Novel Synthetic Polyamines Are Effective in the Treatment of Experimental Microsporidiosis, an Opportunistic AIDS-Associated Infection

Synthesis of polyamine analogues.The syntheses of the tetramines and pentamines shown in Fig. 2 were recently reported (31, 37). The syntheses of the oligoamines shown in Fig. 3 were carried out by a general procedure. ¹N-Monoethyl tetramides, pentamides, hexamides, and heptamides in which the amino groups are protected by mesitylenesulfonyl residues (prepared as described elsewhere [31, 37]) were dimerized by reaction with (E)- or (Z)-2-butene-1,4-diyl-bis(mesitylenesulfonate) to give the corresponding octamides, decamides, dodecamides, and tetradecamides. The protecting groups were then removed and the oligoamines were isolated as hydrochlorides. In a typical reaction, a tetramide (3 mmol) and the butenesulfonate (1.5 mmol) were mixed in 20 ml of dimethylformamide kept at 5°C, sodium hydride (3.6 mmol) was added, and the mixture was kept at 20°C for 18 h. The solvent was evaporated; the residue was partitioned between chloroform and a concentrated ammonium chloride solution; and the product, extracted into the organic layer, was purified by flash chromatography on silica gel (hexane-ethyl acetate [6:4]). The octamide thus obtained (83% yield) was deprotected by dissolution (0.50 mmol) in methylene chloride (20 ml), followed by addition of phenol (37 mmol) and 33% hydrogen bromide in glacial acetic acid (20 ml). The mixture was kept at 20°C for 18 h; further workup was by previously published procedures (32). Octamines SL-11157 and SL-11158 were thus obtained (as their hydrochlorides) in 89% yields. Reduction of the double bond was achieved by hydrogenation of an aqueous solution of either SL-11158 or SL-11157 over platinum oxide at 50 lb/in² for 2 h; SL-11160 was thus obtained in an 85% yield. Analytical data supported the assigned structures.

• Open in new tab
• Download powerpoint

FIG. 2.

Structures of polyamine analogues (tetramines and pentamines) with restricted conformations.

• Open in new tab
• Download powerpoint

FIG. 3.

Structures of oligoamines and pentamidine analogues.

The three pentamidine analogues (SL-11134, SL-11136, and SL-11137) were obtained by modifying known procedures. Thus, 1,5-bis(p-cyanobenzyl)-diaminobutane, 1,6-bis(p-cyanobenzyl)-diaminopentane, and 1,7-bis(p-cyanobenzyl)-diaminohexane were obtained by previously described procedures (33). The corresponding amidines were prepared by the following procedure. 1,6-Bis(p-cyanobenzyl)-diaminopentane (1 mmol) was dissolved in 6 ml of anhydrous tetrahydrofuran under argon, lithium hexamethyl disilazide (4 mmol) dissolved in tetrahydrofuran was added, and the mixture was stirred at 20°C for 2 h. It was then cooled at 5°C; 10 ml of 1 M hydrogen chloride in ether was added; and the white precipitate was filtered, washed with ether, and crystallized from ethanol-ether; 0.5 g (94% yield) of SL-11134 tetrahydrchloride was obtained. Analytical data supported the assigned structures.

E. cuniculi culture and drug assay. E. cuniculi culture and the drug assay were performed as described previously (1). RK-13 cells (5 × 10⁴ cells/well) were plated into each well of a Falcon multiwell, 24-well tissue culture plate and were allowed to incubate for 3 days until they were confluent (8). These cells were then infected with spores of E. cuniculi at a multiplicity of infection of 4:1 (10⁶ organisms per well). This protocol resulted in infection of 50 to 80% of cells in the absence of drug treatment. Drugs were added to duplicate wells and the plates were incubated for 7 days, with the drug-containing medium changed at 3 and 6 days. On day 8, the wells were fixed overnight and stained with Giemsa, and the cells were counted with an inverted microscope at ×400 magnification. Five confluent fields (240 cells/field) were counted for control as well as for drug-treated wells. The percentage of infected cells in the presence of the compound was compared to the percentage of infected control cells. Fifty percent inhibitory concentrations (IC₅₀s) were expressed as micromolar of drug (8). Toxicity to the host cell monolayer was determined by examination of Giemsa-stained, uninfected, drug-treated monolayers for abnormal morphology, such as deformed fibroblasts, ragged holes in the monolayer, detachment of the monolayer, and evidence of lysis.

