Efficacy of TAK-457, a Novel Intravenous Triazole, against Invasive Pulmonary Aspergillosis in Neutropenic Mice

Abstract

TAK-457 is an injectable prodrug of TAK-456, which is a novel oral triazole compound with potent antifungal activity. The in vivo efficacy of TAK-457 was evaluated in two models of invasive pulmonary aspergillosis with CDF₁ mice and CBA/J mice with transient neutropenia induced by cyclophosphamide. Against the infection in CDF₁ mice, treatment with 10 mg of TAK-457 and 1 mg of amphotericin B/kg reduced the fungal burden in lungs and rescued all mice. In the infection model with CBA/J mice, TAK-457 at a dose of 10 mg/kg significantly prolonged the survival time of mice, showing significant reduction of lung chitin levels and the plasma β-d-glucan levels. On the other hand, amphotericin B at 1 mg/kg which was a maximum tolerable dose showed slight but not significant prolongation of survival time of mice, although it also reduced the lung chitin levels and the plasma β-d-glucan levels to a lower extent but still significantly. These results suggest that TAK-457 is a promising candidate for development for the treatment of invasive aspergillosis in humans.

Aspergillus fumigatus is a fungal pathogen of major clinical concern (3, 10, 30). Invasive aspergillosis is a leading cause of morbidity and mortality in immunocompromised patients, particularly in patients receiving chemotherapy against hematologic malignancy or immunosuppressive therapy for bone marrow transplant (8, 18). Its incidence has increased dramatically over the past two decades concomitantly with an increase in the number of immunocompromised hosts (1, 8, 15). Since invasive aspergillosis is a serious disease and most patients with invasive aspergillosis are not able to take drugs orally, an injectable antifungal drug is indispensable. Amphotericin B has been used for the treatment of invasive aspergillosis despite its severe toxicity because other injectable antifungal agents such as fluconazole are ineffective against invasive aspergillosis (11, 25, 27, 32). Although new formulations of amphotericin B have been developed to reduce renal toxicity, its improvement is still insufficient (28, 33, 34, 35). An intravenous formulation of itraconazole, a triazole antifungal agent, has been recently introduced. However, the effectiveness of itraconazole, like amphotericin B, is not enough against invasive aspergillosis in the severely immunosuppressed patients (4, 5, 8, 9, 18). Accordingly, new injectable antifungal agents with low toxicity and high effectiveness against invasive aspergillosis are strongly demanded. We have previously reported that TAK-456, an oral triazole antifungal agent, has potent in vitro and in vivo antifungal activities against major pathogenic fungi, including Aspergillus species, Candida species, Cryptococcus neoformans, and dermatophytes (12, 16; N. Tsuchimori, R. Hayashi, N. Kitamoto, Y. Iizawa, T. Kitazaki, K. Itoh, and K. Okonogi, unpublished data). The solubility of TAK-456 in water, however, was not sufficient for injectable formulation. TAK-457, a quarternary triazolium salt with a biologically labile acetoxymethyl moiety, is an injectable prodrug of TAK-456 with improved water solubility (Fig. 1). TAK-457 was rapidly converted to TAK-456 in the blood of mice, rats, and humans (13). A rapid increase in plasma levels of TAK-456 was observed after intravenous administration of TAK-457 into rats (13). Therefore, it is expected that TAK-457 would be a useful injectable drug against serious fungal infections such as invasive pulmonary aspergillosis. This report describes the in vivo activity of TAK-457 in mouse models of invasive pulmonary aspergillosis.

FIG. 1.

Chemical structure of TAK-457.

(This work was presented in part at the 40th Interscience Conference on Animicrobial Agents and Chemotherapy, Toronto, Canada, 17 to 20 September 2000 [Y. Iizawa, R. Hayashi, N. Kitamoto, N. Tsuchimori, K. Asai, K. Itoh, and K. Okonogi, Abstr. 40th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1086].)

MATERIALS AND METHODS

Antifungal agents.

TAK-457 was synthesized at Takeda Chemical Industries, Ltd., Osaka, Japan. Doses of TAK-457 were shown as doses of its parent compound TAK-456. Fluconazole (Pfizer, Inc., Tokyo, Japan) was extracted from commercially available intravenous injection in our laboratories. The purity of all azoles was confirmed to be >99% by high-performance liquid chromatography. Amphotericin B was obtained commercially from Bristol-Myers Squibb, Inc., Tokyo, Japan. TAK-457, fluconazole, and amphotericin B were dissolved in 5% glucose solution (Otsuka Co., Tokyo, Japan).

