A humanized antibody against mucormycosis targets angioinvasion and augments the host immune response

Abstract

Mucormycosis is a fungal infection caused by Mucorales fungi that cause severe disease and fatality, especially in immunocompromised individuals. Although vaccines and immunotherapeutics have been successful in combating viral and bacterial infections, approved anti-fungal immunotherapies are yet to be realized. To address this gap, monoclonal antibodies targeting invasive fungal infections have emerged as a promising approach, particularly for immunocompromised patients that are unlikely to maximally benefit from vaccines. The Mucorales spore coat proteins (CotH) have been identified as crucial fungal invasins that bind to glucose-regulated protein 78 (GRP78) and integrins of host barrier cells. Previously, we described a murine monoclonal antibody, anti-CotH C2, which protected diabetic ketoacidosis (DKA) and neutropenic mice from mucormycosis. Here, we advanced the development the C2 IgG1 by humanizing it, establishing a stable Chinese hamster ovary (CHO) cell line producing the antibody at commercial yields, and carried out optimization of the upstream and downstream manufacturing processes. The resultant humanized IgG1 (VX-01) exhibited a ten-fold increase in binding affinity to CotH proteins and conferred comparable in vitro and in vivo efficacy when compared to C2 antibody. The mechanism of protection was reliant on prevention of angioinvasion and enhancing opsonophagocytic killing. VX-01 demonstrated acceptable safety profiles with no detectable damage to host cells in vitro and weak or moderate binding to only cytoplasmic proteins in ex vivo Good Laboratory Practices (GLP)-human tissue cross reactivity studies. Our studies warrant continued development of VX-01 as a promising adjunctive immunotherapy.

One Sentence Summary:

An anti-CotH humanized IgG1 antibody protects against lethal mucormycosis when given as adjunctive therapy to immunocompromised mice.

INTRODUCTION

Mucormycosis, caused by filamentous fungi belonging to the Mucorales order, was first identified in 1885(1). These pathogenic fungi are ubiquitous, found in various environmental sources like soil and vegetation; they are also a major cause for food spoilage, thereby making exposure to these pathogens common. Several risk factors contribute to the susceptibility to mucormycosis, including leukemia, immunosuppression in organ transplant recipients, blood dyscrasias (e.g. myelodysplastic syndrome), and diabetes mellitus(2). The incidence of diabetes-related cases has been on the rise in recent decades, fueled by global obesity rates, with an estimated 382 million people affected in 2013 and projected to reach 592 million by 2035(3). Additionally, malnourishment, acidosis, and prematurity are also considered as potential risk factors for mucormycosis(4). Notably, mucormycosis has become the second most common invasive mold infection in major U.S. transplant centers, trailing only aspergillosis(1). The infection is highly invasive, characterized by a mortality rate exceeding 40% despite the use of antifungal therapy and aggressive surgical intervention(2). The mortality rate escalates to nearly 100% in cases of disseminated diseases and prolonged neutropenia(2, 5). Consequently, the urgent need for innovative treatment modalities to combat mucormycosis is apparent.

Monoclonal antibodies (mAbs) have emerged as a crucial immunotherapy against infectious diseases, successfully targeting viral and bacterial pathogens(6). Because mucormycosis usually afflicts patients with severely compromised immunity, these patients are more likely to benefit from mAb treatment instead of a vaccine-based immunotherapy(7). Some of the best antigens to target for mAb therapy are those expressed on the fungal cell surface or secreted and are known to contribute to the pathogenesis of the disease(8). In this context, CotH3, a fungal cell surface protein, has been identified as the key ligand that binds to glucose-regulated GRP78 on human cells (e.g. nasal epithelial cells and umbilical vein endothelial cells [HUVECs]), playing a vital role in fungal-mediated cell invasion hematogenous dissemination(9, 10). We have shown that murine mAbs raised against CotH3 protein can protect mice from mucormycosis(11).

We humanized one of the most protective murine mAbs (named VX-01), expressed it in CHO cells, and optimized and completed manufacturing processes. VX-01 enhanced the ability of antifungal drugs to protect mice from mucormycosis through prevention of invasion and enhancing opsonophagocytosis. Finally, in a GLP-human tissue cross-reactivity study, VX-01 was shown to be safe. Thus, our study introduces a therapeutic approach against mucormycosis with the potential to treat patients with this severe infection.

RESULTS

Comparative analysis of binding affinity between grafted and chimeric mAbs targeting CotH3

To enhance the potential therapeutic application of the anti-CotH3 C2 antibody in treating mucormycosis, we pursued a humanization strategy to maintain or improve its binding affinity to its fungal CotH3 antigen. Initially, we obtained the sequences of the heavy chain (HC) and light chain (LC) variable regions of the C2 antibody through hybridoma cell line mRNA sequencing. N-terminal protein sequencing of the murine mAb was also conducted, revealing identical sequence of the first 10 amino acids of both the kappa light chain variable region (V_κ) and the heavy chain variable region (V_H), to the encoding regions of the cloned mRNA sequence of the hybridoma cells(12).

To achieve humanization, we performed alignments of the murine HC and LC variable regions with human germline antibody variable regions to identify the most suitable frameworks for the mouse C2 mAb V_H and V_κ. Based on the alignment analysis, we selected the human VH4-59*01 framework for the HC humanization and the IGKV2-29*02 (A18 V_κ) framework for the LC humanization. Initially, we generated a chimeric mAb containing the mouse V_H and V_κ fused to the human constant regions (fig. S1). We also generated a grafted antibody in which the six complementary determining regions (CDRs) of the C2 mouse V_H and V_κ were grafted in the selected human framework to generate a humanized clone without mutations in the framework. This step was intended to minimize the antigenicity of the antibody when given to humans.

The DNA encoding the V_H variable region was cloned into a pcDNA3.4-IgG1 vector, and sequencing confirmed the HC sequences of the mouse-human chimera and grafted variants. Similarly, the V_κ region for both constructs was cloned into a pcDNA3.4-kappa vector, with sequencing confirming the LC sequences.

The expression of the light and heavy chains was achieved in ExpiCHO cells, and both the chimera and grafted antibodies were purified to >95% purity based on SDS-PAGE (fig. S2). We then performed biolayer interferometry (13) to assess the binding kinetics of each mAb for the CotH3 antigen (table S1). Whereas the chimeric mAb exhibited an approximately ten-fold improvement in affinity over the original murine mAb [e.g. a dissociation constant (K_D) of 5.7 nM binding affinity for the chimera versus 59 nM for the C2 parental mAb], the grafted humanized variant showed no detectable binding to CotH3 antigen (Fig. 1A).

Fig 1. A chimeric mAb, but not grafted ones, recognized R. delemar CotH3.

(A) Binding kinetics of the mouse C2, chimera, and grafted antibodies targeting CotH3 antigen. The on rates were measured between 0 and 300 seconds, and the off rates were measured after washing the probe at 300 seconds. Dissociation constants (K_D) are shown. The variation of the response is shown as the experimental (red) and fitting curves (blue, green, or cyan); the dotted red line divided the association and dissociation steps. (B) Immunoblot analysis depicting the binding of mouse C2 antibody, grafted antibody, and the chimeric mAb to denatured CotH3 antigen. The bands indicate the specific binding of each antibody to the CotH3 antigen. (C) Immunoblot analysis evaluating binding of the chimeric mAb to CotH3 antigen or a human lung epithelial cell line total lysate. (D) Flow cytometry analysis showing binding of mouse C2 antibody, the chimeric mAb, and the grafted antibody to R. delemar.