Mouse model of microsporidiosis.Two well-validated models of microsporidiosis were used, one with nude (nu/nu) BALB/c mice (10) which are completely immunosuppressed and the other with C57BL/6J CD8-knockout (ΔCD8) mice (19). The latter immune defect is specific, allowing growth of microsporidia. ΔCD8 mice (19) were infected with 3 × 10⁷ spores (intraperitoneally [i.p.]) 24 h before treatment, and nude mice were infected with 10⁶ spores (i.p.) 24 h before treatment (10). Animals were treated i.p. with 1 to 10 mg of drug per kg of body weight for 5 days, followed by 2 days without drug and then 5 more days of drug treatment. This protocol has been used for the administration of these compounds in animal tumor models and for initial toxicity testing. Animals were considered cured of infection if they survived more than 28 days postinfection with no evidence of the presence of microsporidia. Tissues were obtained from mice at the ends of the observation periods and were then fixed and embedded by standard protocols. Tissue sections were stained for microsporidia with hematoxylin-eosin and tissue chromotrope stains. These sections were examined in a blinded fashion for pathology and the presence of organisms.

PCR assay for microsporidia.As published previously (27), the parasite loads in the tissues were estimated by using a semiquantitative PCR with DNA extracted from tissues (liver and kidney) at the termination of each experiment. DNA was extracted by use of the Qiamp tissue kit (Qiagen, Chatsworth, Calif.), and 3 μg of each sample was analyzed. The PCR was performed with a pair of primers, 5′-ATGAGAAGTGATGTGTGCG-3′ and 5′-TGCCATGCACTCACAGGCATC-3′, that amplify a 549-bp fragment of the small-subunit rRNA gene of E. cuniculi (GenBank accession no. L17072 ). A 510-bp competitive internal standard was generated by the method of Kirisits et al. (21), as reported previously (27). PCR was performed using the following conditions: 35 cycles of denaturation at 94°C for 45 s, annealing at 53°C for 1 min, and elongation at 72°C for 30 s. Amplification was performed with the Eppendorf Scientific Inc., Westbury, N.Y.) master kit and with dGTP, dATP, dTTP, and dCTP each at a concentration of 0.2 mM and each E. cuniculi-specific primer at a concentration of 0.4 μM. Various amounts of the internal standard were added to each reaction mixture to determine the relative amounts of parasite rRNA in each sample. Amplicons were analyzed by electrophoresis on a 1.5% agarose gel and were visualized with ethidium bromide. The number of parasites was determined by amplification of a known amount of parasite with a dilution of internal standard by using the same PCR conditions described above (27). The internal standard was used as a quality control to determine the ratios of the amount of competitive template-directed product to the amount of microsporidian template-directed product.

Toxicity studies of polyamine analogues in mice.All agents were administered i.p. for 5 days, followed by 2 days without drug and then another 5 days of treatment. The mice (groups of three mice each) weighed 23 to 33 g and were matched as to approximate weight. The weight loss 1 day after the end of treatment is expressed in comparison to the weight at the start of treatment. Control (untreated) animals gained 1 to 3 g in 14 days, depending on their starting weight. Daily observations on clinical toxicity (e.g., decreased movement, ruffled fur, and decreased grooming) were obtained for all animals. The general dose range used was 1 to 150 mg/kg/day, depending on the agent.