Organisms.

A. fumigatus 437 (Roche culture collection) was a generous gift from A. Polak, Hoffman-La Roche, Basel, Switzerland. The MICs of TAK-456, fluconazole, and amphotericin B, determined by the NCCLS method (23), were 1, >64, and 0.5 μg/ml, respectively.

Animals.

We used two strains of mice: 5-week-old Crj:CDF₁ female mice and CBA/JNCrj female mice (Charles River Japan, Inc., Kanagawa, Japan) weighing 16 to 22 g. Since it is well known that there is a strain difference in susceptibility to an experimental infection (7, 14, 31), we established two types of infection models with the two strains of mice. Indeed, the response to pulmonary aspergillosis was quite different in the two strains. We described their course of infection in the results section. They were caged in groups of 5 to 10 and given food and water ad libitum. When mice were rendered transient neutropenia, 200 mg of cyclophosphamide (Shionogi Co., Osaka, Japan)/kg was injected intraperitoneally 4 days before fungal challenge. It has been reported that a 200-mg/kg treatment of cyclophosphamide to mice results in an immediate depression of peripheral blood leukocytes (22). All procedures were performed in accordance with the standards approved by Takeda Institutional Animal Ethics Committee (approval no. E1-019).

Preparation of inocula.

A. fumigatus was grown on a malt agar (Difco Laboratories, Detroit, Mich.) plate for 10 days at 28°C. Conidia were collected with 0.1% Tween 80 (Sigma Chemical Co., St. Louis, Mo.) in saline, washed twice with 0.01% Tween 80 in saline, suspended in Trypticase soy broth (Becton Dickinson Co., Cockeysville, Md.) containing 0.01% Tween 80 and 10% glycerol, and stored at −80°C. The suspension of conidia was thawed at room temperature and diluted with 0.01% Tween 80 in saline shortly before use.

Invasive pulmonary aspergillosis. (i) Infection procedure.

Neutropenic CDF₁ mice and CBA/J mice were used for invasive pulmonary aspergillosis. On the day of infection, the mice were anesthetized by administering ketamine hydrochloride (Sankyo Inc., Tokyo, Japan) at 100 mg/kg intraperitoneally and were inoculated intranasally with 0.04 ml of conidial suspension of A. fumigatus 437 (8.4 × 10⁵ to 1.1 × 10⁶ CFU/mouse). To prevent bacterial infection, the mice were given drinking water containing 500 mg of tetracycline hydrochloride (Sigma) per liter, as well as 10 mg of vancomycin (Shionogi) and 10 mg (imipenem equivalent) of imipenem-cilastatin (Banyu Co., Tokyo, Japan)/kg subcutaneously on the day after inoculation of A. fumigatus conidia. TAK-457 and reference antifungal agents were administered into the lateral tail vein of mice (0.2 ml over 5 s) once a day for 5 days starting 2 h after the fungal challenge. Infected control mice received 5% glucose intravenously. The mice were observed for their survival over 7 days after infection.

(ii) Determination of lung weight and fungal burden in lungs.

The lung weight and fungal burden in lungs of CBA/J mice and CDF₁ mice were determined 2 and 3 days after infection, respectively. Both mice were killed by bleeding from abdominal arteries and veins under ether anesthesia 24 h after the last treatment, and then the lungs were removed aseptically. After we measured the weight, the lungs were homogenized in 3 ml of distilled water. Lung homogenates were serially diluted with distilled water, and 0.1 ml of the diluted homogenates was plated onto Sabouraud dextrose agar (Eiken Chemical Co., Tokyo, Japan). The plates were incubated at 28°C for 40 to 48 h, and the CFU of A. fumigatus were enumerated. The lower detection limit is 150 CFU/lung.

(iii) Determination of β-d-glucan level in plasma.