To further verify the binding affinity of the antibodies, we conducted an immunoblot using recombinant CotH3 protein (expressed in Saccharomyces cerevisiae(11)) which showed intense recognition of the CotH3 antigen at 65 kDa by both the mouse C2 mAb and the chimeric mAb, but less recognition by the grafted clone (Fig. 1B). The binding of the chimeric mAb to CotH3 antigen is specific since the mAb only recognized the CotH3 protein band from total lysate of S. cerevisiae expressing the Flag-tagged CotH3. The generated chimeric mAb did not recognize any proteins extracted from a human alveolar epithelial cell line (A549 cells) (Fig. 1C). Next, we compared the ability of the chimera and grafted mAbs with the C2 parent mouse clone in recognizing CotH antigens on R. delemar. Consistent with the binding affinity and immunoblot data, the chimeric mAb showed strong binding to R. delemar spores that exceeded the binding activity of the parent C2 mouse clone, whereas the grafted clone showed less binding to R. delemar spores (Fig. 1D). Collectively, these studies show that the chimeric mAb retains recognition of the CotH3 on R. delemar, whereas the grafted mAb lost most of its ability to recognize the antigen. Because our goal is to optimize the therapeutic benefit of these antibodies for treating mucormycosis by minimizing the mouse sequences in the generated mAb, we decided to introduce key mouse framework residues back to the grafted variant to restore its binding activity.

Back mutations to LC and HC restored the binding affinity of the humanized mAb to recombinant CotH3 protein

As the first step towards restoring binding activity of the grafted and humanized mAb, we conducted a computer modeling step that utilized the human antibody frameworks and mouse antibody CDRs. For this purpose, Protein Data Bank (PDB) human structures 1XGP (heavy chain) and 2W60 (light chain) were chosen as homology templates for the murine variable regions, and 4LKX was used for the human framework with grafted murine CDRs. An in-depth structural assessment of the homology models was then conducted, where visual analysis and residue sequence and structure comparisons between the mouse and humanized models guided us in identifying regions where back-mutations might be warranted(14). The model of the light chain V_κ region and heavy chain V_H variable region is depicted in Fig. 2A.

Fig. 2. Back mutations enhanced binding of humanized mAb to CotH3 protein.

(A) Model representation of the LC variable region (teal blue framework) and HC variable region (green framework) of the 4LKX PDB structure. The grafted CDRs from the parental C2 mAb are shown. In the left-hand panel, the CDRLs are highlighted in magenta and the CDRHs are shown in brown. The right-hand panel highlights potential regions of conflict (yellow). (B) Summary of the three sets of framework residues (upper set: LC; lower set: HC) that could impact the binding pocket formed by the three CDRs. These residues are potential candidates for back-mutation from human to murine residues. (C) Steady-state analysis was performed by BLI whereby a 1:2 serial dilution of CotH3 from 480nM to 7.5nM (Specifically 480, 240, 120, 60, 30, 15, and 6.5nM) was compared with buffer alone; Shown are dissociation constant (K_D as an assessment for binding affinity) values for the humanized variants (VX-01, VX-02, and VX-03) compared with the chimeric mAb.

In the case of the light chain, we identified three groups of residues within the human A18 framework that could potentially impact the folding of the CDRs. The first group consisted of residues S53, R54, and F55, which were located at the beginning of framework 3 of the human chain, adjacent to CDR2. These residues were predicted to force a different conformation of light chain complementarity determining region 2 (CDRL2) compared with the lysing, leucine, and aspartic acid (KLD) residues at these positions in the murine V_K. Consequently, the first LC variant, named LC1, was designed with the S53K (changing S to K), R54L, and F55D back-mutations at these positions. The second construct (LC2) included additional mutations on residues Y34 and Y36, situated at the beginning of framework 2 of the human acceptor framework, representing contact residues that could impact CDR orientation. Thus, the LC2 had mutations of LC1 (S53K, R54L, and F55D back mutations) and additional backmutations Y34N and Y36L. Finally, the third construct LC3 included an additional mutation at residue L46R in the human framework, this residue was predicted to potentially play a role in binding or stabilizing CDR conformations. Thus, LC3 had the following six backmutations S53K, R54L, F55D, Y34N, Y36L, and L46R backmutations (Fig. 2B).

We also identified a total of four key residues that could impact the conformation of the CDR heavy chain and the antigen-binding pocket. Thus, we generated three HC constructs. In HC1, we introduced a backmutation of V71R to maintain the conformation of CDRH2. In the HC2 construct we identified W47 and N58 human amino acids which are likely located within the Vernier zone of CDRH2 and its space-filling partner at position 58, respectively, which were considered to potentially impact CDR orientation and the antigen binding pocket. Thus, these two residues were back mutated to its mouse corresponding amino acids of W47H and N58F, and HC2 contained these two backmutations in addition to V71R. Finally, in the construct HC3, the human residue S35 was also considered to potentially impact the CDR conformation and the antigen binding pocket. Thus, it was back mutated to its mouse corresponding amino acid of N (S35N) and the HC3 construct contained V71R, W47H, N58F and S35N backmutations (Fig. 2B).

The designed humanized LCs and HCs were synthesized and cloned into antibody expression vectors. A total of 16 different HC/LC antibody clones were generated by pairing each LC and HC variant, including the HC and LC of the grafted antibody without any back mutations in the framework. Transfection of ExpiCHO cells with small-scale plasmid preps was conducted, and the yield of each antibody clone was estimated using protein G probe capture. The chimeric mAb was also produced to serve as the reference control. This comprehensive rational design approach led to the development of engineered humanized variants that were assessed for their binding activity to CotH3 recombinant protein using Bio-Layer Interferometry (BLI) (13). Each variant was loaded onto Protein G sensors in the binding buffer, and then 30 nM of CotH3 antigen was introduced, allowing us to monitor the on-rates for 300 seconds, followed by the dissociation step for another 300 seconds. Only three humanized variants, namely LC3:HC3 (VX-01), LC3:HC2 (VX-02), and LC2:HC3 (VX-03), exhibited distinct binding profiles to CotH3, similar to the reference chimera (table S2). No detectable binding was observed for any of the variants with the grafted LC or with LC-1, as well as with the variants containing the grafted HC or HC-1. Furthermore, all three variants, along with the chimera, demonstrated negligible dissociation.

To further characterize the binding kinetics, we performed steady-state analysis using a serial dilution of CotH3 from 480 nM to 7.5 nM for all three humanized variants and the chimera(15). The K_D values for the variants were similar, ranging from 4 to 6 nM. Both the chimera and VX-01 exhibited K_D values of 5.2 and 5.4nM, respectively, with high R² values of 0.94 and 0.95, indicating robust and reliable binding (Fig. 2C). Collectively, these data show that the grafted antibody's binding affinity was successfully restored through backmutations and the objective of achieving an improvement in binding affinity above the parental murine mAb was successfully accomplished with the humanized variants.

The back mutated mAbs specifically bind to CotH3 protein expressed on Mucorales spores

We investigated the specificity of the three generated mAb variants (VX-01, VX-02, and VX-03) in binding to CotH3 protein. Flag-tagged CotH3 protein was expressed in S. cerevisiae and the entire cell lysate was separated on an SDS-PAGE and immunoblot analysis was performed using each of the generated mutated mAbs (VX-01, VX-02, and VX-03). Whereas VX-01 and VX-02 recognized only the CotH3 recombinant protein at 65kDa, VX-03 recognized two additional bands at approximately 48 kDa and 35 KDa. These two additional bands recognized by VX-03 likely represent degradation of CotH3 protein since VX-03 did not recognize any bands from S. cerevisiae transformed with empty plasmid (Fig. 3A). None of antibody variants recognized proteins extracted from the A549 human alveolar cell line (Fig. 3A).

Fig. 3. Humanized antibody variants bind to CotH3 antigen and Mucorales fungi.

(A) Immunoblot analysis testing binding of VX-01, VX-02, and VX-03 to CotH3 antigen or a human lung epithelial cell line total lysate. (B) Binding of the control human isotype-matched IgG1, VX-01, VX-02, VX-03, and the parental mouse C2 antibodies to R. delemar, Cunninghamella, M. circinelloides, Rhizomucor, and Lichtheimia was measured by flow cytometry.

