Structure-Guided Synthesis of FK506 and FK520 Analogs with Increased Selectivity Exhibit In Vivo Therapeutic Efficacy against Cryptococcus

Abstract

Calcineurin is an essential virulence factor that is conserved across human fungal pathogens, including Cryptococcus neoformans, Aspergillus fumigatus, and Candida albicans. Although an excellent target for antifungal drug development, the serine-threonine phosphatase activity of calcineurin is conserved in mammals, and inhibition of this activity results in immunosuppression. FK506 (tacrolimus) is a naturally produced macrocyclic compound that inhibits calcineurin by binding to the immunophilin FKBP12. Previously, our fungal calcineurin-FK506-FKBP12 structure-based approaches identified a nonconserved region of FKBP12 that can be exploited for fungus-specific targeting. These studies led to the design of an FK506 analog, APX879, modified at the C-22 position, which was less immunosuppressive yet maintained antifungal activity. We now report high-resolution protein crystal structures of fungal FKBP12 and a human truncated calcineurin-FKBP12 bound to a natural FK506 analog, FK520 (ascomycin). Based on information from these structures and the success of APX879, we synthesized and screened a novel panel of C-22-modified compounds derived from both FK506 and FK520. One compound, JH-FK-05, demonstrates broad-spectrum antifungal activity in vitro and is nonimmunosuppressive in vivo. In murine models of pulmonary and disseminated C. neoformans infection, JH-FK-05 treatment significantly reduced fungal burden and extended animal survival alone and in combination with fluconazole. Furthermore, molecular dynamic simulations performed with JH-FK-05 binding to fungal and human FKBP12 identified additional residues outside the C-22 and C-21 positions that could be modified to generate novel FK506 analogs with improved antifungal activity. IMPORTANCE Due to rising rates of antifungal drug resistance and a limited armamentarium of antifungal treatments, there is a paramount need for novel antifungal drugs to treat systemic fungal infections. Calcineurin has been established as an essential and conserved virulence factor in several fungi, making it an attractive antifungal target. However, due to the immunosuppressive action of calcineurin inhibitors, they have not been successfully utilized clinically for antifungal treatment in humans. Recent availability of crystal structures of fungal calcineurin-bound inhibitor complexes has enabled the structure-guided design of FK506 analogs and led to a breakthrough in the development of a compound with increased fungal specificity. The development of a calcineurin inhibitor with reduced immunosuppressive activity and maintained therapeutic antifungal activity would add a significant tool to the treatment options for these invasive fungal infections with exceedingly high rates of mortality.

Keywords: X-ray crystallography; antifungal; antifungal resistance; antifungal susceptibility testing; ascomycin; calcineurin; tacrolimus.

FIG 1

Crystal structure of Aspergillus fumigatus FKBP12 bound to FK520. (A) Chemical structures of FK506 (tacrolimus) and FK520 (ascomycin). Structural differences at the C-21 residue are highlighted in red. (B) Protein X-ray crystal structure of A. fumigatus FKBP12 (purple) bound to FK520 (gray) characterized at a resolution of 1.7 Å. (C) Higher-resolution focus of FK520 bound to the FKBP12 ligand-binding pocket. Both C-22 and C-21 of FK520 approach key fungus-specific residue F88 in the 80s loop of FKBP12.

FIG 2

Chemical synthesis of novel FK506 and FK520 analogs. (A) One-step synthesis for C-22-modifed analogs from FK506 or FK520 as the starting material. (B) Panel of six FK506/FK520 analogs with acylhydrazone groups at C-22. Acylhydrazones differ in the R² group. Compounds JH-FK-01, JH-FK-02, JH-FK-03, and JH-FK-04 are derivatives of FK506. Compounds JH-FK-05 and JH-FK-07 are derivatives of FK520.

FIG 3

JH-FK-05 shows reduced immunosuppressive activity in vitro and in vivo and demonstrates fungal specificity. (A) In vitro immunosuppressive activity of calcineurin inhibitors was measured in a primary murine T-cell model. Dose response curves were generated from IL-2 expression in cells exposed to increasing concentrations of inhibitors. All analogs were reduced for immunosuppression compared to controls FK506 (black) and FK520 (pink). Error bars indicate standard error of the mean (SEM). (B) Therapeutic index scores for FK506/FK520 analogs were calculated by comparing the ratio of antifungal (MIC for C. neoformans) to immunosuppressive activity (IL-2 IC₅₀) to the same ratio for FK506. FK506 therapeutic score was set to 1.0, and increasing scores indicate increasing fungal specificity. (C) In vivo immunosuppression of FK506 and JH-FK-05 was measured in a murine model of a T cell-dependent response. Animals were treated daily via i.p. injection with vehicle, FK506 (5 mg/kg), or JH-FK-05 (60 mg/kg) beginning on day −1. NP-OVA was administered subcutaneously on day 0 to promote T cell-dependent proliferation of T follicular helper cells (Tfh) and germinal center B (GC B) cells. The proportions of proliferated Tfh cells and GC B cells were measured in animal lymph nodes on day 7. Tfh and GC B cells were measured from TCRβ⁺ CD4⁺ CD44^hi cells and TCRβ⁻ CD19⁺ cells, respectively. (D) Representative flow cytometry plots are presented from individual animals in each treatment group. Gating indicates a region containing Tfh cells (left, oval) or GC B cells (right, square). Error bars indicate standard deviation. For all plots, **, P < 0.01; *, P < 0.05.

FIG 4

JH-FK-05 treatment reduces organ fungal burden in multiple models of C. neoformans infection. (A) Experimental timelines for intranasal and intravenous models of C. neoformans infection. For both models, treatment groups included vehicle (gray), JH-FK-05 60 mg/kg (blue), fluconazole 12 mg/kg (red), and combination (purple). Treatment began on day 0 via i.p. injection and continued daily until the indicated endpoint. Each treatment group contained 5 female A/J mice. (B) Lung tissue fungal burdens from animals in the intranasal model. (C) Brain tissue fungal burdens from animals in the intravenous model. For all plots, ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05. Error bars indicate standard deviation.

FIG 5

JH-FK-05 treatment extends animal survival in murine model of cryptococcosis. (A) Survival plot of animals infected with C. neoformans H99 through intranasal instillation. Treatment groups included vehicle (black), JH-FK-05 60 mg/kg (blue), fluconazole 12 mg/kg (red), and combination (purple). Animals received treatment via i.p. injection daily for 14 days, and animals were monitored for health following treatment termination. Each treatment group consisted of 10 A/J female mice. (B) Survival curve pairwise comparisons by log-rank (Mantel-Cox) test.

FIG 6

Ligand atoms and protein residue contacts observed in MD simulations. Analysis of the significance of the observed contacts of human FKBP12 (hFKBP12) and C. neoformans FKBP12 (cnFKBP12) with JH-FK-05. (A) Stick representation of JH-FK-05 labeled for atoms making more significant contacts to hFKBP12 (magenta circles) and C. neoformans (cyan circles). (B) Residues of hFKBP12 (magenta) and C. neoformans (cyan) making more significant interactions with the ligand are shown in stick format on the ribbon protein structures. (C) Z-score graphs. Black circles represent Z-scores of C. neoformans FKBP12 contacts minus hFKBP12 contacts. More significant contacts for C. neoformans (ΔZ-score > 1) are labeled in cyan, while contacts more significant for the human protein (ΔZ-score < 1) are labeled in magenta.