Extrapolating Antifungal Animal Data to Humans – Is it reliable?

Abstract

Purpose of Review

This article aimed to review animal models of antifungals and identifies human literature to assess if the extrapolation of results is reliable.

Recent Findings

Animal studies have helped identify AUC/MIC targets for new drugs and formulations such as isavuconazole and delayed release posaconazole that have translated to successful outcomes in humans. Models have also been influential in the identification of possible combination therapies for the treatment of aspergillosis, such as voriconazole and echinocandins. However, challenges are endured with animal models when it comes to replicating the pharmacokinetics of humans which has been exemplified with the newest itraconazole formulation. Additionally, animal models have displayed a survival benefit with the use of iron chelators and amphotericin for mucormycosis which was not demonstrated in humans.

Summary

Animal models have been a staple in the development and optimization of antifungal agents. They afford the ability to investigate uncommon diseases, such as invasive fungal infections, that would otherwise take years and many resources to complete. Although there are many benefits of animal models there are also shortcomings. This is why the reliability of extrapolating data from animal models to humans is often scrutinized.

Introduction

Invasive fungal infections are associated with high mortality rates, especially in immunocompromised patients. [1, 2] Although these infections are not frequent, their prevalence and mortality incidence has increased due to a greater number of at-risk patients on immunosuppressive therapy and chemotherapy. [2, 3] As a result, animal models have been utilized to recreate these infections and test antifungals in a more complex environment than in vitro studies. The goal of these animal models is to predict the drug’s action in humans and forecast clinical outcomes in infected patients. There are pros and cons to applying animal data to humans. [4]

Animal models are practical and have been paramount in determining that antifungals have concentration-dependent activity. [4] They also allow investigators to measure drug concentrations in specific target tissues. This is particularly relevant for central nervous system (CNS) infections where adequate brain tissue concentrations would be necessary. Additionally, animal models provide a host to assess how certain combinations of antifungals can work synergistically, or antagonistically when used together. This area has become important for treatment strategies for resistant fungal infections. [4]

Mice are the most commonly used model but other animals including rats, guinea pigs, rabbits, and zebrafish are also used. Rabbits can undergo repeat blood sampling, mice can be genetically modified, and zebrafish offer the advantage of being transparent preventing need for invasive imaging. [3] When designing an animal model study, the following must be taken into consideration: host species, fungal strain, inoculum size and timing, route of inoculation, and immunosuppression. [3, 5] Similar to humans, immunocompetent mice do not easily succumb to fungal infections. Neutropenia can be induced by immunosuppressive agents to allow for infection to occur. [3] Timing of antifungal administration relative to inoculation influences disease severity and influences the potential for antifungal therapy to control the infection. [5] With the exception of Candida species, life-threatening fungi often cause infection by inhalation and invasion of sinopulmonary tissues in humans. Dissemination to extrapulmonary sites is also possible. To more closely mimic human pathogenesis, animals are often inoculated by sinopulmonary route. [3] Due to the growing use of these animal models the aim of this review is to discuss recent literature of various antifungal therapies and assess the ability to extrapolate this animal data for human use.

Triazole Antifungals

A previous review has discussed pros and cons of extrapolating animal data of triazoles antifungals, especially fluconazole and voriconazole, to human use. [4] Briefly, in animal studies, triazoles exhibit concentration-dependent activity meaning that a higher dose or area under the concentration curve to minimum inhibitory concentration ratio (AUC/MIC) is associated with greater treatment benefits (Table 1). Correlation of animal models with antifungal pharmacodynamics (PD) in humans has been best validated with fluconazole. This relationship has not been as well studied with other antifungals, but the principle is generally accepted in practice (Table 2). Factors such as site of infection, healthy versus infected animal models, species and other variables all play a role the interpretation of animal studies in the context of human use. [4]

Table 1.

Summary of the discussed animal and human data and their comparison.

Drug	Animal Data	Human Data	Similarities or Differences1
Triazoles
Voriconazole	• Combination with echinocandin shows either no difference, additive effects, or synergy [6, 7] • Rapid elimination and autoinduction of metabolism in mice [8, 9]	• Benefit seen with echinocandin combination for aspergillus [11••]	Similar benefit with combination echinocandin
Posaconazole	• AUC/MIC > 100 effective for mucormycosis and aspergillus [16•]	• Targeting AUC/MIC >100 effective in single patient case with mucormycosis and aspergillus infection [18] • IV and delayed release posaconazole reach higher concentrations [18, 19]	Similar efficacy with AUC/MIC target > 100
Itraconazole	• Higher concentrations and improved outcomes with SUBA-itraconazole for guinea pigs and rats [30, 31•] • Mice GI tract not optimized for SUBA-itraconazole absorption [31•]	• Higher bioavailability and less patient variability compared to standard itraconazole [29]	Similar for species other than mice
Isavuconazole	• Efficacy demonstrated against aspergillosis and mucormycosis [35, 39] • High brain tissue concentrations [45] • Strong dose-response correlation [34]	• Efficacy demonstrated against aspergillosis and mucormycosis [36••, 40] •Low CSF levels but good treatment success [41–43] • No statistically significant dose-response correlation but confounders present [38]	Similar efficacy Difference in presence of dose-response correlation
Echinocandins	• Paradoxical effect at higher doses [48] • High dose extended interval regimen efficacious [57–59]	• No evidence of paradox or at least not shown to be clinically relevant [52, 54, 55] • High dose extended interval shown effective for fungal prophylaxis and esophageal candidiasis [66, 68]	Difference in the impact of paradoxical effect
Amphotericin B	• Iron chelators and L-AMB reduced mucormycosis burden and improved survival [73]	• Iron chelator and L-AMB combination had higher mortality rate but confounding variable present [75••]	Difference in efficacy with iron chelator combination

¹Similar indicates that findings between animal models and clinical studies are consistent. Difference indicates discordance between animal and human data.