Next, we studied the binding of the humanized mAbs to Mucorales fungi by immunostaining and flow cytometry, with the anti-CotH3 C2 as the positive control. The results indicated that all three humanized antibodies (VX-01, VX-02, and VX-03) demonstrated recognition of various Mucorales fungi including Rhizopus, Cunninghamella, Mucor circinelloides, Rhizomucor, and to a lesser extent Lichtheimia (Fig. 3B, table S3). This reduced binding to Lichtheimia can be attributed to the reduced identity of the 16-mer peptide in Lichtheimia (5 mismatched amino acids) when compared with the other Mucorales fungi (fig. S3A) to which the original C2 murine mAb was raised against. Alternatively, the reduced binding can be attributed to the reduced expression of CotH ortholog in Lichtheimia spores when compared with R. delemar (fig. S3B). Nonetheless, these data also show that the humanized variants retained the binding specificity of the parental mouse antibody and universally recognize CotH3 on Mucorales fungi.

The humanized antibodies enhance opsonophagocytic ex vivo and protect mice from mucormycosis

We previously showed that anti-CotH3 polyclonal antibodies enhance polymorphonuclear leukocyte (PMN)-mediated damage of R. delemar through opsonophagocytic killing (11). To further investigate the activity of the humanized antibodies in enhancing the phagocytosis of R. delemar, we conducted in vitro experiments using human polymorphonuclear leukocytes (PMNs) isolated from human peripheral blood cells of healthy individuals. The extracted neutrophils were then incubated with R. delemar at a 1:1 or 1:2 ratio (PMN:spores) in the presence of either the control isotype IgG or the humanized antibodies at 37°C for 2-3 hours. Only VX-01, but not VX-02 or VX-03, more than doubled the ability of neutrophils to kill R. delemar indicating that this humanized variant is opsonophagocytic and further validate its potential as a potent immunotherapeutic agent against mucormycosis (Fig. 4A).

Fig. 4. The VX-01 humanized Ab enhanced OPK activity of R. delemar and protected mice from mucormycosis.

(A) The ability of PMNs harvested from healthy volunteers to phagocytize R. delemar was measured after 3 hours of incubation in the presence of VX-01, VX-02, and VX-03 (PMN:spore ratio 1:1 or 1:2, n=8). An isotype IgG1 was used as a control. Data are median ± interquartile range. Data were analyzed by one-way ANOVA with Dunnett’s multiple comparisons test. (B) Shown is the survival of neutropenic mice infected intratracheally with R. delemar. After 24 hours, mice were treated with 30 μg of VX-01, VX-02, VX-03, mouse C2 antibodies, or the human isotype-matched control IgG1 (n=10 per group, except for n=5 for uninfected mice). Data were analyzed by Log-rank (Mantel-Cox) test comparing each treatment with the isotype IgG1 control.

We next evaluated the ability of the humanized antibodies to protect against murine mucormycosis. We used our established and clinically relevant neutropenic mouse model in which infection with R. delemar is induced through the intratracheal route. Treatment with single dose of 30 μg of an isotype-matched IgG1, the murine mAb (C2), or its humanized versions (VX-01, VX-02, or VX-03) were given 24 hours post infection through the intraperitoneal (IP) route and survival of mice was determined over a 21-day period. The dose of the antibody was selected based on historical data obtained with the murine C2 antibody(11). Although VX-01 and VX-02 both significantly (P<0.05) prolonged overall survival of mice infected with R. delemar compared with isotype control, only VX-01 resulted in protection identical to what is afforded by the C2 clone. Specifically, C2 mouse antibody or VX-01 humanized antibody resulted in a median survival time of >21 days and 60% overall survival versus 7 days median survival time and 0% overall survival for mice treated with isotype-matched IgG1. Treatment with VX-02 resulted in a 10-days median survival time, and an overall survival of 30% (Fig. 4B).

We also investigated if the generated humanized antibodies with improved binding to CotH3 protein had any direct effect on the viability or germination of R. delemar. We focused on VX-01 because it gave us the most consistent results in enhancing opsonophagocytic killing activity of human neutrophils and had superior protection in mice to VX-02 and VX-03. Incubating R. delemar in the presence of varying concentrations of VX-01 (ranging from 0.1 to 100 μg/ml) or the mouse C2 mAb for 24 hours had no effect on the metabolic activity of the fungus as determined by XTT assay (fig. S4). Similarly, germination of R. delemar spores in the presence 100 μg/ml of VX-01 or mouse C2 mAbs for 6 hours had no effect on the ability of the fungus to germinate or change the length of the germ tube (fig. S5).

VX-01 reduces R. delemar ability to invade and damage endothelial cells and abrogates R. delemar-mediated vascular leak

Angioinvasion and vascular leak are hallmarks of mucormycosis(2). We previously showed that anti-CotH3 antibodies protected mice infected intratracheally with R. delemar from angioinvasion and subsequent hematogenous dissemination to the brain(11). Thus, we investigated if VX-01 would protect human umbilical vein endothelial cells (HUVECs) from R. delemar-induced invasion and subsequent damage. Consistent with our previously reported data, VX-01 had no effect on the ability of R. delemar to adhere to HUVECs but significantly (P<0.0001) reduced R. delemar-mediated invasion (endocytosis) by >30% when compared with isotype-matched IgG1 (Fig. 5A). Further, incubating HUVECs with 50 μg/ml of either an isotype-matched IgG1 or VX-01 had no damaging effect on these mammalian primary cells in the absence of R. delemar infection, indicating that the humanized VX-01 antibody has favorable safety profiles. Incubating R. delemar infected HUVECs cells with 50 μg/ml VX-01 resulted in approximately 40% protection of these cells versus cells infected and incubated with the isotype-matched IgG1 (Fig. 5B). Furthermore, using a permeability assay in which HUVEC were grown on membrane inserts in 24-well transwell plates and using fluorescent dextran to determine migration of the fluorescence signal from the upper chamber to the lower chamber after addition of R. delemar (Fig. 5C), we show that VX-01 antibody almost completely abrogated the ability of R. delemar to cause vascular leak when compared with isotype-matched IgG1 or E. coli lipopolysaccharide (LPS), a potent inducer of vascular permeability (Fig. 5D). Collectively, these findings underscore the potential of VX-01 as an effective agent in neutralizing angioinvasion and vascular leak, two processes are essential to mucormycosis pathogenesis and are mediated by CotH proteins(9, 16).

Fig. 5. VX-01 prevents HUVECs cell damage and vascular permeabilization by R. delemar.

(A) Adherence and invasion (endocytosis) of R. delemar to HUVECs was measured in vitro in the presence of isotype IgG1 or VX-01 (50 μg/ml). Adherence and endocytosis were determined by the differential fluorescence of the R. delemar under the microscope (n=30 per group) after 1.5 hours of incubation with VX-01 or isotype-matched control IgG1. HPF, high powered field. (B) The ability of R. delemar to injure HUVECs after 6 hours of incubation was measured in the presence of VX-01 (50 μg/ml) by the ⁵¹Cr release method. The human isotype-matched IgG1 served as a control (n=6 wells from 2 experiments). (C) A diagram depicting the cell permeability assay. (D) The ability of R. delemar to induce permeability of HUVECs after 4 hours of incubation was measured in the presence of VX-01 (50 μg/ml) by measuring the amount of fluorescent dextran migrating from the upper chamber to the lower chamber of the transwell (n=9 to 18 per group from 2 individual experiments). Human isotype-matched IgG1 served as a negative control, and LPS was used as an inducer of vascular permeability. Data are presented as median ± interquartile range. Data were analyzed by unpaired, two-sided Student’s t test.

