Fosmanogepix (APX001) Is Effective in the Treatment of Immunocompromised Mice Infected with Invasive Pulmonary Scedosporiosis or Disseminated Fusariosis

There are limited treatment options for immunosuppressed patients with lethal invasive fungal infections due to Fusarium and Scedosporium. Manogepix (MGX; APX001A) is a novel antifungal that targets the conserved Gwt1 enzyme required for localization of glycosylphosphatidylinositol-anchored mannoproteins in fungi. We evaluated the in vitro activity of MGX and the efficacy of the prodrug fosmanogepix (APX001) in immunosuppressed murine models of hematogenously disseminated fusariosis and pulmonary scedosporiosis.

assay, which assessed log 10 conidial equivalents/gram tissue (CE), as well as histological improvement.

RESULTS
Antifungal susceptibility. The activities of MGX and comparators were evaluated against a panel of 6 Scedosporium, 3 Lomentospora prolificans, and 10 Fusarium clinical isolates, including strains that were utilized in the efficacy models (Table 1). Susceptibility was evaluated using the CLSI M38-A2 broth microdilution method for filamentous fungi (22). The MGX MEC endpoint criteria were as described for the echinocandins. MIC values were determined for the comparators VORI, POSA, and AMB. The range of MGX concentrations evaluated in the antimicrobial susceptibility studies was 0.015 to 8.0 g/ ml. The MEC value of MGX was 0.03 g/ml for all individual Scedosporium and Lomentospora isolates tested, including S. apiospermum DI16-478, used in the efficacy model. The MEC value of MGX ranged from Յ0.015 g/ml to 0.03 g/ml for the 10 F. solani isolates (Table 1). An additional 19 F. oxysporum strains were also evaluated, and all isolates demonstrated MGX MEC values of Յ0.015 g/ml (data not shown). MIC 90 values of comparator drugs for these strains were the following: 2 g/ml AMB, 8 g/ml VORI, and 8 g/ml POSA (data not shown). These data are consistent with ranges of MGX MEC values reported previously for Fusarium spp. (0.015 to 0.5 g/ml) and Scedosporium spp. (0.03 to 0.25 g/ml) (14). The POSA and AMB MIC values against S. apiospermum DI16-478 were 0.5 g/ml and 1.0 g/ml, respectively, whereas the VORI and AMB MIC values against F. solani 95-2478 were 2 g/ml (Table 1).
Fosmanogepix demonstrates efficacy in a highly immunocompromised pulmonary scedosporiosis model. To assess the effect of fosmanogepix in the treatment of pulmonary scedosporiosis, ICR mice were immunosuppressed with cyclophosphamide (200 mg/kg) and cortisone acetate (500 mg/kg) on days Ϫ2 and ϩ3 relative to infection. This treatment regimen results in ϳ10 days of leukopenia with a total white blood cell count dropping from ϳ13,0000/cm 3 to almost no detectable leukocytes, as determined by the Unopette system (BD, NJ) (23). Mice were intratracheally infected with 3 ϫ 10 7 spores of S. apiospermum DI16-478 on day 0. This strain is susceptible to azoles as well as to micafungin and previously has been validated in a neutropenic mouse model of scedosporiosis (24). Treatment with placebo (diluent control), fosmanogepix (26,   continued daily for 7 or 11 days for fosmanogepix or POSA and 4 days for L-AMB. To extend the half-life of MGX, 4 of the mouse cohorts were also administered 50 mg/kg of the cytochrome P450 inhibitor 1-aminobenzotriazole (ABT) 2 h prior to fosmanogepix administration, as previously described (19) and as labeled in Fig. 1 and Table 2.