Table 2.

Animal and human PK/PD targets

Antifungal	Animal model target attainment	Human model target attainment or clinically relevant data
Fluconazole [79, 80]	AUC/MIC > 25	Dose/MIC > 50
Voriconazole [9, 12•, 81–83]	Candida  Free AUC/MIC > 24 Aspergillus  Free AUC/MIC > 36	Prophylaxis and treatment  Trough 1–5.5 mg/L Severe infection  Trough 2–5.5 mg/L Toxicity  Trough > 5.5 mg/L
Posaconazole [12•, 18, 84]	Aspergillus  AUC/MIC > 300 Mucormycosis  AUC/MIC > 100	Prophylaxis  Trough > 0.7 mg/L Treatment  Trough 1–3.75 mg/L  Cavg > 1.25 mg/L associated with 75% response rate
Itraconazole [12•, 85]	Levels by bioassay Peak > 6 mg/L	Prophylaxis  Trough 0.5–4 mg/L (HPLC)  Trough 3–17 mg/L (bioassay) Treatment  Trough 1–4 mg/L (HPLC)  Trough 3–17 mg/L (bioassay) Toxicity  Trough > 17.1 mg/L (bioassay)  Corresponds to ∼ 4 mg/L (HPLC)
Isavuconazole [12•, 34, 86–88]	AUC/MIC > 50	TDM not routinely recommended Treatment  Trough 2–3 mg/L reasonable
Echinocandin [47, 89]	Non-Candida albicans  Free AUC/MIC > 7 Candida albicans  Free AUC/MIC > 20 Aspergillus  Cmax/MEC > 10	Further elucidation required
Amphotericin B deoxycholate [90–92]	Candida  Cmax/MIC > 10 Aspergillus  Cmax/MIC > 2.4 L-AMB ~ 4 to 6-fold higher PD targets	Further elucidation required

AUC area under the concentration curve; C_avg average concentration; C_max maximal concentration; HPLC high-performance liquid chromatograph; L-AMB liposomal amphotericin B; MEC minimum effective concentration; PD pharmacodynamics; TDM therapeutic drug monitoring

Voriconazole

Voriconazole is considered first line therapy for the treatment of invasive aspergillosis. However, despite voriconazole’s efficacy, mortality rates remain high. [6] As a result, the role of combination therapy for the treatment of invasive aspergillosis has been investigated. In vitro studies have shown that voriconazole and echinocandins have synergistic effects against aspergillus. [7] However, animal studies have had mixed outcomes with results of either no difference, additive, or synergistic effects. [7] One immunocompetent mouse model compared combination therapy with voriconazole and anidulafungin versus monotherapy for two different Aspergillus fumigatus isolates. There were wild type isolates (MIC 0.25 mg/L) and isolates with a cyp51 mutation inferring voriconazole resistance (MIC 4 mg/L). The investigators found that combination therapy improved survival and was synergistic in isolates that were susceptible to voriconazole, but the effect was only additive for isolates that were voriconazole resistant. [6] Of relevance, voriconazole undergoes rapid elimination in mice and multidose experiments demonstrate evidence of autoinduction, most pronounced in mice and rats. These pharmacokinetic (PK) differences make mice a less optimal animal model for voriconazole compared to other species such as guinea pigs. [8, 9] Strategies of higher doses and administration of agents known to inhibit voriconazole metabolism have been utilize in an attempt to overcome this set back with mice models. [9, 10] Taking this into consideration efficacy outcomes with voriconazole mice models can be difficult to apply to human use due to alteration of voriconazole PK.

A randomized clinical trial in humans compared combination therapy of voriconazole and anidulafungin to voriconazole monotherapy for aspergillosis in patients with hematological malignancies or hematopoietic stem cell transplantation (HSCT). Investigators found that in patients with radiographic findings and positive galactomannan, combination therapy was associated with a lower mortality rate than voriconazole monotherapy (15.7% vs 27.3%, P=0.037). [11••] Overall the randomized control trial seems to mirror the results of animal studies, which was improved outcomes with combination therapy. Based on this data it is reasonable to consider treating patients with documented invasive aspergillosis with combination therapy of voriconazole and an echinocandin, especially in regions where azole resistance rates are > 10%. [12•, 13]