CBA/J and CDF₁ mice were killed 2 and 3 days after infection, respectively, as described above. Blood samples were taken by drawing into heparinized plastic syringe that was inserted into the abdominal artery under ether anesthesia, and the level of β-d-glucan in plasma was measured by a (1→3)-β-d-glucan determination assay kit, the Fungitec G Test (Seikagaku Co., Tokyo, Japan), according to the manufacturer’s instructions. Briefly, a plasma sample was added to 0.32 M perchloric acid and incubated at 37°C for 20 min. After centrifugation, the supernatant was neutralized with 0.18 mol of NaOH/liter. The samples were then added to the G test reagent, and the mixture was incubated at 37°C for 30 min. The absorbance values at 540 and 620 nm were measured after diazo coupling. The lower detection limit was 45 pg/ml.

(iv) Determination of chitin level in lungs.

CBA/J and CDF₁ mice were killed by bleeding from abdominal arteries and veins under ether anesthesia 2 and 3 days after infection, respectively, and the lungs were removed aseptically. The chitin level in the lungs, which has been reported to correlate with the weight of Aspergillus hyphae (29), was determined as described previously (17). Briefly, lung homogenates were centrifuged. The pellets were resuspended in sodium lauryl sulfate (3% [wt/vol]) and heated at 100°C for 15 min. After centrifugation, pellets were washed with distilled water, resuspended in KOH (120% [wt/vol]), and then heated at 130°C for 60 min. After cooling, ice-cold ethanol (75% [vol/vol]) was added. Tubes were kept on ice for 15 min and Celite545 was added. After centrifugation, pellets were washed and suspended in solution containing NaNO₂ (5% [wt/vol]) and KHSO₄ (5% [wt/vol]). Samples were mixed gently for 15 min and then centrifuged. The supernatant was mixed with ammonium sulfate (12.5% [wt/vol]) and 3-methyl-2-thiazolone hydrazone HCl monohydrate. The mixture was heated at 100°C for 3 min. After cooling, FeCl₃ · 6H₂O (0.83% [wt/vol]) was added to the mixture, and the opitical density at 650 nm was measured after 25 min. Glucosamine (10 μg/ml) was used as a standard. The chitin level was expressed as an arbitrary unit based on the glucosamine equivalent.

Statistical analysis.

Survival time of mice was analyzed by the log-rank test. A one-way analysis of variance was performed on the ranks of the lung weight, fungal burden in lungs, chitin level in lungs, and β-d-glucan level in plasma. Pairwise comparisons of each treatment group to the control group were adjusted by using Dunnett’s multiple-comparison test. The statistical significance was defined as P < 0.05.

RESULTS

(i) Pulmonary aspergillosis in CDF₁ mice.

Effects of TAK-457 and reference antifungal agents on the survival rate of mice in the infection model with CDF₁ mice are shown in Fig. 2. Each group consisted of 10 mice. In this model, local but severe inflammation occurred in lungs of mice after infection and little β-d-glucan, which has been reported to correlate with the extent of Aspergillus infection (19) and is used for the diagnosis in humans (20, 24), was detected (data not shown). Nontreated mice died within 4 days after infection (Fig. 2). All mice survived with TAK-457 treatment at a dose of 10 mg/kg and amphotericin B treatment at a dose of 1 mg/kg, although neither 3 mg of TAK-457 nor 10 mg of fluconazole/kg was effective. In a separate experiment with this model, the lung weights and fungal burdens in the lungs were examined 3 days after infection by using live mice (Table 1). On day 3 after infection, the numbers of survivors were 5 of 15 (untreated group), 5 of 8 (fluconazole-treated group), 6 of 8 (group treated with 3 mg of TAK-457/kg), and 8 of 8 (groups treated with 10 mg of TAK-457 and amphotericin B/kg). The lungs of mice before infection weighed ca. 150 mg (data not shown), and their weights increased twofold or more by day 3 after infection. This increase was suppressed significantly by the treatment with 10 mg of TAK-457 and 1 mg of amphotericin B/kg (Table 1). The fungal burden of mice treated with 10 mg of TAK-457 and 1 mg of amphotericin B/kg was also significantly lower than that of vehicle-treated mice (Table 1).

FIG. 2.

Effects of TAK-457 and reference antifungal agents against pulmonary infection caused by A. fumigatus 437 in neutropenic CDF₁ mice. Antifungal agents were administered intravenously once a day for 5 days starting 2 h after infection (n = 10). Agents: vehicle (•), TAK-457 at 10 mg/kg (○) and 3 mg/kg (◊), amphotericin B at 1 mg/kg (□), fluconazole at a dose of 10 mg/kg (▵).