(i) Survival. Survival of mice was assessed over 21 days (n ϭ 10 to 20 mice/cohort). The median survival time for vehicle control cohorts with and without ABT were equivalent (P ϭ 0.566): 7 and 8 days, respectively (Table 2). Neither POSA nor L-AMB prolonged median survival time versus the placebo (P Ͼ 0.5). In contrast, dosing regimens of 52, 78, and 104 mg/kg fosmanogepix plus ABT for 7 days significantly demonstrated prolonged median survival times of 17, 13, and 11 days, respectively ( Table 2 and Fig. 1). Of note is that the surviving mice looked healthy. Further, these fosmanogepix treatments enhanced overall survival by day 21, when the experiment was terminated (30% to 50% for 52 to 104 mg/kg versus 10% for placebo or POSA). Survival after treatment with fosmanogepix for 7 days or 11 days was not different (data not shown). Dosing regimens yielding lower exposures (26 mg/kg with ABT, 52 mg/kg BID without ABT, and 104 mg/kg BID without ABT) did not result in statistically significant prolonged median survival times (P ϭ 0.7, P ϭ 0.16, and P ϭ 0.36, respectively) ( Table 2 and Fig. 1), consistent with a dose response.  (ii) Tissue burden. The effect of drug treatment on tissue fungal burdens was examined using qPCR to evaluate log 10 conidial equivalents (CE)/gram of tissue. Mice (n ϭ 10) were infected and treated as in the survival studies and then sacrificed on day ϩ4 (8 h after the last treatment), and lungs (primary target organ), kidneys, and brains were processed and evaluated for CE/gram of tissue (25). In this experiment, we evaluated several doses above the highest dose used in the survival study (104 mg/kg plus ABT) in order to determine whether additional reductions in CE would be observed at higher doses. These higher doses (156, 208, and 264 mg/kg, all plus ABT) can only be used in studies in which mice are sacrificed at an early time point (e.g., 4 days) but not in longer-term survival studies due to mouse-specific toxicity observed in highly immunocompromised mice (18).
At all doses (104 to 264 mg/kg), fosmanogepix plus ABT significantly reduced tissue fungal burden in lung (P Ͻ 0.04) and brain (P Ͻ 0.007) compared to that in placebotreated mice and were comparable to that of L-AMB treatment (P Ͼ 0.18) (Fig. 2). In kidney, all doses of 104 mg/kg to 264 mg/kg plus ABT significantly reduced fungal burden compared to placebo control (P Ͻ 0.007), but only the highest dose (264 mg/kg plus ABT) demonstrated statistical significance versus L-AMB treatment (P ϭ 0.02) (Fig.  2). Thus, a dose-response curve generally was not observed at these higher dosing levels. All fosmanogepix treatments resulted in an ϳ2-log reduction in lung, brain, and kidney CE.
(iii) Histological observations. Brain tissues taken from placebo mice processed at the same time of the tissue fungal burden experiment (4 cohorts of 104, 156, 208, and 264 mg/kg plus ABT) and collected at day ϩ4 postinfection (n ϭ 3 mice/group) showed that S. apiospermum spores had disseminated from the lungs to the brain with fulminant growth and inflammation (Fig. 3) that coincided with neurological symptoms, including torticollis, spinning, and barrel rolling. In contrast, no signs of infection or inflammation could be detected in any of the brains collected from mice treated with fosmanogepix (104 to 264 mg/kg plus ABT) or L-AMB (10 mg/kg) (Fig. 3). These results confirm the similar efficacy of fosmanogepix and L-AMB in this scedosporiosis murine model. Surprisingly, we could not detect any signs of infection in the lung (the primary infection target), indicating the rapid dissemination of the infection to the brain and/or lack of fungal filamentation in this organ (data not shown).
Fosmanogepix demonstrates efficacy in an immunocompromised disseminated fusariosis model. We investigated the efficacy of fosmanogepix in the treat- ment of disseminated fusariosis, as previously described (26). Immunosuppression of ICR mice was performed as described above for the scedosporiosis model. Mice were infected with 8.1 ϫ 10 2 spores of F. solani 95-2478 (MEC, 0.03 g/ml) by tail vein injection on day 0. This strain was from a patient blood sample and was previously used in neutropenic mouse models of disseminated fusariosis (26,27). Treatment with placebo (diluent control), fosmanogepix (78 or 104 mg/kg, p.o.) plus 50 mg/kg ABT, L-AMB (15 mg/kg, i.v.), or VORI (40 mg/kg, p.o.) began 16 h postinfection and continued for 8 days for fosmanogepix or VORI and 4 days for L-AMB. To enhance the half-life of the molecules, 50 mg/kg ABT was administered 2 h prior to fosmanogepix, and grapefruit juice (50%) was added in the drinking water for VORI-treated mice (28). Mice were sacrificed on day ϩ4 and organs processed to determine CE by qPCR.