Posaconazole

Posaconazole is one of only two triazole antifungals that has activity against both Aspergillus species and mucormycosis. [14] AUC/MIC has been shown to be a strong predictor of treatment outcomes of posaconazole use for these infections. [15] An animal model of neutropenic mice sought to define relationships between posaconazole exposure and treatment outcomes for mucormycosis and correlate them with targets of Aspergillus. The study assessed A. fumigatus (MIC 0.5 mg/L) and Rhizopus oryzae (MIC 2 mg/L) across a dosing range of 0.1 to 80 mg/kg/day. Higher posaconazole doses were required to achieve maximum reductions in lung fungal burden for R. oryzae compared to A. fumigatus (50 mg/kg/day vs 10 mg/kg/day, respectively). Investigators determined that an AUC/MIC ratio associated with stasis was 57 for R. oryzae and 27 for A. fumigatus. It was concluded that a provisional AUC/MIC target of > 100 (or free drug AUC/MIC > 1) should achieve sufficient exposure to treat A. fumigatus or R. oryzae infection. [16•] Posaconazole oral solution would only be able to obtain this target for isolates with MICs < 0.06 mg/L by EUCAST (European Committee on Antimicrobial Susceptibility Testing). However, it was predicted that the newer posaconazole formulations, delayed release tablet and intravenous (IV) solution, would reliably achieve an AUC/MIC > 100 in 95% of patients for MICs up to 0.125 mg/L. [16•]

As mentioned previously, the rarity of fungal infections, particularly mucormycosis, limits available human research. As a result, the use of posaconazole for mucormycosis has not been assessed in a randomized clinical trial, however there are numerous published case reports. In a case series of 96 patients, posaconazole was used as either monotherapy (n=21), generally after an initial treatment period with another antifungal, combination therapy (n=60), often with amphotericin B, or as salvage therapy (n=14). Authors found a complete response in 64.6% of patients and an overall mortality of 24%. They concluded that posaconazole was successful for the treatment of mucormycosis both as combination therapy and as second line salvage therapy. [17] However, these case reports occurred before the release of the newer posaconazole formulations. The delayed release tablet and IV solution are now the preferred dosage forms because they achieve higher AUCs, have less PK variability, and higher rates of goal trough attainment compared to the oral suspension formulation. [18–20] These formulations have not been widely studied in animal models. Instead, investigators of animal studies may give varying amounts of posaconazole oral suspension to mimic human exposure levels seen with the newer posaconazole formulations. [16•] Clinicians should be mindful of this detail when extrapolating data from animal models to human use with these newer posaconazole formulations. However, in a recent publication the new formulations of posaconazole have shown promise in treating patients with mucormycosis. Patients receiving posaconazole monotherapy and patients on amphotericin B plus posaconazole were matched with amphotericin B based treatment controls. The new posaconazole formulations were associated with a lower 42-day all-cause mortality rate compared to the amphotericin B control group. This suggests the utility of these new posaconazole formulations for the treatment of mucormycosis in humans. [21]

The only identified study assessing posaconazole AUC/MIC targets for mucormycosis in humans was in a single patient case. The patient was co-infected with A. fumigatus, Aspergillus flavus, Rhizopus microsporus, and Lichtheimia corymbifera (MICs 0.01, 0.12, 0.25, and 0.25 mg/L, respectively). The patient was treated with the delayed release posaconazole tablet. The patient received a loading dose of 300 mg every 12 hours for 2 doses, then 400 mg every 24 hours for days 2 through 14. After this point the dose was decreased to 300 mg once daily which was administered for 4 months. The study found the AUC of posaconazole maintenance doses was 3 to 4-fold higher than the AUC after the loading dose. As a consequence of this lower AUC, the loading dose was not able to achieve the AUC/MIC target specified in previous murine model studies (i.e. >100 for mucormycosis). However, the maintenance doses of 400 and 300 mg once daily were adequate to meet these targets. Therefore, it may be necessary to increase the loading dose if the patient has a known infection with elevated MICs. In either event, the patient was successfully treated with posaconazole delayed release using AUC/MIC targets that had been identified in the previously discussed murine model which used oral solution posaconazole. [18]

Cyclodextrin toxicities with IV Voriconazole and IV Posaconazole

Both IV formulations of voriconazole and posaconazole require the additive sulfobutylether-B-cyclodextrin (SBECD) for solubility. [22] SBECD is cleared renally in both animals and humans and accumulates in renal dysfunction. [23] In mice and rat models SBECD was given as a single dose up to 2000 mg/kg IV with no adverse effects. However, with repeated doses of ≥ 160 mg/kg/day given over 1 to 6 months renal cellular toxicity in the form of renal vacuolation and tubular necrosis was seen. This effect was reversed within 1 month after SBECD discontinuation with the exception of a few rats that still had renal tubular vacuolation. [23] As a result of these findings oral voriconazole and posaconazole are recommended instead of the IV formulations for patients with a CrCl < 50 mL/min. [24, 25] However, there is insufficient evidence to support that this adverse effect occurs in humans. A systematic review of IV voriconazole in patients with renal impairment identified 7 original articles that varied in methodologies and study definitions. Overall the studies found that IV voriconazole was not a risk factor for worsening renal function in patients with baseline renal impairment. Additionally, the rates of worsening renal dysfunction were equivalent to oral voriconazole and other antifungal agents. [22] Recently, a retrospective review of 101 patients with hematological disease on voriconazole IV for fungal prophylaxis was completed. Median duration of voriconazole was 27 days which is longer than other clinical studies. Multivariate analysis showed cumulative voriconazole IV doses ≥ 400 mg/kg were associated with worsening renal function in patients without baseline renal impairment. [26] However, it should be noted that there were confounding factors, including the administration of other nephrotoxic agents, that could have impacted these findings. Voriconazole IV has been assessed in patients undergoing continuous renal replacement therapy (CRRT). SBECD was effectively cleared at rate that was similar to the ultrafiltration rate. [27] Posaconazole IV has also been assessed in CRRT in a patient case study. Investigators found that posaconazole IV could be given at standard doses to patients receiving CRRT without significant risk for SBECD accumulation. [28] Data surrounding SBECD shows that animal studies may not have accurately predicted toxicities in humans. Through further studies would be needed this shows how there is potential to draw wrong conclusions from animal models and thereby prevent the use of IV therapy in patients that could benefit from it.

