or capsid proteins (plT, p24, p9, p7) could be considered as possible targets for the interaction with the bicyclams (Fig. 12). Originally (13,14), we envisaged the capsid protein p7 (Fig. 13) as a possible target for the bicyclams, as this protein contains two zinc fingers that could possibly make zinccoordination complexes with the bicyclams.
After painstaking efforts, we succeeded in obtaining mutants that were resistant to the bicyclams JM2763 and JM3100 (Fig. 14) (15,16). Sequence analysis of these mutants revealed the presence of several mutations within the gpl20 glycoprotein located in the V3-V4 region (Fig. 15) (17). Although it is not clear yet which and how many of these mutations are required for engendering resistance, it is obvious that the primary site of interaction for the bicyclams is the gp 120 rather than any of the other viral glycoproteins or capsid proteins.
The role of several metals in the interaction of the bicyclam JM3100 with HIV has been assessed. From Fig. 16, it is evident that Zn facilitates the binding of the bicyclams to the virus.
Equilibrium analysis studies revealed that the optimal binding of JM3100 with the virus is achieved at Zn concentrations of 0.2 to 0.6 mM (Fig. 17). That Zn may play a key role in the anti-HIV activity of the bicyclams is also evident from Table 6. Only the Zn complex with the bicyclam JM3100 (i.e. JM3479) was equipotent to JM3100. The other metal bicyclam complexes (JM3462, JM3469, JM3461 and JM3158) containing Ni, Cu, Co, Pd, respectively, showed gradual loss in activity, the Pd complex being inactive as an anti-HIV compound. JM3100 has been found efficacious in vivo, in decreasing the virus load in the SCID-hu Thy/Liv mice (that is, severe combined immune deficient mice reconstituted with human fetal thymus and liver) infected with HIV (18)      G-octet-forming oligonucleotides The oligonucleotide 5'GTGGTGGGTGGGTGGGT3' which forms a G-octet with potassium in the middle (Fig. 18) (19) has been shown to exhibit activity against different HIV strains in cell culture. As shown in Table 7, the G-octet forming oligonucleotide T30177 (also referred to as AR177 or Zintevir) shows activity against the HIV-I(III) and HIV-I(RF) strain at an EC0 of 0.15-2.8 and 0.03-0.3 laM, respectively (20). The anti-HIV activity of T30177 in cell culture persists for several weeks after an initial 4-day exposure of the cells to the compound and subsequent removal of the drug. This contrasts with the behavior of other anti-HIV compounds which rapidly loose their activity when removed after the original 4-day exposure (Fig. 19). It has been shown that the T30!77 inhibits the HIV-1 integrase (Fig. 20). The HIV DNA integration is an highly complicated process involving at least 3 steps (endonuclease, strandtransfer and DNA ligation) (Fig. 21) and the T30177 would interfere with the first step (endonuclease) of the integration process (21,22).
However, it is doubtful that the inhibitory effect of T30177 on the HIV integrase would account for the anti-HIV activity observed with the compound in cell culture experiments, as timeof-addition experiments with T30177 have indicated that the compound inhibits HIV replication, at a step which coincides with virus adsorption and/or fusion (Fig. 22). Also, resistant HIV strains selected under continuous pressure of T30177 revealed the presence of mutations in the gpl20 molecule, but not in the integrase gene (23), again pointing to the viral adsorption/fusion process as the primary target for the anti-HIV action of T30177.
Clinical studies with zintevir (T30177) have been initiated. When administered intravenously as-singleor repeat-doses to cynomolgus monkeys, zintevir did not cause significant hemodynamic toxicity (unlike other oligonucleotides) at plasma drug concentrations that have shown anti-HIV activity in vitro (24,25).    To enter the cells, HIV must first interact through its viral envelope gpl20 with the CD4 receptor of the host cell (virus adsorption) (Fig. 23) before the viral envelope can fuse with the outer cell membrane (Fig. 24). This virus cell fusion is made possible only after the viral envelope gpl20 has also interacted with the second receptor [ioe. fusin (CXCR4)]. Our current investigations are aimed at elucidating the interactions of the compounds described here (polyanions, bicyclams, zintevir) with the viral envelope gp 120 glycoprotein and the cellular receptors that are involved in the virus adsorption/fusion process.