(i) Survival. Treatment of mice (n ϭ 10/group) with fosmanogepix plus ABT or L-AMB significantly (P Ͻ 0.01) enhanced median survival time versus placebo (12 and 10 days for 78 and 104 mg/kg of fosmanogepix, 10 days for L-AMB treatment) (Table 3). Furthermore, fosmanogepix plus ABT and L-AMB treatments equally enhanced overall survival by day 21 when the experiment was terminated (40% for L-AMB or fosmanogepix at 78 mg/kg and 20% for fosmanogepix at 104 mg/kg) (Fig. 4). VORI was not effective in this model in terms of median survival time or overall survival by day 21 (Table 3 and Fig. 4).
(ii) Tissue burden. Administration of L-AMB (15 mg/kg) and fosmanogepix (78 mg/ kg, 104 mg/kg, and 130 mg/kg) plus ABT resulted in reductions in tissue burden CE that were statistically different from values for the placebo control (kidney, P Ͻ 0.0003; brain, P Ͻ 0.0154) (Fig. 5) (n ϭ 10/group). Compared to placebo, treatment with 78, 104, or 130 mg/kg fosmanogepix plus ABT reduced kidney counts by 2.10, 2.21, and 3.14 log 10 CE, respectively, while L-AMB treatment resulted in a 3.96-log 10 reduction in  .04 log 10 CE, respectively, versus 3.76 log 10 for L-AMB. L-AMB demonstrated activity that was statistically equivalent to both 104 mg/kg and 130 mg/kg fosmanogepix in reduction of brain CE (P Ն 0.07). VORI did not demonstrate statistically significant reductions in CE versus placebo control in brain (0.48 log 10 CE) or kidney (0.68 log 10 CE), and fosmanogepix outperformed this azole in treating hematogenously disseminated murine fusariosis (Fig. 5).
(iii) Histological observations. Kidney tissues that were harvested from a parallel tissue burden experiment that utilized dosing regimens of 78, 104, and 130 mg/kg fosmanogepix (plus ABT) and 15 mg/kg L-AMB were also processed for histology (n ϭ 3 mice/group). Consistent with the tissue fungal results, kidneys from placebo-or VORItreated mice showed diffused and fulminant filamentation of F. solani (Fig. 6). In Grapefruit juice (50%) was added into the drinking water of the VORI treatment group from day Ϫ2 to day ϩ4 to enhance the drug half-life; 50 mg/kg ABT was administered daily 2 h prior to fosmanogepix dosing. *, P Ͻ 0.05; **, P Ͻ 0.002. P values were assessed versus the placebo control using the log rank test.

FIG 5
Reduction in kidney and brain tissue burden in a disseminated Fusarium model. Mouse kidney and brain tissue burdens in mice (n ϭ 10) were measured at 4 days postinfection. Male ICR mice were infected and treated as described for Fig. 4 with a dose of 50 mg/kg ABT administered orally 2 h prior to daily fosmanogepix dosing. Mice were sacrificed 8 h after the last dose, and kidney and brain tissues were harvested and processed for fungal burden by qPCR. Fungal burden data (presented as medians Ϯ interquartile ranges) were log 10  contrast, less diffusion of fungal hyphae was evident in kidneys treated with fosmanogepix at 78 mg/kg, while fungal elements were confined to a small abscess in kidneys treated with fosmanogepix at 104 mg/kg. Finally, kidneys harvested from mice treated with fosmanogepix at 130 mg/kg or L-AMB at 15 mg/kg had similar and consistently less infection than other groups (Fig. 6).

DISCUSSION
There are limited treatment options for the management of hyalohyphomycosis, especially in the case of disseminated disease in highly immunocompromised patients. Fosmanogepix is an i.v. and orally available antifungal prodrug that is currently in clinical development for the treatment of life-threatening invasive fungal infections, including candidiasis and aspergillosis. The broad-spectrum nature of this first-in-class agent previously has been demonstrated in vitro as well as in vivo (15)(16)(17)(18).