TABLE 1.

Effects of TAK-457 and reference antifungal agents on lung weight and fungal burden in lungs of neutropenic CDF₁ mice inoculated intranasally with A. fumigatus 437^a

Drug	Dose (mg/kg)	No. of surviving mice/total no. of mice	Mean ± SEMb
			Lung wt (mg)	Fungal burden in lungs (log CFU/g)
Vehicle	0	5/15	420 ± 29.6	5.07 ± 0.07
TAK-457	3	6/8	381 ± 21.9	4.65 ± 0.14
TAK-457	10	8/8	299 ± 46.1**	3.12 ± 0.27**
Fluconazole	10	5/8	387 ± 13.4	5.07 ± 0.06
Amphotericin B	1	8/8	295 ± 11.3**	4.16 ± 0.19*

^aAntifungal agents were administered intravenously once a day on days 0, 1, and 2 of infection, and the weight and the number of A. fumigatus cells were determined for the surviving mice on day 3.

^b*, P < 0.05; **, P < 0.01 (versus the vehicle-treated group).

In a separate experiment, lung chitin levels were examined 3 days after infection with live mice (Fig. 3). The numbers of survivors were four, two, five, and five of five mice in the untreated, fluconazole-treated, TAK-457-treated, and amphotericin B-treated groups, respectively. In contrast to chitin levels of vehicle-treated mice and mice treated with 10 mg of fluconazole/kg, the chitin levels of all mice treated with TAK-457 and three of five mice treated with amphotericin B were below the detection limit (Fig. 3).

FIG. 3.

Effects of TAK-457 and reference antifungal agents on lung chitin levels of neutropenic CDF₁ mice inoculated intranasally with A. fumigatus 437. Antifungal agents were administered intravenously once a day on days 0, 1, and 2 of infection, and chitin levels were determined for the surviving mice on day 3. One vehicle-treated and three fluconazole-treated mice died by day 3 after infection (n = 5). Circles indicate chitin levels of individual mouse, and bars indicate means. *, P < 0.05 versus the vehicle-treated group. AMPH-B, amphotericin B; FLCZ, fluconazole.

(ii) Pulmonary aspergillosis in CBA/J mice.

The effects of TAK-457 and reference antifungal agents on the survival rate of mice in the infection model with CBA/J mice are presented in Fig. 4. Each group consisted of 10 mice. Most of the nontreated mice died within 3 days of infection. Fluconazole was ineffective in this model, and even 1 mg of amphotericin B/kg did not prolong the survival time of mice significantly. Although TAK-457 at a dose of 10 mg/kg could not rescue the mice, it prolonged the survival time significantly (P < 0.01). In a separate experiment, we examined the lung weights and the fungal burdens in lungs in this infection model 2 days after infection, when all mice were alive (five mice per group). Although TAK-457 and amphotericin B suppressed significantly (P < 0.01) the increase of lung weight at doses of 10 and 1 mg/kg, respectively (Table 2), these treatments did not decrease the number of CFU in the lungs. In the infection model with CBA/J mice, no obvious inflammation occurred in the lungs (data not shown), but a large amount of β-d-glucan was detected in blood (Fig. 5). In this model, the level of β-d-glucan, a soluble cell wall component of Aspergillus, was determined 2 days after infection (eight mice per group) (Fig. 5). β-d-Glucan levels in plasma in mice treated with 10 mg of TAK-457 and 1 mg of amphotericin B/kg were significantly (P < 0.01 in both drugs) lower than those of vehicle-treated mice. In a separate experiment, we examined the lung chitin levels in order to estimate the degree of proliferation of A. fumigatus in lungs of CBA/J mice (five mice per group). The lung chitin levels in vehicle-treated mice 2 days after infection ranged from 24.2 to 65.7 U (Fig. 6). TAK-457 at a dose of 10 mg/kg and amphotericin B at a dose of 1 mg/kg reduced the lung chitin levels significantly (P < 0.01 in both drugs). The treatment with fluconazole did not affect the β-d-glucan and chitin levels at all.

FIG. 4.

Effects of TAK-457 and reference antifungal agents against pulmonary infection caused by A. fumigatus 437 in neutropenic CBA/J mice. Antifungal agents were administered intravenously once a day for 5 days starting 2 h after infection (n = 10). Agents: vehicle (•), TAK-457 at 10 mg/kg (○), amphotericin B at 1 mg/kg (□), fluconazole at 10 mg/kg (▵).