Itraconazole

A new formulation of itraconazole was FDA approved in 2018. Super-BioAvailable (SUBA)-itraconazole was designed to increase absorption of itraconazole by using a pH-dependent polymeric matrix. [29] Population PK analysis in humans has shown that SUBA-itraconazole capsules have a relative bioavailability that is 173% higher than the standard itraconazole capsules (Sporanox®). SUBA-itraconazole also provides more consistent exposure with 21% less variability between subjects. These results indicate that SUBA-itraconazole doses between 53 and 65 mg will produce similar exposure to 100 mg of Sporanox®. [29] SUBA-itraconazole was studied in rats and found to have greater drug absorption compared to itraconazole oral solution. [30] This led to SUBA-itraconazole being studied in various neutropenic mice and guinea pig models. [31•] In non-infected mice investigators were unable to find a correlation between SUBA-itraconazole dose and itraconazole serum and lung tissue concentrations. When SUBA-itraconazole was assessed in mice infected with A. fumigatus there was not a statistically different decrease in lung fungal burden and worse survival rates when compared to the positive control, posaconazole. This contrasts the findings in guinea pigs which demonstrated a dose-concentration relationship. In guinea pigs infected with A. fumigatus SUBA-itraconazole improved survival. Rationale for the wide difference in outcomes between mice and guinea pigs stems from differences in gastrointestinal (GI) tracts. Compared to humans and other animal models, mice generally have a higher stomach pH (pH 4) and a more acidic intestinal pH (pH 5). SUBA-itraconazole is designed to sustain the lower pH values of the stomach and dissolute when the pH rises in the intestine. Aside from pH differences SUBA-itraconazole also needs fluid to dissolve, and mice have low fluid volumes in the intestine thus further preventing drug absorption. [31•] From this comparative animal study it can be concluded that mice are not the optimal animal model for SUBA-itraconazole, and their results should be assessed with caution. Animals, such as guinea pigs and rats, that have GI tracts that closer resemble humans in terms of pH and fluid should be utilized instead.

Isavuconazole

One of the newest triazole antifungal agents available is isavuconazole. The prodrug, isavuconazonium sulfate, is broken down by plasma esterases to become the active component, isavuconazole. [32] It has a broad spectrum of activity that includes Aspergillus species and Mucorales. [2, 32] Isavuconazole has been studied in many animal models. Its role in invasive pulmonary aspergillosis was studied using a neutropenic murine model. [33] The study included 10 A. fumigatus isolates, 4 of which were wild type and the remaining 6 had a cyp51 mutation. This particular mutation is associated with decreased susceptibilities to triazoles. [33] Isolates were tested in accordance to the Clinical and Laboratory Standards Institute (CLSI) to identify MICs. For wild type isolates the MICs ranged from 0.25 to 1 mg/L and from 0.125 to 8 mg/L for cyp51 mutants. The investigators assessed isavuconazonium given for 7 days in doses ranging from 40 to 640 mg/kg every 12 hours (or about 20 to 307 mg/kg of isavuconazole). Stasis occurred at isavuconazonium doses 65 to 617 mg/kg/12 hours. The corresponding median isavuconazole AUC/MIC ratio to achieve stasis for total and free drug was 503 and 5, respectively. Similar to other antifungals the AUC/MIC ratio was a strong predictor of treatment outcomes. Stasis was seen for all isolates with MIC ≤ 1 mg/mL regardless of cyp51 mutation presence. Overall, a dose-response relationship was seen. Higher isavuconazole doses were associated with a larger microbiologic effect. In the case of elevated MICs a higher isavuconazole dose was needed to achieve a similar effect compared to isolates with lower MICs. [33]

A PD study of invasive A. fumigatus was performed in immunocompetent mice. [34] Four different strains consisting of a wild type and 3 mutants where included. The isolates had MICs by EUCAST that ranged from 0.5 to 8 mg/L. In mice, the half-life of isavuconazole was dose dependent and ranged from 1 to 3 hours. The mice model indicated that AUC/MIC was the primary driver of treatment efficacy. A PD target for 50% survival was determined to be total drug AUC/MIC_CLSI ratio of 50.5 (or AUC/MIC_EUCAST ratio of 24.7). The authors concluded that normal human maintenance dosing of isavuconazole 200 mg daily would cover A. fumigatus strains with MICs ≤ 0.5 mg/L and likely remain effective for strains with MICs up to 2 and possibly 4 mg/L. [34]