A previous study evaluated the efficacy of fosmanogepix in the treatment of disseminated fusariosis and pulmonary scedosporiosis using less immunocompromised models where 200 mg/kg of 5-fluorouracil was administered 5 or 6 days prior to infection (29). In the two survival experiments, the nadir of the neutrophil counts occurred approximately on the day of infection (30), with neutrophils recovering early during the course of the infection. The current regimen results in pancytopenia for at least 9 days from the first administered dose (23).
In this study, successful treatment outcomes utilized ABT plus oral dosing of 78 mg/kg or 104 mg/kg fosmanogepix, where total area under the plasma drug concentration-time curve (tAUC) values were ϳ200 and ϳ280 g·h/ml (K. J. Shaw, unpublished), exposures that were shown in pharmacokinetic/pharmacodynamic (PK/ PD) studies to be associated with A. fumigatus stasis and 1-log kill, respectively (31). To achieve Ն90% survival in the earlier study that utilized 5-fluorouracil with three-timesper-day (TID), intraperitoneal (i.p.) administration dosing regimens (without ABT) resulted in total tAUC values that ranged from ϳ12 g·h/ml (fusariosis) to ϳ46 g·h/ml (scedosporiosis), considerably lower than the exposures needed for efficacy in the current study, where animals were severely immunocompromised with a cyclophosphamide-cortisone acetate regimen. Consistent with the increased severity of the current model, fosmanogepix dosed at 104 mg/kg BID without ABT (tAUC of ϳ50 g·h/ ml; Shaw, unpublished) did not result in statistically significant prolonged median survival times in the scedosporiosis model. In terms of clinical relevance, the exposures reached in the current study (ϳ200 and ϳ280 g·h/ml) are consistent with exposures achieved in fosmanogepix phase 1 singleand multiple-ascending-dose studies (9,21) and anticipated in future fosmanogepix clinical trials. Efficacy, as measured by median survival time versus placebo control, was observed in both the S. apiospermum pulmonary infection model and the F. solani disseminated model using both of these dosing regimens (P Յ 0.05).
Although clinical guidelines for Scedosporium include VORI as the preferred agent for first-line therapy (7), the data from the pulmonary scedosporiosis model showed that there was no benefit in overall survival or in median survival time versus the placebo control (P ϭ 0.08) of mice treated with a high dose of POSA (30 mg/kg BID), which resulted in exposures in mice that are approximately six times the clinically relevant dose (32). Of note is that this high dose of POSA was efficacious in a pulmonary aspergillosis model (18), highlighting the difficulty of successful treatment of pulmonary scedosporiosis despite the POSA MIC value against both molds being 0.5 g/ml ( Table 1) (18). Similarly, a high dose of 10 mg/kg L-AMB, which is typically used in invasive mold infection models as a comparator, was ineffective in extending survival in this model versus the placebo control (P ϭ 0.65).
In the disseminated fusariosis model, the control antifungal agents utilized were also consistent with first-line standard of care for patients. In a previous study, Graybill et al. utilized grapefruit juice to enhance the half-life of VORI, resulting in the observation of prolonged survival and reduced numbers of CFU in a disseminated F. solani 95-2478 model when 10 mg/kg or 20 mg/kg VORI was administered (28). However, in this study, 40 mg/kg VORI plus grapefruit juice was not efficacious (P ϭ 0.42). Although the same strain of F. solani and the same strain of mice (ICR) were used in the two studies, there are notable differences in the two experiments. The Graybill et al. study utilized a 3-log higher inoculum (8 ϫ 10 5 CFU/mouse versus 8.1 ϫ 10 2 /mouse); however, the mice were immunocompromised with 150 mg/kg 5-fluorouracil on day Ϫ1 rather than utilizing the cyclophosphamide-cortisone acetate regimen utilized here. Consistent with the increased level of immunosuppression of the current model, the higher dose (40 mg/kg) of VORI was not effective in the current model. In contrast, in this study, mice administered a high dose of 15 mg/kg L-AMB demonstrated survival that was significantly different from that of the placebo control (P ϭ 0.001) and was equivalent to fosmanogepix dosed at 78 and 104 mg/kg plus ABT (P Ͼ 0.5).
