We have previously described the potent and selective inhibition of several strains of human immunodeficiency virus type 1 (HIV-1) and type 2 (HIV-2) by JM2763, an n-propyl-linked dimer of the 1,4,8,11-tetraazamacrocyclic (cyclam) ring system. Upon further investigation, we have also found that incorporating an aromatic rather than aliphatic linker leads to analogs with higher antiviral potency. The prototype, JM3100 (19a, isolated as the octahydrochloride salt), which contains a p-phenylenebis(methylene) moiety linking the cyclam rings, inhibited the replication of HIV-1 (IIIB) and HIV-2 (ROD) at EC50's of 4.2 and 5.9 nM, respectively, while remaining nontoxic to MT-4 cells at concentrations exceeding 421 microM. In order to identify the structural features of bis-tetraazamacrocycles required for potent activity, we have prepared a novel series of phenylenebis(methylene)-linked analogs, in which the macrocyclic ring size was varied from 12 to 16 ring members. Depending upon the substitution of the phenylenebis(methylene) linker (para or meta), sub-micromolar anti-HIV activity was exhibited by analogs bearing macrocycles of 12-14 ring members but with varying cytotoxicity to MT-4 cells. Furthermore, while we found that identical macrocyclic rings are not required for activity, substituting an acyclic polyamine equivalent for one of the cyclam rings in 19a resulted in a substantial reduction in anti-HIV potency, clearly establishing the importance of the constrained macrocyclic structure. A short series of transition metal complexes of 19a were also prepared and evaluated. Complexes of low kinetic stability such as the bis-zinc complex retained activity comparable to that of the parent compound. Finally, the activity of bicyclam analogs appears to be insensitive to the electron-withdrawing or -donating properties of substituents introduced onto the linker, but sterically hindering groups such as phenyl markedly reduced activity. As a result, several analogs with anti-HIV potency comparable to that of 19a have been identified.

Bis-tetraazamacrocycles such as the bicyclam AMD3100 are a class of potent and selective anti-HIV-1 and HIV-2 agents that inhibit virus replication by binding to the chemokine receptor CXCR4, the co-receptor for entry of X4 viruses. With the aim of optimizing the anti-HIV-1 and HIV-2 activity of bis-azamacrocycles, a series of analogues were synthesized which contain neutral heteroatom (oxygen, sulfur) or heteroaromatic (of lower pK(a) than a secondary amine) replacements for the amino groups of AMD3100. The introduction of one or more heteroatoms such as oxygen or sulfur into the macrocyclic ring of p-phenylenebis(methylene)-linked dimers (to give N(3)X or N(2)X(2) bis-macrocycles) gave analogues with substantially reduced anti-HIV-1 (III(B)) and anti-HIV-2 (ROD) potency. In addition, the bis-sulfur analogue was also markedly more cytotoxic to MT-4 cells. However, bis-tetraazamacrocycles featuring a single pyridine group incorporated within the macrocyclic framework exhibited anti-HIV-1 and HIV-2 potency comparable to that of their saturated, aliphatic counterparts. The p-phenylenebis(methylene)-linked dimer of the py[14]aneN(4) macrocycle inhibited HIV-1 replication at a 50% effective concentration (EC(50)) of 0.5 microM while remaining nontoxic to MT-4 cells at concentrations approaching 200 microM. A series of analogues containing macrocyclic heteroaromatic groups of varying pK(a) were also synthesized, and their ability to inhibit HIV replication was evaluated. Replacing the pyridine moiety of the py[14]aneN(4) macrocyclic ring with pyrazine or pyridine groups substituted in the 4-position (with electron-withdrawing or -donating groups) either reduced antiviral potency or increased cytotoxicity to MT-4 cells. Finally, we synthesized a series of analogues in which the ring size of the bis-pyridyl macrocycles was varied between 12 and 16 members per ring including the py[iso-14]aneN(4) ring system, an isomer of the py[14]aneN(4) macrocycle. The p-phenylenebis(methylene)-linked dimer of the py[iso-14]aneN(4) (AMD3329) displayed the highest antiviral activity of the bis-azamacrocyclic analogues reported to date, exhibiting EC(50)'s against the cytopathic effects of HIV-1 and HIV-2 replication of 0.8 and 1.6 nM, respectively, that is, about 3-5-fold lower than the EC(50) of AMD3100. AMD3329 also inhibited the binding of a specific CXCR4 mAb and the Ca(2+) flux induced by SDF-1alpha, the natural ligand for CXCR4, more potently than AMD3100. Furthermore, AMD3329 also interfered with virus-induced syncytium formation at an EC(50) of 12 nM.

A series of bicyclam analogs connected through a heteroaromatic linker have been synthesized and evaluated for their inhibitory effects on HIV-1 (IIIB) and HIV-2 (ROD) replication in MT-4 cells. The activity of pyridine- and pyrazine-linked bicyclams was found to be highly dependent upon the substitution of the heteroaromatic linker connecting the cyclam rings. For example, 2,6- and 3,5-pyridine-linked bicyclams were potent inhibitors of HIV-1 and HIV-2 replication, whereas the 2,5- and 2,4-substituted pyridine-linked compounds exhibited substantially reduced activity and, in addition, were found to be highly toxic to MT-4 cells. We have subsequently discovered that these effects are not unique; amino-substituted linkers also have the potential to deactivate phenylenebis(methylene)-linked bicyclams. A model is proposed to explain the deactivating effects of the pyridine group in certain substitution patterns based on the ability of the pyridine nitrogen to participate in pendant conformations (complexation) with the adjacent azamacrocyclic ring, which may involve hydrogen bonding or coordination to a transition metal. The introduction of a sterically hindering group such as phenyl at the 6-position of the 2,4-substituted pyridine-linked bicyclam appears to prevent pendant conformations, providing an analog with comparable anti-HIV-1 and anti-HIV-2 activities to the parent m-phenylenebis(methylene)-linked bicyclam. The results of this study have been used to develop a quantitative structure-activity relationship model with improved predictive capability in order to aid the design of antiviral bis-azamacrocyclic analogs.