TABLE 2.

Effects of TAK-457 and reference antifungal agents on lung weight and fungal burden in lungs of neutropenic CBA/J mice inoculated intranasally with A. fumigatus 437^a

Drug	Dose (mg/kg)	No. of surviving mice/total no. of mice	Mean ± SEMb
			Lung wt (mg)	Fungal burden in lungs (log CFU/g)
Vehicle	0	8/8	264 ± 16.8	4.50 ± 0.07
TAK-457	10	8/8	196 ± 6.57*	4.37 ± 0.03
Amphotericin B	1	8/8	202 ± 7.75*	4.47 ± 0.08

^aAntifungal agents were administered intravenously once a day on days 0 and 1 of infection, and the weight and the number of A. fumigatus cells were examined for surviving mice on day 2.

^b*, P < 0.01 versus the vehicle-treated group.

FIG. 5.

Effects of TAK-457 and reference antifungal agents on plasma β-d-glucan levels of neutropenic CBA/J mice inoculated intranasally with A. fumigatus 437. Antifungal agents were administered intravenously once a day on days 0 and 1 of infection, and the β-d-glucan levels were determined on day 2. Data are expressed as means ± the standard error of the mean (SEM; n = 8). **, P < 0.01 versus the vehicle-treated group as determined by Dunnett test. AMPH-B, amphotericin B; FLCZ, fluconazole.

FIG. 6.

Effects of TAK-457 and reference antifungal agents on lung chitin levels of neutropenic CBA/J mice inoculated intranasally with A. fumigatus 437. Antifungal agents were administered intravenously once a day on days 0 and 1 of infection, and the chitin levels in lungs were determined on day 2 (n = 5). Circles indicate the chitin levels of individual mice, and the bars indicate means. **, P < 0.01 versus the vehicle-treated group. AMPH-B, amphotericin B; FLCZ, fluconazole.

DISCUSSION

Since patients with severe fungal infection are usually not able to take an oral medicine, we concentrated our efforts on developing an injectable form of an antifungal agent for the treatment of severe infections, such as those caused by A. fumigatus (13). In this study, we compared the efficacy of an injectable triazole, TAK-457, with that of the standard injectable antifungal drug amphotericin B by using two murine models of pulmonary aspergillosis.

Against the pulmonary aspergillosis in CDF₁ mice, both TAK-457 and amphotericin B were effective at doses of 10 and 1 mg/kg, respectively. In this model, the number of organisms present at the time therapy was initiated was ca. 10⁵ CFU/lungs (data not shown), and TAK-457 clearly reduced the fungal burden in lungs. It has been reported that the fungicidal activity of triazole agents against A. fumigatus is weaker than that of amphotericin B (21). In the present infection model, inflammatory cells appear to be involved in the fungicidal effect of TAK-457 because it is well known that polymorphonuclear leukocytes attack the hyphae and kill A. fumigatus by damaging their membranes (26).

In the infection model with CBA/J mice, on the other hand, TAK-457 at a dose of 10 mg/kg but not amphotericin B at a dose of 1 mg/kg prolonged the survival of mice significantly. A dose of amphotericin B higher than 1 mg/kg could not be used because of its toxicity. It has been reported that fungi can extend their hyphae freely in the absence of inflammatory cells (2, 6, 17). Since no apparent inflammation could be observed in lungs in this infection model, we believe that antifungal agents must suppress the fungal proliferation strongly in order to rescue the mice. Although neither TAK-457 nor amphotericin B reduced the numbers of CFU of infecting fungi in the lungs in this model, these agents clearly suppressed the chitin levels in lungs and β-d-glucan levels in plasma. These results suggest that both drugs suppressed the extension of hyphae and were effective against the infection; it has been reported that chitin levels but not CFU reflect hyphal growth (2, 17, 29) and that the elevation in the β-d-glucan level in plasma paralleled the development and extent of Aspergillus infection in the animal model (19).

TAK-457 showed activity similar to that of amphotericin B against invasive pulmonary aspergillosis in mice. Since this new injectable triazole may prove useful for the treatment of severe aspergillosis in humans, further clinical studies to assess the potential of TAK-457 therapy as an alternative to amphotericin B therapy for invasive aspergillosis are warranted.