Histology and CFU number reduction were also evaluated in the S. apiospermum and F. solani models using dosing regimens that were of concentrations equal to or higher than those used in the survival study in order to determine whether greater organism clearance would be observed at higher doses. Compared to placebo control mice in the pulmonary scedosporiosis model, all doses of 104 mg/kg to 264 mg/kg fosmanogepix plus ABT resulted in a significant reduction in brain, lung, and kidney burdens. At these higher doses, no clear dose response was observed, since there was no significant difference between the fosmanogepix treatment regimens when evaluated pairwise, suggesting that maximal activity was achieved at the dose of 104 mg/kg fosmanogepix plus ABT or was outside the dynamic range of the assay.
Consistent with the survival results using the disseminated fusariosis model, compared to the placebo control group, significant reductions in brain and kidney fungal burden were observed for fosmanogepix (78 mg/kg, 104 mg/kg, and 130 mg/kg) plus ABT and 15 mg/kg L-AMB but not for VORI. In brain, again no statistical significance was observed between fosmanogepix dosing groups; however, kidney burden was statistically lower for the highest fosmanogepix dosing group (130 mg/kg) and equivalent to that for L-AMB. The activity of fosmanogepix was also confirmed by the lack of detection of any fungal elements in the brains of mice with scedosporiosis and with the reduced fungal elements seen in kidneys of mice infected with F. solani.
In this study, we showed that the active moiety MGX was highly active against S. apiospermum, S. boydii, L. prolificans, and F. solani in vitro. Using highly immunocompromised mouse models of pulmonary scedosporiosis and disseminated fusariosis, we demonstrated that the efficacy of fosmanogepix was equivalent to or better than the current standard of care of antifungal agents. Findings from both oral and i.v. fosmanogepix phase 1 clinical studies have shown favorable PK, allowing once-daily dosing with high bioavailability (ϳ90%) and no food effect (9,21). Given the previous demonstration of efficacy across a broad range of yeasts and mold mouse models, including Candida (15,16,19) and Aspergillus (18,31), activity against azole-and echinocandinresistant isolates of Aspergillus species (Cyp51 and Fks1, respectively) (31), as well as the data described here for Fusarium and Scedosporium, further investigations into the development of this first-in-class agent for invasive fungal infections is highly warranted, especially against difficult-to-treat infections.

Microorganisms.
Nine clinical isolates of Scedosporium/Lomentospora and 9 clinical isolates of Fusarium solani were obtained from the Fungus Testing Laboratory at the University of Texas Health Sciences Center at San Antonio (UTHSCSA) ( Table 1). Fusarium solani 95-2478 is a blood isolate provided by P. Ferrieri (University of Minnesota) (26). Fungal strains used in this study were routinely grown on Sabouraud dextrose agar plates for 5 to 10 days until confluent at 37°C. Conidia were collected by flooding the plates with sterile phosphate-buffered saline containing 0.01% (vol/vol) Tween 80. The conidia were concentrated by centrifugation and washed in the same buffer, diluted, and counted using a hemocytometer.
Antifungal agents. For in vitro studies, the active moiety MGX (Amplyx Pharmaceuticals) was used along with AMB (Fisher Scientific, Hampton, NH), POSA (Sigma-Aldrich Corp., St. Louis, MO, USA), and VORI (Sigma-Aldrich Corp., St. Louis, MO, USA). For efficacy studies, the water-soluble N-phosphonooxymethyl prodrug fosmanogepix (Amplyx Pharmaceuticals) was used. The final prodrug solution was in 5% dextrose and dosed orally (p.o.) per gram of mouse body weight. A 5-mg/ml solution of ABT (Fisher Scientific, Hampton, NH) in water was administered orally 2 h prior to infection as 10 l per gram of mouse body weight, resulting in a dose of 50 mg/kg, and this served as the placebo plus ABT group. For in vivo efficacy studies, POSA (Merck & Co., Inc., Rahway, NJ) was purchased as an oral suspension (200 mg/5 ml) and kept at room temperature. L-AMB (Gilead Science, Foster City, CA) and VORI (50-mg tablet from Ajanta Pharma USA, Bridgewater, NJ) were obtained from local pharmacies. The final L-AMB solution was dissolved in 5% dextrose and administered intravenously, while crushed VORI tablets were suspended in irrigation water and administered to mice by oral gavage.