Pharmacokinetics of VX-01 and the parental murine antibody using the neutropenic mouse model

To facilitate the clinical development of the VX-01 antibody and optimize dosing, we conducted a pharmacokinetics (PK) study to characterize its behavior in the mouse model. To mimic the clinical scenario, we compared the PK of VX-01 to the mouse C2 antibody using the neutropenic mouse model infected with R. delemar. VX-01 or the parental C2 antibody were administered intravenously (IV) at different doses (10 μg, 30 μg, or 100 μg). Blood samples were collected at 0.5 hours, 2 hours, 4 hours, 8 hours, 24 hours, 72 hours, 120 hours, and 168 hours after antibody administration. Both VX-01 and C2 displayed bi-exponential serum concentration-time profiles, characterized by a short distribution phase and a long elimination phase (Fig. 6, A to C). When examining the average concentration during the elimination phase (after 8 hours of IV injection), we observed that the average concentrations for the 10 μg, 30 μg, and 100 μg of VX-01 injections were 0.57 μg/ml, 4.08 μg/ml, and 8.94 μg/ml, respectively. In contrast, the average concentrations for the 10 μg, 30 μg, and 100 μg murine C2 antibody injections were 2.51 μg/ml, 9.68 μg/ml, and 25.6 μg/ml, respectively. These results indicate that the murine antibody displayed greater stability in mouse serum compared with the humanized antibody. Despite lower concentrations of VX-01 in the elimination phase, it had comparable serum half-life to C2 antibody that ranged from approximately 3 to 7 days (Fig. 6C). We also compared the efficacy of the three VX-01 doses given intravenously (IV) in protecting mice from R. delemar intratracheal infection. Although all doses resulted in a significant (P<0.05) improvement in survival of mice, the 30 and 100 μg doses resulted in comparable enhanced survival of 40% when compared with mice infected with isotype-matched IgG1 (Fig. 6D). Because the IV and the IP routes showed similar protection in neutropenic mice infected with R. delemar, we investigated the breadth of protection of VX-01 in neutropenic mice infected with second most common cause of mucormycosis, Mucor circinelloides(2). VX-01 and C2 resulted in prolonged survival of 70% and 40%, respectively versus 10% for isotype-matched placebo mice (Fig. 6E). Taken together, these results show that VX-01 has comparable PK parameters to C2 and produces similar protection to mice from mucormycosis due to R. delemar or M. circinelloides regardless of the antibody route of administration.

Fig. 6. VX-01 has comparable pharmacokinetics and in vivo efficacy to C2 mouse mAb.

(A and B) VX-01 (A) or C2 mouse IgG1 (B) were given intravenously at a single of 10 μg/each neutropenic mouse (green), 30 μg (gray), or 100 μg/each (black). The solid line represents the simulated pharmacokinetic (PK) profile using in vitro parameters, and the dots represent the observed concentrations (n=4 per time point). Data are presented as median ± interquartile range. (C) PK data for VX-01 and mouse C2 antibody, including Cmax (peak concentration), T_1/2 (terminal half-life), and AUC (area under the curve). Results are presented as mean values. (D and E) Shown is the survival of neutropenic mice (n=10 per group, except for n=5 for uninfected mice) infected intratracheally with R. delemar (D) or M. circinelloides (E) and 24 hours later treated with 10 μg, 30 μg, 100 μg of VX-01 or human isotype-matched IgG1 as the control group. P values are versus isotype-matched IgG1 treated mice. Survival data were analyzed by Log-rank (Mantel-Cox) test.

VX-01 protects neutropenic and DKA mice from pulmonary mucormycosis primarily by prevention of angioinvasion and enhancement of phagocyte killing

Having optimized the dose and route of administration of VX-01, we embarked on a detailed study to characterize the therapeutic benefits of the humanized antibody. We tested VX-01 in the neutropenic mouse model of R. delemar infection and used survival, tissue fungal burden, and histopathological examination as endpoints. As expected, both VX-01 and the mouse C2 antibodies resulted in similar protection characterized by a median survival time of 14 and 13 days, respectively, and an overall survival of 38% when compared with isotype-matched IgG1-treated placebo mice (Fig. 7A). Additionally, both antibodies resulted in similar reduction of lung fungal burden of approximately 1-log versus isotype-matched placebo mice when the organ was harvested on day 4 post infection (Fig. 7B). Finally, histopathological examination of lung tissues sampled on the day 4 post infection showed that mice treated with the isotype IgG displayed wide fungal infiltration and abscesses with tissue edema. In contrast, the lungs of mice treated with either VX-01 or C2 showed no visible fungal presence and normal lung architecture (Fig. 7C).

Fig. 7. VX-01 protects neutropenic and DKA mice from R. delemar infection and enhances neutrophil influx.

(A and B) Neutropenic mice were treated with a 30-μg single dose (IP route) of VX-01 or mouse C2 antibodies at 24 hours after intratracheal infection. Survival (A) and lung fungal burden (B) were measured (n=8 to 9 per group, except for n=5 for uninfected mice). (C to F) Neutropenic mice were treated as in (A), and tissues were collected 96 hours after intratracheal infection. (C) Grocott’s methenamine silver (GMS) staining of the lung is shown to evaluate hyphae invasion. Scale bars, 200 μm (low magnification), 50 μm (high magnification, inset). (D) Immunohistochemistry staining showing VEGF expression (red) in lungs of neutropenic mice infected with R. delemar and treated with Isotype IgG1 control or VX-01. Scale bars, 500 μm. (E) Quantification of fluorescence intensity of the VEGF staining in micrographs of (D) (n=20 random fields). (F) The number of macrophages in the lungs of infected mice was quantified by flow cytometry after staining with cell surface marker of macrophages (n=5 per group) (G) DKA mice (n=20/group, except for 10 for uninfected mice) were infected with R. delemar and treated as in (A) and survival was measured. (H) MPO abundance was measured in the lungs and spleens of DKA mice (n = 5 or 10 mice per group) treated with VX-01 or isotype control. Lungs were harvested 96 hours after intratracheal infection with R. delemar. A single dose of 30 μg of VX-01 or human isotype-matched IgG1 was given 24 hours after infection. MPO values are presented as raw values in nanograms per gram of the organs. (I) RT-PCR of Mpo gene expression in lungs and spleen of DKA mice (n=4 mice per group) infected with R. delemar and treated with either VX-01 or isotype-matched IgG1 as in (H). Data in (B) is presented as median + interquartile range and was analyzed by one-way ANOVA with Tukey’s multiple comparisons test. Data in (E, F, H, and I) are presented as median + interquartile range and were analyzed by the unpaired, two-sided Student’s t test. Data in (A and G) were analyzed by Log-rank (Mantel-Cox) test comparing with isotype IgG1 control.

We investigated the mechanism by which VX-01 protected neutropenic mice from mucormycosis. To investigate if VX-01 prevented fungal invasion, we conducted immunohistochemistry staining of lung sections of mice infected with R. delemar and treated with either VX-01 or isotype matched IgG1 for vascular endothelial growth factor (VEGF), a potent angioinvasion and microvascular permeability enhancing cytokine(17). VX-01 treatment significantly (P<0.0001) resulted in reduced expression of VEGF in the lungs of mice infected with R. delemar compared with lung sections from infected mice treated with isotype-matched IgG1 (Fig. 7, D and E). Consistent with VX-01 effect on reducing lung fungal burden and decreasing fungal angioinvasion (Fig. 7, B to E), we detected fewer lung macrophages in lungs of infected mice treated with VX-01 versus those treated with isotype-matched IgG1 (Fig. 7F). These results indicate that the VX-01 mechanism of protection in neutropenic mice is reliant on prevention of angioinvasion.

