Paclitaxel (Taxol)-Induced Killing of Leishmania major in Murine Macrophages

Abstract

The antitumor drug paclitaxel (Taxol) has been demonstrated to be a lipopolysaccharide mimetic in murine macrophages. In this study, the capacity of paclitaxel to activate macrophages to become microbicidal for Leishmania major was examined. Paclitaxel and gamma interferon synergized to kill intracellular L. major in Lpsⁿ, but not Lps^d, macrophages by a nitric oxide (NO·)-dependent mechanism.

In 1990, Ding and colleagues (3, 5) demonstrated that the antitumor agent paclitaxel (Taxol) induced in murine macrophages secretion of tumor necrosis factor alpha (TNF-α) and involution of TNF receptors, two actions also elicited by gram-negative lipopolysaccharide (LPS). They also demonstrated that, like LPS, paclitaxel effects were restricted to macrophages derived from mice that possessed normal LPS sensitivity (e.g., Lpsⁿ) and were not observed in Lps^d macrophages (3). Subsequent studies extended the LPS-mimetic activities of paclitaxel to include induction of LPS-inducible genes, secretion of other LPS-inducible cytokines, tyrosine phosphorylation of mitogen-activated protein kinases, translocation of NF-κB, and autophosphorylation of Lyn kinase (2, 4, 8, 24; reviewed in references 18 and 28). The finding that LPS structural antagonists blocked the LPS-mimetic activities of paclitaxel suggested that paclitaxel and LPS may share a common signaling apparatus (17). Finally, studies in which certain paclitaxel analogs were found not to induce LPS-like effects, yet still induce the well-characterized microtubule-stabilizing effects of paclitaxel, led to a functional dissociation of these two phenomena (12, 27). The latter findings were strengthened by the finding that paclitaxel induced normal microtubule bundling in macrophages derived from C3H/HeJ (Lps^d) mice in spite of a failure of paclitaxel to induce LPS-inducible actions in vitro (14).

Another activity of LPS that is shared by paclitaxel is the capacity to synergize with gamma interferon (IFN-γ) to induce tumoricidal activity in vitro, which was found to be dependent upon the induction of inducible nitric oxide synthase (iNOS) mRNA and the subsequent release of NO· (16). This finding led us to hypothesize that paclitaxel might act analogously to kill NO-˙sensitive intracellular pathogens. Therefore, we examined the effect of paclitaxel on the survival of the intracellular parasite Leishmania major. As was observed for the induction of macrophage tumoricidal activity, paclitaxel synergized with IFN-γ to induce a NO·-dependent inhibition of intracellular parasite replication. However, except at extremely high concentrations of paclitaxel, this reduction in survival was not apparent on parasites cultured in vitro in the absence of macrophages, suggesting that a direct effect on parasite microtubule formation is unlikely to be the principal cause. These data suggest that the direct cytotoxic effects of paclitaxel usually ascribed to its β-tubulin binding can be superseded by the activation of macrophages to produce microbicidal mediators, such as NO·.

Standard methods were utilized for this study. Briefly, thioglycollate-induced peritoneal exudate macrophages from 5- to 6-week-old female C3H/OuJ and C3H/HeJ mice (Jackson Laboratory, Bar Harbor, Maine), iNOS knockout (KO) mice (13) (the kind gift of Carl Nathan, Cornell University, New York N.Y.), and (C57BL/6 × 129)F₁ control mice (Jackson Laboratory) were cultured as described elsewhere (14, 15, 24). The iNOS KO mice used in our experiments were obtained from homozygous inbreeding in the F₂ generation (129SvEv × C57BL/6). The experiments reported herein were conducted according to the principles set forth in Guide for the Care and Use of Laboratory Animals (11). Macrophages were plated in 24-well tissue culture plates at a final concentration of 10⁶ cells/well. Macrophages were allowed to adhere for ∼4 h, washed gently to remove nonadherent cell types, and then treated as indicated below and in the figure legends. Protein-free Escherichia coli K235 LPS was prepared by the hot phenol-water extraction method of McIntire et al. (20), and protein-rich, butanol-extracted LPS (LPS-But) was prepared by the method of Morrison and Leive (21). Murine recombinant IFN-γ was provided by Genentech, Inc. (South San Francisco, Calif.). Paclitaxel was provided by the Drug Synthesis and Chemistry Branch, National Cancer Institute, National Institutes of Health (NIH). l-N-Monomethylarginine (L-NMMA) was purchased from Sigma Chemical Co. (St. Louis, Mo.). Metacyclic L. major promastigotes (kindly provided by David Sacks, NIH) were prepared as described elsewhere (26). Macrophages were infected with promastigotes at a multiplicity of infection of ∼1.