In vitro testing. The in vitro susceptibility of manogepix (drug concentrations of 0.015 to 8.0 g/ml) against agents of scedosporiosis and fusariosis was evaluated using the Clinical Laboratory and Standards Institute (CLSI) M38-A2 method using minimum effective concentration (MEC) endpoints for the echinocandins (22). Three clinical isolates each of Scedosporium apiospermum, S. boydii, and Lomentospora prolificans were evaluated along with 10 strains of F. solani. MIC values were determined for POSA, VORI, and AMB (22). All in vitro testing was conducted with drug powders with known potencies.
Efficacy models. ICR mice (Envigo) were immunosuppressed with cyclophosphamide (200 mg/kg) and cortisone acetate (500 mg/kg) on days Ϫ2 and ϩ3 relative to infection (n ϭ 10/group). To prevent bacterial infection, enrofloxacin (50 g/ml) was added to the drinking water from day Ϫ3 to day 0. Ceftazidine (5 g/dose/0.2 ml) replaced Baytril treatment on day 0 and was administered daily by subcutaneous injection from day 0 until day ϩ8. All drug treatments were initiated 16 h postinfection and continued for 8 consecutive days, given by oral gavage for POSA, VORI, or fosmanogepix, but L-AMB was given intravenously through day ϩ4. To extend the half-life of MGX, mice were administered 50 mg/kg the cytochrome P450 inhibitor ABT 2 h prior to fosmanogepix administration for most of the cohorts. The prodrug fosmanogepix was dosed at 26 mg/kg (with ABT), 52 mg/kg (with and without ABT), 78 mg/kg (with ABT), 104 mg/kg (with and without ABT), 156 mg/kg (with ABT), 208 mg/kg (with ABT), and 264 mg/kg (with ABT) in the various experiments. Using a conversion factor of 1.3 to account for the methyl phosphate group in the prodrug, the doses were equivalent to MGX at 20 mg/kg, 40 mg/kg, 60 mg/kg, 80 mg/kg, 120 mg/kg, 160 mg/kg, and 203 mg/kg, respectively.
(ii) Disseminated F. solani. Mice were infected with a targeted inoculum of 8.1 ϫ 10 2 cells of F. solani by tail vein injection. For survival studies, treatment with placebo (diluent control), fosmanogepix (78 or 104 mg/kg, p.o.) plus ABT, L-AMB (15 mg/kg, i.v.), or VORI (40 mg/kg, p.o.) began 16 h postinfection and continued for 8 days for fosmanogepix or VORI and 4 days for L-AMB (n ϭ 10/group). To enhance the half-life of VORI, grapefruit juice (Ocean Spray) was added to the drinking water to a final concentration of 50% and was given to VORI-treated mice. To assess tissue fungal burden, mice were infected and treated as described above, except that mice were sacrificed on day ϩ4 (4 days of treatment), 8 h following the last treatment (n ϭ 10/group). Organs were processed for CE by qPCR using 28S primers (F, TAAATGGACCAGGGCGCAAA; R, AGAGGGAACGAGATGGGTT).
For both models, tissues harvested from mice to assess fungal burden were also processed for histopathological examination. Briefly, tissues were fixed in 10% zinc-buffered formalin, paraffin embedded, sectioned, and stained with Grocott's methenamine silver (GMS) stain for microscopic examination (n ϭ 3/group).
Statistical analysis. The nonparametric log rank test was used to determine differences in survival times. Differences in lung, kidney, and brain CFU numbers were compared by the nonparametric Wilcoxon rank sum test. A P value of Ͻ0.05 was considered significant.
All animal-related study procedures were compliant with the Animal Welfare Act, the Guide for the Care and Use of Laboratory Animals (34), and the Office of Laboratory Animal Welfare and were conducted under an IACUC-approved protocol by The Lundquist Institute at Harbor-UCLA Medical Center.