Another study investigating isavuconazole for A. fumigatus was conducted in a neutropenic rabbit model. [35] MIC_CLSI for isavuconazole was determined to be 1 mg/L. Infected rabbits were given a loading dose of isavuconazole and then received one of three maintenance dosing options (isavuconazole 20, 40, or 60 mg/kg once daily). These doses yielded AUCs of 60 ± 14, 141 ± 32, and 197 ± 27 mg·h/L, respectively. Higher doses of 40 and 60 mg/kg showed significant decreases in fungal burden and improved survival. Isavuconazole 20 mg/kg was less active compared to the other regimens. The authors concluded that there is a concentration-dose relationship and that the findings may suggest that administration of higher doses of isavuconazole for some isolates with elevated MICs could be advantageous. [35]

Overall in vivo models indicate that isavuconazole’s antifungal activity is strongly correlated with AUC/MIC, with higher ratios inferring greater efficacy. The elimination half-life seen in the prior studies was short (1–3 hours). [34] In general, this is a known phenomenon seen with smaller animals. [4]

Isavuconazole for the treatment of Aspergillus species has been studied in the SECURE trial, a phase 3 clinical trial comparing isavuconazole with voriconazole. [36••] The findings of this randomized trial indicated that isavuconazole was non-inferior to voriconazole for the treatment of invasive aspergillosis. All cause 42-day morality was 19% vs 20%, respectively. [36••]

A population PK study was preformed using data from 9 phase 1 trials and the phase 3 SECURE trial. [37] As expected a longer half-life was seen when studying isavuconazole in humans (130 hours vs 1–3 hours in mice). [34, 37] Clinical dosing of isavuconazole (200 mg every 8 hours for 6 doses followed by 200 mg daily) was found to yield an average AUC of 97 ± 48 mg·h/L. [37] Comparing this to the AUCs in the aforementioned rabbit study, 97 mg·h/L lies in between the AUCs seen with 20 and 40 mg/kg doses. Interesting in the animal model, a dose of 40 mg/kg was associated with considerably better survival and microbiological resolution compared to 20 mg/kg. [35] With this information it could be conceived that an AUC of 97 mg·h/L is not adequate to produce notable improvement in human outcomes. However, the SECURE trial shows at doses expected to yield this AUC, isavuconazole was non-inferior to voriconazole. [36••] This may be a sign of differences between animal and human studies, but there are other factors that could explain this as well such as differences in MICs, or the possibility that a dose between 20–40 mg/kg would have produced outcomes more akin to that seen with 40 mg/kg. An alternate point of view is that perhaps higher isavuconazole doses could have demonstrated superiority over voriconazole in the SECURE trial.

Authors of the population PK study also performed a probability of target attainment (PTA) analysis. It was concluded that with isavuconazole 200 mg daily, >90% of simulated patients will achieve adequate exposures to treat infections with MICs ≤ 1 mg/L by EUCAST or MICs ≤ 0.5 mg/L by CLSI. [37] This was similar to the conclusions made by Seyedmousavi and colleagues when they extrapolated their in vivo findings to predict target attainment at clincal doses of isavuconazole. [34]

As previously mentioned, animal studies have shown a strong correlation with AUC/MIC ratios and isavuconazole efficacy. A post-hoc analysis of the SECURE trial was conducted to assess for an exposure-response relationship in humans. In contrast to in vivo data, this study was not able to demonstrate a statistically significant association between measures of drug exposure (i.e. steady state AUC, trough concentrations) and outcomes such as 42-day all-cause mortality. When the authors assessed AUC/MIC ratios there was only a small number of patients who had both PK and MIC data available (n=36). Possible explanations for the lack of a statistically significant relationship include that drug exposure variability between patients was overall low to moderate thus preventing a wide concentration range for curve assessment. Additionally, it was presumed that most Aspergillus isolates were wild type, but if more mutant isolates with elevated MICs were included it might have better defined an exposure-response curve. Taken together, a lack of variability in isavuconazole exposure and fungal MICs differs from animal studies where these factors are often diversified. This may explain why AUC/MIC ratio and efficacy relationships were found in animals but not humans. For now, there is no indication to monitor isavuconazole concentrations, but ongoing information from the post-approval database may enable an exposure-response curve to be better defined in humans. [38]

Isavuconazole has been compared with liposomal amphotericin B (L-AMB) in both humans and in vivo models for the treatment of invasive fungal infections caused by mucormycosis. [39, 40] In an animal model, neutropenic mice were infected with Rhizopus delemar (MIC_CLSI of 0.25 mg/L) and then treated with either isavuconazole (100 mg/kg 3 times daily), L-AMB (15 mg/kg once daily), or placebo. Survival rates at 21-days were 65%, 40%, and 15%, respectively. It was concluded that isavuconazole was as effective as L-AMB for mucormycosis. The VITAL trial was a clinical study of isavuconazole for mucormycosis when compared to matched controls who received amphotericin B. Similar to the mouse model, the VITAL investigators found that isavuconazole had similar efficacy to amphotericin B for treatment of Mucorales molds. All-cause 42-day morality was 33% with isavuconazole compared to 39% seen with amphotericin B. [40]