Parasite numbers were quantified in macrophage cultures lysed by incubation for 30 min in 0.1% saponin at 37°C. Lysates were titrated in complete M199 medium (GIBCO, Grand Island, N.Y.) (supplemented with 2 mM glutamine, antibiotics, 30% fetal calf serum, and 50 mM 2-β-mercaptoethanol) over blood agar in 96-well plates (6). Wells were scored after 1 and 2 weeks as positive or negative for the presence of parasites. Values from titrations were expressed as percentages of the numbers of parasites recovered from control, unmanipulated cultures.

NO· production was assayed by determining the increase in nitrite concentration by the Griess reaction adapted to microwell plates, with a sodium nitrate standard (19, 23). TNF-α levels were measured by a two-site sandwich enzyme-linked immunosorbent assay (1).

RESULTS AND DISCUSSION

Macrophages from Lpsⁿ C3H/OuJ and Lps^d C3H/HeJ macrophages were cultured with paclitaxel in the absence or presence of IFN-γ overnight, prior to infection with L. major. Culture supernatants were collected and the cells were lysed for enumeration of parasites at 24 h after infection. As found in previous studies (16) carried out in the absence of parasites, 1 and 10 μM paclitaxel synergized with IFN-γ to release NO· in C3H/OuJ but not C3H/HeJ macrophages (Fig. 1A). Indeed, macrophages derived from LPS-hyporesponsive C3H/HeJ mice failed to show any increase in NO· release above that induced by IFN-γ alone. Figure 1B shows the responses of the same macrophage cultures to either a highly purified protein-free LPS preparation (10 ng/ml), demonstrated in previous studies to discriminate clearly between Lpsⁿ and Lps^d macrophages, or a protein-rich LPS preparation, LPS-But (10 μg/ml), which stimulates both Lpsⁿ and Lps^d macrophages due to the presence of contaminating endotoxin-associated proteins (9, 10, 15). As expected from previous studies, C3H/OuJ macrophages responded synergistically to both LPS and LPS-But in combination with IFN-γ, whereas the C3H/HeJ macrophages responded only to LPS-But plus IFN-γ to release NO·. Macrophages stimulated with paclitaxel or LPS plus IFN-γ in the absence of parasites consistently produced levels of NO· comparable to those released in the presence of parasites (data not shown).

FIG. 1.

Induction of NO· release and L. major killing in C3H/OuJ and C3H/HeJ macrophages by paclitaxel (Tx) or LPS and IFN-γ. Macrophages were treated with combinations of paclitaxel (1 or 10 μM) and/or 5 U of IFN-γ per ml (A and C) or with LPS (protein free) (10 ng/ml) or LPS-But (protein rich) (10 μg/ml) and IFN-γ (B and D) and then infected with L. major. NO· was measured from the supernatants (A and B), and the number of L. major parasites was quantified from macrophage lysates (C and D). Results are derived from a single experiment representative of six separate experiments.

Figure 1C and D illustrates the corresponding recoveries of L. major from C3H/OuJ and C3H/HeJ macrophages stimulated as described for Fig. 1A and B. Killing of L. major paralleled the production of NO· in the same macrophage cultures, illustrating that, like LPS, paclitaxel synergizes with IFN-γ to elicit a microbicidal effect in Lpsⁿ macrophages.

Paclitaxel has recently been demonstrated to inhibit the growth of Plasmodium spp. (25, 27) and Toxoplasma gondii (7) directly in vitro, an effect that has been attributed to its ability to block microtubule depolymerization, which, in turn, interferes with mitosis and parasite growth. Prolonged treatment of cell-free cultures of L. major (>72 h) with higher concentrations of paclitaxel (35 μM) led to a decrease in the number of viable parasites recovered at the end of culture. When observed microscopically, a significant proportion of the parasites thus treated appeared to be rounded and slightly enlarged, with greatly decreased motility. Control cultures did not show these changes, suggesting a direct effect of paclitaxel on parasite viability in addition to the effects mediated through macrophage activation (data not shown). However, since these changes were not apparent at lower doses of paclitaxel (either in the absence or presence of IFN-γ) or with shorter incubation times, it appears that a direct effect of paclitaxel on parasite viability cannot account for the decreased parasite viability observed in the presence of macrophages.