We also validated the therapeutic benefits of VX-01 compared with C2 using the DKA mouse model of mucormycosis(18). A single dose of 30 μg of VX-01 or C2 significantly (P<0.0001) prolonged overall survival of DKA mice by 60% to 0% survival of mice treated with the isotype-matched IgG1 (Fig. 7G). Because DKA mice have intact neutrophil count and since we previously showed that anti-CotH3 antibodies enhance mouse neutrophil function(11), we evaluated the role of VX-01 on neutrophil chemotaxis. We focused on myeloperoxidase (MPO) which is recognized for its pivotal role in neutrophil function and antimicrobial defense(19). As expected, the DKA mice treated with the VX-01 exhibited substantially higher concentrations of MPO in the lungs and spleen, but not in the brain (Fig. 7H), likely because VX-01 prevents angioinvasion and dissemination to secondary target organs(11). We further investigated the mechanism by which VX-01 augments MPO production in target organs. Lungs and spleens collected from infected mice and treated with VX-01 or isotype-matched IgG1 were gently disrupted to produce single cells followed by flow cytometry analysis to identify live neutrophil populations. There was no difference in the number of live neutrophils in lungs or spleen harvested from mice treated with VX-01 or isotype-matched IgG1 (fig. S6, A and B). Furthermore, there was no difference in intracellular MPO staining among neutrophils identified in lungs or spleen of mice treated with VX-01 or isotype-matched IgG1 (fig. S6, C and D). To investigate if enhanced mRNA expression of the gene encoding MPO in infected and treated organs with VX-01 or isotype matched IgG1 accounted for the increased detection of MPO in the lungs and spleen of infected mice lungs and spleen that have been treated with VX-01, we harvested lungs or spleens from mice infected with R. delemar and treated with a single dose of VX-01 on day 4 post infection and conducted reverse transcription polymerase chain reaction (RT-PCR) targeting the Myeloperoxidase (Mpo) gene. Indeed, lungs or spleen harvested from VX-01 treated mice showed 1.5-to-5-fold increase in Mpo gene expression when compared with corresponding organs harvested from mice treated with isotype-matched IgG1 (Fig. 7I). These results show that VX-01 augments expression of Mpo in live neutrophils and underscores the role of VX-01 in opsonophagocytic killing activity in the DKA mouse model.

VX-01 can be combined with antifungal drugs to protect mice from mucormycosis

Because any immunotherapy-based drug is likely to be used as an adjunctive therapy with currently used antifungal drugs, we tested the effect of administering VX-01 with either liposomal amphotericin B (LAMB; Fig. 8, A to C) or posaconazole (POSA; Fig. 8, D to F) using the neutropenic mouse model. LAMB is considered as first-line therapy for mucormycosis, whereas POSA is reserved for salvage or step-down therapies(20). We used a severe model of infection of initiating fungal intratracheal inoculation and starting treatment 48 hours post infection to mimic delayed diagnosis of the disease and to enhance the chance of detecting improved efficacy of combined therapy. Treatment with VX-01 alone demonstrated reduced but significant (P<0.05) protection (20% survival in Fig. 8, A and D). Conversely, treatment with LAMB or posaconazole alone did not improve survival, with 20% and 0% survival, respectively (Fig. 8, A and D). However, mice treated with a combination of VX-01 and either LAMB or posaconazole achieved 70% survival (Fig. 8, A and D) and remained healthy throughout the experiment. In an independent experiment, tissue fungal burden of lung and brain (18) corroborated the survival studies, with combination therapy with LAMB (Fig. 8, B and C) or POSA (Fig. 8, E and F) outperforming monotherapy to reducing lung and brain fungal burden by approximately 2-log. Finally, we confirmed these results against murine mucormycosis due to M. circinelloides by using the same delayed and severe model. Again, combination therapy resulted in enhanced survival of neutropenic mice over mice treated with isotype-matched IgG1 or monotherapy with approximately 40% overall survival for either LAMB or POSA combinations (fig. S7, A and B). These findings collectively demonstrate that VX-01 immunotherapy enhances the efficacy of antifungals for treating mucormycosis.

Fig. 8. VX-01 treatment shows improved efficacy when administered with clinically used antifungal drugs.

(A to C) Neutropenic mice (n=10 per group, except for n=5 for infected mice) were infected with R. delemar, and 48 hours later were treated with either a single dose of 30 μg of VX-01, LAMB (10 mg/kg, qd) for 4 days, or a combination of both. Endpoints were survival (A) or tissue fungal burden in the lung (B) or brain (C) as determined by qPCR. (D to F) The same model was used with similar treatment except that LAMB treatment was substituted with posaconazole (30 mg/kg bid for 7 days) (n =10 per group, except for n=5 for uninfected mice), Endpoints were survival (D) or tissue fungal burden in lung (E) or brain (F) as determined by qPCR. Survival data in (A and D) were analyzed by Log-rank (Mantel-Cox) test. P values in the survival curve are versus isotype-matched IgG1 treatment. Data in (B, C, E, and F) are presented as median ± interquartile range. Statistical comparisons in (B, C, E, and F) were made by one-way ANOVA with Tukey’s multiple comparisons test.

Prolonged exposure of R. delemar to VX-01 does not induce resistance

Antibody resistance, characterized by the ability of pathogens to evade the effects of therapeutic antibodies, is a concern that could limit the effectiveness of antibody-based treatments especially with repeated dosing that is likely to be needed for treating patients with mucormycosis(21). To investigate the possibility of Mucorales developing resistance to VX-01, R. delemar was serially passaged in the presence of VX-01 or human isotype-matched IgG1 for 20 generations, followed by flow cytometry analysis to confirm sustained binding of VX-01 to the fungal spores. VX-01 maintained recognition of R. delemar that has been passaged for 20 generations in the presence of VX-01 thereby indicating the lack of resistance to VX-01 in this fungal isolate (fig. S8A). Next, we compared the protective activity of VX-01 against R. delemar strain that was passaged for 20 generations to wild-type R. delemar using the neutropenic mouse model of intratracheal infection. VX-01 has similar protective activity of 40% overall survival against the wild-type R. delemar and the strain that was passaged for 20 generations in the presence of VX-01 (fig. S8B) (P<0.01). Therefore, these data suggest that treatment with VX-01 may not induce resistance.

VX-01 does not bind to cell surface proteins but has weak to moderate binding affinity to cytoplasmic proteins of human tissues

To further investigate the toxicity of VX-01, we conducted GLP-tissue cross reactivity studies using human samples from at least three donors per tissue (Charles River Laboratories). VX-01 was applied at two concentrations (2.5 and 0.5 μg/ml) to cryosections of normal tissues. VX-01 showed cross-reactive binding in epithelial cells of multiple tissues including neurons and glial cells in the central nervous system (CNS); mononuclear cells in lymphoid tissues including mucosal associated lymphoid tissue and peripheral blood; and in male and female reproductive tissues (stromal and germ cells). All the tissue elements with test article binding had granular cytoplasmic staining only with no membrane labeling. In general, the character and distribution of cellular staining was similar at both concentrations, however, reduced in intensity and frequency, or absent at 0.5 μg/ml (fig. S9).

Transfected and clonally derived CHO cells can produce commercial yields of VX-01

Because of these encouraging results, we initiated the manufacturing process development of VX-01. Stable clonal cell lines expressing VX-01 were generated using JOINN-CHO host cells (CHO-K1 derivative). Top clones showed a productivity range of 2.2 to 3.2 g/L with the peak viable cell density ranging from 3.4 to 7.1 x 10⁶/mL in a shake flask fed-batch evaluation. Research cell bank (RCB) of three top clones were prepared and were free from mycoplasma and were sterile. The cell count, viability, and recovery rate matched with those upon freezing and are within the acceptable ranges (Cell viability of > 90% before and after thawing and total number of cells of approximately 10 x10⁶ per vial). Additionally, using the National Institute of Allergy and Infectious Diseases (NIAID) preclinical service, a master cell bank (MCB) was established and used for conducting upstream and downstream manufacturing processes; the production process successfully scaled-up to a 200 L bioreactor, and production was followed by conventional downstream purification for generation of bulk-drug substances. These studies showed that top clone of CHO producing VX-01 could reach a viable cell density (VCD) of 3.0 to 5.5 x 10⁶ cells/mL with viability of >85% after 14 days of culturing in different bioreactor scales from Ambr250, 2L, 10L and 50L (fig. S10, A and B). All bioreactors in different scales produced acceptable titers of 2.5 to 3.5 g/L during the production stage (fig. S10C).