To date 4 patient case reports of isavuconazole therapy for the treatment of CNS related fungal infections exist. [41–43] Two patients had invasive mucormycosis infection, the remaining 2 patients had an unspecified fungal meningitis. All 4 patients had tried and failed at least 2 other antifungal therapies before moving to isavuconazole salvage therapy. All patients were successfully treated with isavuconazole and remained disease free for the duration of study follow-up. [41–43] Isavuconazole dosing was increased above the maintenance dose (200 mg daily) for 2 of the 4 patients. One patient had early disease progression, so the dose was increased, the other patient had a measured trough level that was lower than desired. Both patients were on a higher isavuconazole dose for over a month before decreasing back to 200 mg daily to finish out therapy. [41, 43] Two patients had CSF levels collected during the treatment period. For both patients isavuconazole CSF concentrations were considerably low; about 100 times lower than serum levels. [41] Regardless, treatment was still effective for both patients. Animal studies may be able to explain this occurrence.

Isavuconazole has been shown to be widely distributed in tissue, including the eye, brain, liver and others. [44] In an animal model isavuconazole was given to noninfected rats as a single dose and blood and brain tissue samples were collected. It was found that isavuconazole brain tissue concentrations were almost double that of serum levels. Maximum concentration in the brain was 6.95 compared to 3.67 mg/L in serum. [45] In another animal model of mice infected with cryptococcus investigators found that the ratio of isavuconazole brain tissue to plasma concentrations was about 1.35, which was consistent across several doses. [46] These animal models suggest that CSF levels of isavuconazole may not adequately reflect the drug’s ability to reach high brain tissue concentrations, which helps explain the patient cases that have had treatment success with isavuconazole. [41–45]

Echinocandins

Echinocandins are a class of antifungals which includes caspofungin, anidulafungin, and micafungin. They exhibit concentration dependent activity (Table 1). [47, 48] In vitro studies of these agents have consistently showed a paradoxical effect whereby the fungal burden increases at doses above a certain threshold. [1, 49] This effect is fungal species and drug specific. [1] It has been seen in Candida and Aspergillus isolates. Caspofungin prompts the greatest number of strains to exhibit this paradoxical effect. The occurrence is less commonly seen with anidulafungin and micafungin. The mechanism is not one of resistance development but rather an adaptation of the fungal cell wall as the result of external stress from the antifungals. [1, 49] This paradoxical effect is seen in animal studies, though not as consistently as with in vitro models. [1, 48–50] One mouse model observed an increase in fungal burden only at the highest studied dose of caspofungin (5 mg/kg). Investigators also saw a massive recruitment of neutrophils and pro-inflammatory response, which was in contrast to lower doses that observed a decrease in fungal burden and inflammatory pathology. [48] Other animal studies of neutropenic mice and rabbits for invasive aspergillus have shown that with high dose caspofungin there was an increase in pulmonary fungal burden. However, these studies have not observed differences in mortality or survival rates with these doses compared to lower doses. [47, 51]

Several clinical studies have assessed effects of different doses of echinocandins. Two randomized clinical trials exist to date. One trial compared standard dose caspofungin (70 mg loading dose followed by 50 mg daily) with high dose caspofungin (150 mg daily) in patients with proven invasive candidiasis. They found that patients in the high dose group had similar rates of favorable responses as those in the standard dose group. They concluded that both regimens were safe and effective. [52] The other randomized trial compared standard dose micafungin (100 mg daily), high dose micafungin (150 mg daily), and caspofungin (70 mg loading dose followed by 50 mg daily) for invasive candidiasis. They found similar rates of treatment success among all three regimens, 76.4%, 71.4%, and 72.3%, respectively. However, there was lower than expected success rates among patients with non-candidemia invasive candida infections who were treated with high dose micafungin. [53] A retrospective review of high dose caspofungin therapy (100 mg daily) in immunosuppressed patients found that it has favorable responses comparable to patients given standard dose caspofungin despite more patients in the high dose group having more frequent poor prognostic factors. [54, 55] Collectively, these studies do not suggest the presence of a paradoxical effect with echinocandin treatment in humans, or at least not one of clinical significance. The clinical interpretation of the paradoxical effect in animal studies is often complicated by experimental design and short duration of antifungal therapy compared to clinical practice. Additionally, it has been demonstrated that echinocandins help augment the clearance of fungal infections by enhancing host immune system to recognize the infection. This might off-set any decrease in efficacy that could arise from this paradoxical effect. [55] Regardless, human studies do not seem to show improved efficacy with higher echinocandin doses so the clinical utility of increasing the dose may not be a high yield decision.