Figure 2 illustrates the time course of paclitaxel-plus-IFN-γ-induced killing of intracellular L. major in C3H/OuJ and C3H/HeJ macrophages. Macrophages were treated with 10 μM paclitaxel and 5 U of IFN-γ per ml, infected 4 h later, and lysed at various times postinfection to determine the rate of parasite killing. Figure 2A shows the levels of nitrite released into the supernatant at the indicated times postinfection, while Fig. 2B shows percent parasite recovery from the same macrophages. As was observed in Fig. 1A and 1C, paclitaxel plus IFN-γ synergized to induce NO· release that was correlated with percent parasite kill. By 24 h postinfection, the parasite recovery in treated versus untreated macrophages was reduced by >90% in C3H/OuJ macrophages. Again, C3H/HeJ macrophages did not respond to paclitaxel plus IFN-γ to produce NO· or to be rendered microbicidal.

FIG. 2.

Time course for induction of NO· release and parasite killing in C3H/OuJ and C3H/HeJ macrophages treated with paclitaxel (10 μM) plus IFN-γ (5 U/ml). Results are derived from a single experiment representative of six separate experiments.

Figure 3 shows that the ability of paclitaxel plus IFN-γ to stimulate NO· release and to kill L. major was reversed in the presence of L-NMMA, an inhibitor of iNOS. These data indicate that the correlation observed in Fig. 1 and 2 between release of NO· and parasite killing is the result of a cause-and-effect relationship. To confirm and extend these findings, macrophages derived from iNOS KO (−/−) mice or wild-type (C57BL/6 × 129)F₁ control (+/+) mice were treated with either LPS plus IFN-γ or paclitaxel plus IFN-γ and infected, and levels of NO·, TNF-α, and parasite killing were measured. Figure 4 (top panel) illustrates that under conditions in which the control wild-type macrophages produced normal levels of NO· in response to LPS or paclitaxel plus IFN-γ, iNOS KO macrophages failed to produce detectable NO·. In contrast, levels of stimulated TNF-α were comparable for the two strains’ macrophages (middle panel). Finally, parasite recovery was reduced to <10% of that measured in medium-treated, wild-type macrophages upon stimulation of +/+ macrophages with LPS or paclitaxel plus IFN-γ. In contrast, iNOS KO macrophages killed L. major only marginally in the presence of paclitaxel plus IFN-γ (<20% of wild-type killing). These data indicate that an intact iNOS generating system is required for efficient killing of L. major induced in macrophages by paclitaxel plus IFN-γ. Similar experiments were carried out with macrophages derived from mice with targeted mutations in both the type I and type II TNF receptor genes (TNFp55p75−/−) and with control C57BL/6J macrophages treated with a neutralizing anti-TNF monoclonal antibody. However, neither NO· release nor parasite killing induced by LPS or paclitaxel (alone or in combination with IFN-γ) was altered (data not shown). Thus, the NO·-mediated killing of L. major by activated macrophages appears to be independent of TNF, consistent with very recent work by Nashleanas et al. (22) that demonstrated that TNFp55p75−/− mice clear L. major normally in vivo.

FIG. 3.

Effects of L-NMMA on NO· release and killing of L. major in paclitaxel (10 μM)- and IFN-γ (5 U/ml)-treated C3H/OuJ and C3H/HeJ macrophages. Results are derived from a single experiment representative of two separate experiments.

FIG. 4.

Induction of NO· release, TNF-α production, and killing of L. major by LPS (10 ng/ml) or paclitaxel (10 μM) and IFN-γ (5 U/ml) in (C57BL/6 × 129)F₁ (+/+) and iNOS KO (−/−) macrophages. Macrophages were treated as described in the legend to Fig. 1.

Taken collectively, these data demonstrate that although paclitaxel may have direct effects on parasite viability as a result of its well-characterized ability to bind β-tubulin in the context of microtubules and prevent their depolymerization (7, 24, 26), paclitaxel, alone or in synergy with IFN-γ, elicits parasite killing by highly activated murine macrophages. This mechanism, like macrophage tumoricidal activity induced by paclitaxel and IFN-γ, is NO· dependent and not inducible in mice that express defective iNOS or Lps genes. Thus, these findings extend the LPS-mimetic properties of paclitaxel to the induction of microbicidal activity. Although the LPS-mimetic effects of paclitaxel have been described largely for murine macrophages, recent evidence suggests that paclitaxel may also modulate gene expression and cytokine secretion in human cell types, including unprimed monocytes (29). Thus, it is possible that patients undergoing paclitaxel chemotherapy, who are likely to be immunosuppressed and to exhibit increased susceptibility to opportunistic pathogens, may benefit not only from paclitaxel’s antitumor effects but also from potential direct or indirect antimicrobial actions of this drug.