DISCUSSION

Mucormycosis, the third most common invasive fungal infection in hematopoietic centers, poses a major global health challenge. Despite antifungal drug treatment and aggressive surgical debridement, mortality rates remain unacceptably high. The coronavirus disease 2019 (COVID-19) pandemic has exacerbated the situation in which thousands of COVID-associated mucormycosis (CAM) cases were reported around the globe(22). Most of the CAM cases were reported in India among patients treated with high doses of corticosteroids with diabetes as an underlying predisposing factor(22). The use of corticosteroids exacerbate hyperglycemia and elevated available serum iron likely resulting in increased expression of GRP78(10, 16), which, besides acting as a co-receptor for SAR-CoV-2(23, 24), binds to CotH3 invasin on inhaled Mucorales spores favoring nasal epithelial cell invasion and subsequent hematogenous dissemination(9, 16, 22, 25). In addition, corticosteroids are known to impair the ability of macrophages to phagocytose microbes and suppress their inflammatory response(26). Therefore, interrupting interactions between CotH fungal ligands and their host receptors have the potential to introduce a potent adjunctive therapy that will minimize host cell invasion during mucormycosis. In this context, the use of antibodies carries the promise of blocking invasion and potentially enhance immune recognition and pathogen elimination by mediating opsonophagocytic killing activity in a host that usually suffers from immunosuppression.

We previously reported on the isolation of a murine anti-CotH mAb (C2) that was protective against murine mucormycosis especially when combined with currently used antifungal drugs to treat the disease(11). Although the mechanism of C2 protection was not fully characterized in that study, polyclonal antibodies targeting CotH3 were found to prevent host cell invasion, and subsequent damage and enhance opsonophagocytic killing activity of phagocytes(11). To fully realize the therapeutic potential of C2, a humanization step is needed to: (1) minimize immunogenicity; (2) enable human effector functions (e.g. opsonophagocytic killing) through the human C region; and (3) increase the antibody half-life in serum(27). Antibody humanization, first described by Jones and colleagues in 1986, involves implanting a mouse CDRs into a full human antibody framework to achieve a grafted humanized antibody with the same antigen-binding specificity(28). In certain cases, loss of recognition of the antibody binding target occurs due to the introduction of the human sequences. In this case and to fully restore the binding affinity, additional contact residues in the human frame must be mutated back to the mouse residues(29). Our humanization process with back mutations to 10 amino acid residues resulted in isolating a humanized clone that avidly binds to CotH3 by approximately 10-fold higher than the murine C2 and has a similar binding affinity with a chimeric mAb. There are several factors that can contribute to the enhanced binding affinity of the humanized VX-01 and chimera antibodies including: (1) the human Fc region used in VX-01 and the chimera antibodies can influence overall binding by providing better stability and a more optimal conformation for the Fab region, potentially enhancing binding affinity; and (2) the glycosylation patterns on the Fc region can affect the overall structure and function of the antibody. Human cells such as CHO cells often produce antibodies with different glycosylation patterns compared with mouse cells(30). Although the enhanced binding of VX-01 was only achieved by mutating 10 amino acids, it is prudent to mention that Food and Drug Administration (FDA) approved mAbs have up to 16 framework mutations(31).

VX-01, was able to block R. delemar-induced HUVEC injury and vascular permeability in vitro, processes that are critical in mucormycosis pathogenesis(20). In the neutropenic mouse model, VX-01 appears to protect mice from mucormycosis by reducing the ability of R. delemar to angio-invade as shown by the reduced expression of VEGF, a marker of angioinvasion and microvascular permeability(17), in the lungs of mice treated with VX-01 compared with lungs treated with isotype-matched control IgG1. These results were corroborated with reduced lung fungal burden in VX-01-treated mice which reduced numbers of lung macrophages (Fig. 7, C to E). In the DKA mouse model, it appears that the mechanism of VX-01-mediated protection against mucormycosis might constitute both prevention of angioinvasion and augmenting of opsonophagocytic killing activity, as demonstrated by enhanced detection of total MPO in lungs and spleens of DKA mice coupled with enhanced expression of the MPO gene in live neutrophils isolated from both organs (Fig. 7, H and I). It is prudent to mention that the number of neutrophils detected in organs harvested from VX-01-treated or isotype IgG1-treated mice were similar (fig. S6), hence suggesting that the augmented expression of Mpo gene in live neutrophils from VX-01-treated mice is followed by enhanced secretion of MPO to the infection milieu in a mechanism that is primarily not reliant on NETosis(32, 33).

Although the immunomodulatory function of VX-01 is expected to be maximized in individuals who have some immune effector functions (e.g. DKA hosts or immunocompetent trauma patients), neutropenic hosts maintain tissue macrophages which are derived from embryonic and not hematopoietic origin(34). Thus, it is possible that opsonophagocytic killing activity of tissue resident phagocytes might be modulated by VX-01 in neutropenic hosts. Importantly, the ability of VX-01 to prevent invasion and damage to host cells and to enhance opsonophagocytic killing activity are features maintained from the murine antibodies and are likely critical for demonstrating improved efficacy in combination with antifungal activity. It is thought that the reduction of tissue infarction by diminishing the ability of Mucorales to invade host tissues and the reduction of the tissue fungal load through opsonophagocytic killing activity are essential for adequate delivery of antifungals to the site of infection.

The field of therapeutic antibodies has experienced rapid growth and widespread application in treating cancer, autoimmune diseases, and infectious diseases(35). Invasive fungal infections pose a persistent threat, necessitating new therapeutic options. mAbs have emerged as promising candidates for eliminating microbial pathogens, as evidenced by our findings demonstrating enhanced neutrophil and peritoneal macrophage activity against R. delemar in vitro and in vivo. Unlike bacterial and viral infections, fungal infections lack efficient vaccines due to their occurrence predominantly in severely immunocompromised patients, making it challenging to develop autologous protection through vaccination. Hence, passive immunization using mAbs remains a promising approach, particularly for mucormycosis, which exhibits resistance to most antifungal drugs and high mortality rates. We found that the combination of humanized mAbs with antifungal drug treatments improved survival of mice and eliminated fungal presence in the lungs and brain, where antifungal drugs alone proved ineffective.

Before advancing to human clinical trials, it is imperative to assess the functionality and potential therapeutic effects of the humanized antibody within a relevant in vivo model(36). Mice are commonly used as a preclinical model due to their physiological similarities to humans and the availability of mouse strains with different genetic backgrounds(37). Previous research has shown that human IgG subclasses exhibit comparable relative affinities for mouse Fcγ receptors (FcγR) as they do for their human counterparts, suggesting that preclinical assessment of human IgG1 in mouse models may remarkably mimic FcγR-mediated effector functions(38). Consequently, evaluating the humanized antibody's functionality in mice has become a standard practice in preclinical research and drug development(39). The PK studies showed that the 30-μg therapeutic dose of either murine C2 or humanized VX-01 has a serum half-life of approximately 4 days, which is consistent with the reported half-life of IgG in mice(40). Further, despite the murine C2 antibody having more than two-times higher concentration in the elimination phase compared with VX-01, both treatments resulted in similar survival rates (Fig. 4B). This may be attributed to the 10-fold increase in the binding affinity of VX-01 to CotH3 antigen on Mucorales when compared with the murine C2 antibody (compare Fig. 1A for C2 and Fig. 2C for VX-01). It is anticipated that VX-01 (an IgG1) is likely to have a serum half-life of approximately 3 weeks in humans(41, 42). Thus, with the higher binding capacity of VX-01, it is anticipated that this antibody could be dosed in humans once every 3 to 4 weeks when given with antifungal therapy at a low dose of about 1.25 mg/kg. Given the fact that our manufacturing processes indicated that the transfected CHO cells can produce VX-01 in titers close to 2 to 3 g/L, a 200 L fermentation run can produce at least 5000 doses of the antibody, assuming an average patient of 60 kg.