Since echinocandins are only available as intravenous (IV) formulations there is growing interest in the efficacy and safety of extended dosing regimens. [56] The concept has been tested in animal models. [57–59] In a neutropenic murine model of pulmonary aspergillosis, large micafungin doses given once improved survival and suppressed fungal burden for up to 7 days after infection. [57] Another model of neutropenic mice infected with Candida glabrata found that a single dose of micafungin ≥ 50 mg/kg resulted in maximal fungal burden decline without regrowth at 7 days. [58] A more recent study assessed neutropenic rabbits with disseminated Candida albicans. They were treated over 12 days with varying micafungin dosing regimens, 1 mg/kg every 24 hours, 2 mg/kg every 48 hours, and 3 mg/kg every 72 hours. The researchers found that all treatment groups demonstrated significant and comparable clearance of fungal infection from target organs. [59] Lepak and colleagues studied micafungin in neutropenic mice for both prophylaxis and treatment of disseminated candidiasis with C. albicans. Dosing regimens were humanized to mimic micafungin PK for doses ranging from 100 mg to 1000 mg. They found that humanized doses of 400 and 600 mg given once prevented organism recovery for up to 6 days. Humanized regimens of ≥ 300 mg given every 6 days achieved net stasis of Candida infection. This data demonstrates the possibility of extended dosing frequency with echinocandins, specifically micafungin. To date there are no extended dosing models of anidulafungin or caspofungin. [56]

Clinical PK dose escalation and safety studies have assessed different echinocandins over a wide range of doses and have found no dose-limiting toxicities. [52, 60–67] In 20 patients with HSCT, anidulafungin was given for fungal prophylaxis at doses of 200 mg every 48 hours or 300 mg every 72 hours. They found similar drug exposures with both regimens. Median AUC_0–144 was 348 and 359 mg·h/L, respectively, which were comparable to standard dosing anidulafungin 100 mg daily (AUC_0–144 397 mg·h/L). [67] Another human study assessing micafungin in 20 patients with hematological malignancies found similar AUCs for micafungin dosed as 100 mg daily and 300 mg twice weekly. Monte Carlo simulations were preformed and projected that micafungin 700 mg once weekly would produce similar drug exposure as the previously mentioned regimens. [65] A observational study over a 5-year period was conducted to assess intermittent high dose micafungin as prophylaxis in acute leukemia and HSCT patients. To be included patients had to receive at least 5 doses of micafungin ≥ 300 mg twice or 3 times weekly. Of the patients receiving micafungin for prophylaxis 5 out of 83 patients (6%) developed breakthrough invasive fungal infection. This is consistent with breakthrough rates of other antifungal clinical trials ranging from 5.3–7.3%. Breakthrough infections were largely caused by Aspergillus species. No Candida species were associated with a breakthrough infection. [66] Two randomized control trials assessing extended interval dosing of micafungin for the treatment of esophageal candidiasis have been performed. Their combined analysis has been reported by Andes and colleagues. The studies compared standard dose micafungin (150 mg once daily) with high dose micafungin (300 mg every other day). Successful mycological response at the end of therapy was seen in 78.8% of patients in the standard dose group and 87.1% in patients in the high dose group. This difference approached statistical significance (P=0.056). [68]

Rezafungin is a new echinocandin currently in phase 3 studies. [1, 56] It has shown in preclinical animal studies to have a longer half-life than other echinocandins, ranging from 28 to 80 hours depending on animal species. [69–71] Human data in phase 1 and 2 studies has demonstrated a long half-life of 130 hours, which enables for once weekly dosing. [72] Animal data has shown favorable outcomes with just one dose of rezafungin. [70, 71] Efficacy studies in humans are pending but it will be interesting to see how this agent with a much longer half-life compares for the treatment and prophylaxis of invasive fungal infections.

In general, caution should be warranted when extrapolating results of extended interval, high dose studies from animals to humans (Table 2). Elimination half-lives are significantly shorter in smaller animals, which naturally complicates the interpretation. [4] However as discussed previously, there are a few clinical studies that suggests positive results with the use of extended interval echinocandins. This supports to a certain degree the reliability of the animal model findings.

Amphotericin B

In animal models amphotericin B exhibits concentration dependent fungicidal activity against a broad-spectrum of fungi. Maximum concentration over MIC (C_max/MIC) ratio appears to have the highest predictive value for reductions in fungal burden (Table 1). Amphotericin B is considered first line therapy for the treatment of mucormycosis infections. [4]

Mucormycosis is associated with a high mortality rate so aggressive management with surgical resection and liposomal amphotericin B (L-AMB) are warranted. Iron overload as well as diabetic ketoacidosis (DKA) are both independent risk factors for mucormycosis infection. [73] Animal studies have looked at iron chelator deferiprone and deferasirox in the treatment of mucormycosis. Deferiprone was compared to L-AMB in DKA mouse models. Deferiprone was found to be as effective as L-AMB. Both regimens were more effective than placebo in non-iron overload animals. Administration of free iron with deferiprone reversed protection, confirming that the mechanism of efficacy was iron chelation. [74] Another animal study in mice with DKA or neutropenia was used to study mucormycosis. The mice were treated with either L-AMB monotherapy, deferasirox monotherapy, or combination L-AMB and deferasirox. Investigators found that deferasirox significantly improved survival and decreased tissue fungal burden and efficacy was similar to that of L-AMB. Perhaps most interestingly is that deferasirox and L-AMB worked synergistically to improve survival and reduced tissue fungal burden. [73]