Polyamine analogs.Four classes of polyamine analogues were evaluated for activity against E. cuniculi, the species used in two murine models of microsporidiosis. The first class comprises the tetramines (Fig. 2), homologues of spermine in which the external aminopropyl residues present in spermine (Fig. 1) were replaced by aminobutyl residues, making a homospermine backbone. The terminal primary amino groups were N-ethylated to prevent oxidation by serum aminooxidases (6). Spermine and homospermine are freely rotating molecules that can assume myriad conformations. The introduction of alicyclic residues in the homospermine backbone, as in SL-11093, SL-11098, SL-11099, and SL-11100 (Fig. 2), restricts free rotation at the central segment of the molecule. This is also the case when one or two cis double bonds are introduced, as in SL-11102, SL-11114, SL-11118, and SL-11119. The introduction of a triple bond or a 1,2-dimethylbenzene residue in the central segment, as in SL-11103 and SL-11104, respectively, confers rigidity to that part of the molecule. Spermine and its homologues are regioselective binders (i.e., they bind selectively to a domain or region) of DNA (12), tRNA (15), chromatin (4), and rRNA (7). Spermine analogues that are conformationally restricted inhibit tumor cell proliferation (32), very likely by bending and kinking the conformations of the nucleic acids to which they bind. The potential utility of conformationally restricted homospermine analogues as inhibitors of replication of microsporidia was therefore explored.

The pentamines (Fig. 2) were the second class of polyamine analogs investigated. Pentamine SL-11061 (also called BE-4-4-4-4), designed in our laboratories several years ago, is a powerful inhibitor of tumor cell replication (3). Its free rotating conformation was also restricted by the introduction of cis double bonds into its hydrocarbon skeleton, as in SL-11121 to SL-11123 and SL-11126 to SL-11130.

Hydrophilic groups were also introduced into the pentamine structure: an alcohol residue in SL-11141 and a nonchiral amino acid in SL-11143.

The third class of polyamine analogues used in the present study consisted of the oligoamines, SL-11144, SL-11157, SL-11158, SL-11159, SL-11160, SL-11172, SL-11175, and SL-11207 (Fig. 3). We coined the name oligoamines for this new class of synthetic octa-, deca-, dodeca-, and tetradecamines since the chemically more correct name of “polyamines” had historically been applied to the natural tetra-, tri-, and diamines (Fig. 1). The rationale behind the synthesis of oligoamines is the well-known fact that spermine (a tetramine), at a concentration range of 0.05 to 0.1 mM and at near physiological ionic strength, leads to the collapse of DNA (29). Oligoamines were found to condense DNA at much lower concentrations (2 to 4 μM) and are indeed powerful inhibitors of human tumor cell proliferation.

The fourth class of polyamine derivatives tested consisted of three diamino analogs of pentamidine (SL-11134, SL-11136, and SL-11137; Fig. 3). Pentamidine is a well-known antiprotozoal drug (7).

In vitro susceptibilities to polyamine analogs.The results of the in vitro screening are summarized in Table 1. E. cuniculi was used as a model organism for determination of the antimicrosporidial activities of the polyamine analogs. It is not feasible to screen agents directly for activity against E. bieneusi due to the absence of in vitro cultivation systems or small animal models. E. cuniculi has been reported in cases of encephalitis, disseminated infection, and hepatitis in humans (44). In previous work (8, 44), other compounds screened have demonstrated similar inhibition profiles against all of the Encephalitozoonidae. In vitro activity was assayed as described previously (8). Polyamine analogs were added to E. cuniculi-infected RK-13 cells, and the cultures were incubated for 7 days, with the drug-containing media changed on days 3 and 6. The percentages of infected cells containing clusters of parasites in treated and untreated cultures were compared.