Mechanisms of antibody resistance include altered antigen expression or binding, impaired complement-mediated cytotoxicity or antibody-dependent cellular cytotoxicity, altered intracellular signaling effects, and inhibition of direct induction of cell death(43). Despite these challenges, mAbs offer promise as targeted therapies due to their high specificity and reduced toxicity. Our studies of passaging R. delemar in the presence of VX-01 for 20 generations and then evaluating the binding ability and therapeutic efficacy of VX-01 against the 20^th generation indicate that resistance to VX-01 is a low probability. This could be attributed to the fact that VX-01 has no static or cidal activity against Mucorales fungi and therefore is less likely to exert any pressure on the fungus.

Furthermore, as the number of mAbs in the market grows, the safety profile of these antibodies, such as host cross-activity and immunogenicity, becomes increasingly important(44). VX-01 had no damaging effect on HUVEC cells even at the high concentration of 50 μg/ml, which still effectively prevented R. delemar invasion and HUVEC damage. Further, human tissue cross-reactivity studies showed a favorable safety profile of weak or moderate binding of VX-01 to only cytoplasmic proteins without any indication of binding to cell surface proteins. Although this result was not expected since CotH antigens are present only in Mucorales fungi and absent from human cells, cytoplasmic protein binding is considered of little toxicologic importance due to the limited ability of antibody therapeutics to access the cytoplasmic compartment in vivo(45, 46). Good laboratory practice (GLP)-toxicity and rat tissue cross-reactivity studies are currently underway to fully evaluate the safety of VX-01 prior to advancing it into clinical trials.

Some limitations to the study exist. Although we have shown that the mouse C2 mAb protects against murine mucormycosis due to Rhizopus, Mucor, and Lichtheimia (albeit to a lesser extent probably due to the lower expression of CotH3 in Lichtheimia)(11), here we only showed efficacy of VX-01 in treating murine mucormycosis due to R. delemar or M. circinelloides. Additionally, the productivity of VX-01 in CHO cells can be improved to higher titers of ≥ 5 g/L. This would be beneficial for clinical use. Finally, although our studies of passaging R. delemar in VX-01 for 20 generations studies showed that fungal resistance to VX-01 did not develop, clinical resistance of R. delemar or other Mucorales fungi to VX-01 cannot be entirely ruled out.

In conclusion, our study highlights the efficacy and reliability of the humanized antibody targeting CotH3 proteins as a therapy, particularly when combined with antifungal drugs, for combating mucormycosis. This research is especially timely, given the surge in mucormycosis cases in India during the COVID-19 pandemic. Future research efforts will focus on optimizing gene-encoded mAb technology in small and large animal models, utilizing fed-batch or perfusion modes to achieve multi-gram production, and developing delivery systems for clinical use based on feedback from large animal data. These endeavors aim to pave the way for improved treatment options and better patient outcomes in the battle against invasive fungal infections.

MATERIALS AND METHODS

Study Design

This study aimed to evaluate the efficacy and safety of humanized mAbs targeting Mucorales fungi. The humanization process involved cloning and humanizing the variable regions of the C2 mAb, which was originally derived from a mouse hybridoma. The resulting humanized antibodies were subjected to a series of in vitro and in vivo tests to assess their binding affinity, functional activity, and therapeutic potential. The sample sizes for in vitro and in vivo experiments were determined based on prior studies and power analyses to ensure reliable effect measurements, with at least two experiments conducted in vitro and 6 to 10 mice per group in vivo. Data collection was guided by pre-defined endpoints such as survival times and fungal burden measurements, with stopping rules in place for early termination upon achieving primary endpoints or detecting adverse effects. The primary and secondary endpoints were pre-specified, and statistical corrections for multiple comparisons were applied. Experimental design included controlled laboratory experiments with randomization of treatment groups and blinding during outcome assessment and data analysis to minimize bias. Randomization was used for group assignments, and blinding was employed to ensure objective measurement and analysis of results.

Study approval

All procedures involving mice were approved by the Institutional Animal Care and Use Committee (IACUC) of The Lundquist Institute for Biomedical Innovations at Harbor-UCLA Medical Center, according to the NIH guidelines for animal housing and care (Protocols #11671 and 056151). Human endothelial cell and neutrophil collection was approved by the Institutional Review Board (IRB) of The Lundquist Institute for Biomedical Innovations at Harbor-UCLA Medical Center. Because umbilical cords are collected without donor identifiers, the IRB considers them medical waste not subject to informed consent. Blood collection for neutrophil harvest was obtained after the volunteer signing an IRB approved consent.

Organisms and culture conditions

R. delemar 99-880 (formerly classified as R. oryzae), M. circinelloides f. jenssenii UTHSCSA DI15-131 are clinical isolates obtained from the Fungus Testing Laboratory at the University of Texas Health Sciences Center at San Antonio (UTHSCSA). Cunninghamella. bertholletiae 182, also a clinical isolate was a gift from T. Walsh (Center for Innovation Therapeutics and Diagnostics, Richmond, Virginia, USA). Lichtheimia corymbifera is another clinical isolate obtained from the DEFEAT Mucor clinical study(47). Finally, Rhizomucor was obtained from a patient seen at the Harbor-UCLA Medical Center. Growth conditions and enumeration of cells and germination were previously described(11, 48).

Cloning the V_H and V_κ regions, humanization of the C2 mAb, and antibody production

The C2 mAb was isotyped to IgG1 kappa light chain. Thus, a reverse primer based on the mouse kappa light chain constant region was used for 5' Rapid Amplification cDNA Ends (5’ RACE) PCR to amplify the light chain variable region. A reverse primer based on the mouse heavy chain constant region was used for 5'-RACE PCR to amplify the heavy chain variable region. The amplified PCR products were cloned into a TOPO vector for sequencing analysis. To confirm the hybridoma sequencing result, the N-terminal protein sequencing was performed by Creative Proteomics using Edman degradation on the heavy and light chains of the murine mAb, that had been separated by SDS-PAGE and transferred to a PVDF membrane. Alignments of the murine V_H and V_κ regions with human germline antibody variable regions were performed to identify the best frameworks as acceptors for the murine C2 V_H and V_κ. Based on the alignment analysis, the human VH4-59*01 was selected for the HC humanization and the A18 V_κ was selected for the LC humanization. The gene encoding the designed grafted humanized variant was custom synthesized by Bio Basic.

A chimera and a grafted humanized IgG were then constructed and generated by AvantGen. The cloned VH was fused with human IgG1 coupled with a V_κ fused to the human kappa constant region to generate the chimera. The grafted humanized version with the selected human antibody frameworks and mouse antibody CDRs. For this, the DNA encoding the V_H variable region was cloned into BamH I and Apa I sites of AvantGen's pcDNA3.4-IgG1 vector and the sequencing was confirmed. V_κ regions for these constructs were cloned into the Afl II and Acc65 I sites of AvantGen's pcDNA3.4-kappa vector and sequencing confirmed. Large-scale plasmid preparations were prepared and used for transfecting ExpiCHO cells. Secreted antibody was purified from the culture medium 10 days post-transfection using Protein A affinity chromatography. The purified antibody was analyzed by SDS-PAGE gel to determine the purity of the antibody, which was estimated to be greater than 95% pure, and the antibody concentration was determined using UV absorbance at 280 nm.