These promising results of animal studies were the rationale for the DEFEAT Mucor trial. Twenty patients with proven or probable mucormycosis were randomized to either L-AMB plus deferasirox or L-AMB plus placebo. Overall the deferasirox group had more patients with active malignancy, neutropenia, corticosteroid therapy, and pulmonary mucormycosis infection (instead of rhino-orbital infection) which are all known factors that increase mortality. The results showed higher mortality rates in patients receiving combination deferasirox therapy at 90 days compared to placebo (82% versus 22%, P=0.01). The confounding baseline characteristics of the two study groups and the limited sample size make it difficult to generalize the results of this trial. Regardless the trial does not support the adjunctive use of deferasirox therapy for mucormycosis. [75••] Here is an example of how animal studies may not accurately predict outcomes in humans (Table 2). It is unclear what caused such a vast difference in findings but being mindful that animals frequently do not have other drug-drug interactions on board and have fewer co-morbidities which can make the translation muddied.

Discussion

Compared to human studies, strengths of animal models include their lower cost and ability to manipulate and assess more aspects of a study, such as fungal strain, drug doses, tissue concentrations, and more. Their use has been important for investigating therapy options for these overall rare invasive infections, especially ones caused by resistant fungal strains. [5] However, there are also many limitations of animal models, most apparent are differences in the absorption, distribution, and elimination of drugs. [5, 76] The PK of antifungals, particularly triazoles, is altered in mice compared to humans. Smaller animals have higher elimination rates meaning that investigators may have to dose more frequently or co-administer drugs that inhibit p-glycoprotein or cytochrome P450 to limit clearance of the drug of interest. [4, 5, 76] This is one of the reasons why animal to human dose extrapolation is not a simple conversion based on body weight. [77] Other examples of PK differences can be illustrated with voriconazole and SUBA-itraconazole. As previously discussed, mice are considered poor models for both drugs but for different reasons. Voriconazole is rapidly metabolized and undergoes autoinduction of metabolism in mice. [8, 9] SUBA-itraconazole has limited absorption in mice secondary to their lower intestinal fluid volumes and pH. [31•] For both drugs guinea pigs were shown to better correlate with human PK. This exemplifies the need for checks and balances among studies of different species. Assessing efficacy and safety in only one animal species may not be sufficient. Additionally, for all antifungals, there is a need for concomitant human PK studies to interpret and confirm similar drug concentrations found in animal models. [5]

Differences in host immune response could influence antifungal outcomes in human and animal models. High doses of echinocandins have been shown to cause a paradoxical effect in animal studies. [48] This effect has not been identified in humans or at least does not impact clinical outcomes. [52, 54, 55] Echinocandins can augment the clearance of fungal infections by enhancing the ability for the human immune system to recognize infection. Difference in human and animal host immune response and enhancement could alter antifungal efficacy in light of this paradoxical effect. [55] Regardless, human studies do not seem to show improved efficacy with higher echinocandin doses so the clinical utility of increasing the dose cannot be recommended at this juncture. However, echinocandin high dose extended interval regimens have shown promise in both animal and human studies. [57–59, 66, 68] Still, we caution clinicians from interpreting animal results of high dose extended interval studies because antifungal PK in animals is vastly different making it difficult to apply to humans.

Safety findings between animal and human data can differ. This has been evident with SBECD where animal studies indicate renal cellular toxicities, but this effect has not been fully identified in humans. Unfortunately, these safety differences can lead to effective therapies being passed over without proof of equivalent toxicity seen in humans. Nonetheless, there is an ethical dilemma with studying potentially toxic drugs in humans. We therefore advise practitioners to be mindful of animal toxicities but recognize that they many not reliably detail human side effects so when it comes to antifungal selection benefits and risk should always be weighted.

Lastly, humans with invasive fungal infections inevitably have other co-morbidities. In contrast, the animals used for studies are otherwise healthy with the exception of possible induced immune suppression depending on the model. Co-morbidities can significantly impact response to antifungal therapy and treatment success. Readers can consider this to be similar to a clinical study in which the treatment groups have extensively different baseline characteristics. Many clinicians will evaluate the results of such a study with skepticism. Also, with co-morbidities comes the need for other drug therapies which carry the potential for drug-drug interactions with antifungals, particularly triazoles. [4] Drug interactions are not common with animal studies, unless it is intentionally done as a PK booster.

Taking a step back, when testing the efficacy of antifungals in animal models, it is not the effects of the antifungal on the animal that investigators truly want to know but rather the effects of the antifungal on the fungi when placed in an animal host. Animal models can be considered overall helpful for predicting antifungal effects in human disease, because the organism that is affected is not the human or animal but actually the fungi. However, differences arise due to how hosts process the drug and the inability for animal models to replicate all covariables present in human disease. These limitations should be noted when extrapolating data from animals to humans. [78]

Conclusion

Given the limited clinical experience and treatment with fungal infections, animal studies provide important guidance concerning which treatment regimens and dosing strategies will have the highest likelihood of clinical success. The PK of antifungals, especial triazoles, are very difference in mice, compared with that in humans. So, it is necessary to confirm results in human models. This can also make interpretation of extended frequency high dose regimens difficult. Overall, animal studies are a helpful bridge from in vitro models to humans. The correlation is overall reliable, but limitations of each model need to be taken into consideration when extrapolating animal data to humans.