The analogues in the first class, namely, the conformationally restricted tetramines SL-11093, SL-11098 to SL-11100, SL-11102 to SL-11104, SL-11114, SL-11118, and SL-11119, were minimally active. With the exception of the fully saturated derivative SL-11061, the analogues in the second class, the pentamines SL-11061, SL-11121 to SL-11123, SL-11126 to SL-11130, SL-11141, and SL-11143, were not very active either; SL-11061, however, inhibited parasite replication at relatively low concentrations (IC₅₀, 42 to 43 μM). The pentamidine analogues (SL-11134, SL-11136, and SL-11137) were also inactive. In contrast, the oligoamines seemed to hold promise as antimicrosporidial agents. The octamines (SL-11157, SL-11158, and SL-11160) displayed low IC₅₀s, the decamines (SL-11144 and SL-11159) were even more active, while the dodecamine (SL-11172) and the tetradecamine (SL-11175) were also very active in the inhibition of parasite growth.

Interestingly, SL-11207, although a nonamine, was effective only at a somewhat higher range of IC₅₀s, similar to those of SL-11061. This could be because the more active oligoamines are a continuous catenation of NH₂⁺ groups (polyamines are protonated at neutral pH) that can wrap and bind (either by charge or by hydrogen bonds [17]) to DNA and cause it to collapse. In SL-11207, the polycationic chain is interrupted by a nonprotonated tertiary amine residue and therefore might not interact with the phosphate groups of DNA in the correct manner. The protons of two NH₂⁺ residues separated by a zig-zag chain of four methylene groups are at a distance of 7.33 Å from each other, a value that fits the distance between successive phosphate anions in the DNA helix (7.3 Å), as well as the distance between phosphate groups (7.96 Å) found in spermine phosphate crystals (36).

Toxicity studies.In tandem with the in vitro studies, we determined the tolerance of select polyamine analogues by mice as a prelude to in vivo studies. The data are shown in Table 2. SL-11061 was tolerated up to at least 10 mg/kg when it was given i.p. daily for 5 days, followed by 2 days with no treatment and then a second 5-day course of therapy. Animals were observed for weight loss and lethality for at least 20 days following drug injection. Host tolerances of the oligoamines SL-11158 and SL-11144 were similar (Table 2). Polyamines with alicyclic rings, e.g., SL-11093 and SL-11099, were tolerated up to 50 mg/kg when the schedule described above was used. The weight loss in animals treated with SL-11093 was negligible.

Efficacies of polyamine analogs in experimental infections.In vivo therapeutic studies were then undertaken with the same strain of E. cuniculi that was used in vitro. The studies were carried out in separate laboratories with C57BL/6J ΔCD8 mice in one laboratory and BALB/c nude (nu/nu) mice in the other (Table 3) (10, 19). In these studies, immunodeficient mice were inoculated (i.p.) with 1 × 10⁶ to 3 × 10⁷ spores and treated (i.p.) with the treatment schedule described above, beginning 24 h after infection. SL-11061 at 1.25 mg/kg or at 10 mg/kg/day cured 10 of 11 BALB/c nu/nu animals, with 1 animal dying of infection on day 20 (Table 3, experiments 1 and 2), and SL-11061 at 1.25 mg/kg cured 2 of 2 C57BL/6J ΔCD8 animals (Table 3, experiment 4). The oligoamines were even more effective. SL-11144 was curative at 1.25 mg/kg (two of two animals) or at 2.5 mg/kg/day (three of three animals), while SL-11158 at 5 mg/kg/day cured four of four animals (Table 3, experiments 3, 4, and 5). In view of the very low toxicity of SL-11093, it was also assayed in vivo. Even though it had only moderate activity in vitro (Table 1), it was curative in vivo when it was given at 5 mg/kg/day by the same treatment schedule mentioned above; it cured five of five animals. Animals considered cured in these studies were observed for evidence of parasitemia for 30 days after untreated control animals died and were also examined for evidence of parasites in liver by histologic staining as well as by PCR (by a semiquantitative assay for E. cuniculi [27]), with no parasites being observed by either technique.