Biolayer interferometry

The purified preparations of the chimera and grafted humanized antibodies were then used to assess the affinity of each clone for the CotH3 antigen by biolayer interferometry (BLI) using the Gator system (ProbeLife). A Protein G coated sensor probe (160006, ProbeLife) was loaded with the bivalent antibody test clone at a concentration of 5 μg/ml in ProbeLife’s kinetic buffer (Phosphate-buffered saline [PBS] pH 7.4, 0.02% Bovine Serum Albumin [BSA], 0.002% Tween-20, 0.005% Sodium azide [NaN₃]). A serial dilution of CotH3 antigen of 400, 200, 100, 50, 25, 12.5, 6.25 nM and buffer only (0 nM) were added to the sensors. All reactions were performed at 25°C. The values were chosen when three or more concentrations yield the same kinetic values. The original murine anti-CotH3 C2 was used as the reference antibody.

Back mutations in the grafted humanized variable regions

To restore the binding activity of the humanized antibody, computer modeling and analysis were performed for the C2 murine mAb and humanized grafted antibodies with the selected human antibody frameworks and mouse antibody CDRs. Protein Data Bank (PDB) structures 1XGP (heavy chain) and 2W60 (light chain) were chosen as homology templates for the murine variable regions and 4LKX for the human framework with the grafted murine CDRs. A structural assessment of the homology models was then carried out - models were visually assessed using Pymol (Schrödinger, Inc.) and residue sequence/structure commonalities and differences between the two models (mouse and humanized V_H and V_κ) were used to guide where backmutations might be warranted. DNA fragments encoding the humanized variants with backmutations were custom synthesized by Integrated DNA Technologies (IDT) and cloned into AvantGen's antibody expression vectors. Transfection of ExpiCHO cells and antibody expression and purification were conducted as above.

Immunoblotting

The Flag tag fused CotH3 antigen construct in the yeast plasmid pXW55-URA3 was previously described in details(11). For Western blotting, 0.1 μg Flag-CotH3 or 10 μg of whole cell lysate was used to separate proteins on an SDS-PAGE. Separated proteins were transferred to PVDF membranes (GE Water & Process Technologies) and treated with Western blocking reagent (Roche) for overnight at 4°C. C2 murine IgG1 or the generated chimera, humanized, or back-mutated humanized antibodies (0.02 μg/ml each), were used as primary antibodies. After 1 h, 0.2 μg/ml of horseradish peroxidase (HRP) conjugated anti-human IgG (Invitrogen, Cat #31412) or HRP conjugated anti-mouse IgG (Invitrogen, Cat # G-21040) secondary antibody were added for another 1 hour at room temperature. Flag-CotH3 bands were visualized by adding the HRP substrate (SuperSignal West Dura Extended Duration Substrate, Thermo Scientific), and the chemiluminescent signal was detected using an In-gel Azure Imager c400 fluorescence system (Azure Biosystems).

Binding of the produced antibodies to CotH proteins expressed on Mucorales

Mucorales spores were collected in endotoxin-free PBS containing 0.01% Tween 80 (PBST). Collected spores were washed with PBS and then fixed in 4% paraformaldehyde. The fixed cells were washed three times with PBS, and 2 x 10⁶ /100 μl were used for each staining. Briefly, cell pellets (100 μl) were blocked using of 2% BSA in PBS per tube. Each antibody (20 μg/ml) was used as the primary antibody in the blocking buffer. After 1 hour, the cells were washed 3 times with PBS + 0.05% Tween 80. Then 10 μg/ml of Alexa Fluor 488- IgG anti-human IgG secondary antibody (Invitrogen, Cat #A11013) in blocking buffer was added for another 1 hour at room temperature followed with 3 times washing with PBS + 0.05% Tween 80. One-milliliter samples of stained spores were analyzed using a LSR II (Becton Dickinson) instrument equipped with an argon laser emitting at 488 nm. Fluorescence emission was read with a 515/40 bandpass filter. Fluorescence data were collected with logarithmic amplifiers. The mean fluorescence intensities of 2 x10⁴ events were calculated using FlowJo v10 software(16).

Effect of antibodies on the growth of R. delemar in vitro and PMN killing assay

R. delemar spores (1 × 10⁵ cells/ml) were cocultured in 96-well flat-bottom plates at 37°C for 6 or 24 hours with or without VX-01, mouse C2 antibody, isotype-matched IgG1, or media alone and processed to determine the effect of the mAbs on the growth of the organism using XTT assay as we previously described(11). For PMN killing assay, R. delemar spores were collected and washed twice with Hanks' Balanced Salt Solution (HBSS) and counted. Spore counts were adjusted to 1 x 10⁵ spores/ml using 1X RPMI-1640 media supplemented with 10% human serum [Gemini Bio-Products], and 1% penicillin/streptomycin. One hundred μl of spores from the above stock were added to Eppendorf tubes. This was followed by adding 3 μl of 100 μg/ml stock of either VX-01, VX-02, VX-03, or the isotype-matched IgG1 prepared in 1 X RPMI-1640 media containing 10% human serum and 1% penicillin/streptomycin, and the tubes were stored on ice for 30 minutes. Next, the tubes were warmed up to room temperature for 30 minutes prior to adding human neutrophils isolated from healthy volunteers after an IRB approved consent form was obtained. Neutrophils were purified as described before(49), suspended in RPMI-1640 media supplemented with 10% human serum and 1% penicillin/streptomycin, and added at 1:1 or 1:2 ratio of neutrophils to spores. Tubes were incubated at 37°C for 3 hours, followed by 1 second sonication to kill all remining neutrophils. Tubes we then stored at 4°C overnight prior to serial dilution and plating on Potato Dextrose Agar (PDA) + 0.1% Triton plates. Surviving colonies were counted and % killing was determined by comparing colony counts with those collected from tubes treated similarly without the addition of neutrophils.

Fungal adhesion, invasion, endothelial cell damage, and transwell permeability assays

The number of R. delemar organisms adhering to and invading HUVECs was determined using a modification of our previously described differential fluorescence assay(50). At least 100 organisms were counted in 20 different fields on each slide. One slide per arm was used for each experiment, and the experiment was performed in triplicates on different days. For the damage assay, HUVECs were collected by the method of Jaffe et al.(51). HUVECs were propagated and processed for Rhizopus-induced damage assay according to our previously described method(48). To study the effect of VX-01 on vascular permeability in vitro, HUVECs were seeded on 24-Corning transwell plates with permeable polyester inserts (0.4 μm, Fisher) coated with fibronectin and processed for the effect of VX-01 on vascular permeability as we previously described using 50 μg/ml VX-01 or isotype-matched IgG1(48).

In vivo efficacy evaluation

All animal studies used either the DKA or neutropenic mouse models using male Institute of Cancer Research (ICR) mice (20 to 23 g) obtained from Envigo(11). Neutropenic or DKA mice were treated with a single dose of 30 μg VX-01 humanized antibody or isotype-matched human IgG1 starting 24 or 48 hours after infection administered by intraperitoneal injection. Survival served as the primary endpoint. As a secondary endpoint and in some experiments, the fungal burden in the lungs and brains (primary and secondary target organs) was determined 96 hours after infection by Quantitative polymerase chain reaction (qPCR) assay, as we previously described(18). Values were expressed as log10 spore equivalent per gram of tissue. Histopathological examination was carried out on sections of the harvested organs after fixing in 10% zinc formalin. The fixed organs were embedded in paraffin, and 5-mm sections were stained with hematoxylin and eosin.

For determining the protective activity of the mAbs with antifungal drugs, neutropenic mice were infected with R. delemar or M. circinelloides(53) and treated with the VX-01 humanized antibody, isotype IgG, posaconazole (Merck and Co.), LAMB (Gilead Sciences), or a combination of the VX-01 and either antifungal drug starting 48 hours after infection. VX-01 was given once by intraperitoneal injection at 30 μg, posaconazole was administered by oral gavage at 30 mg/kg twice daily for 7 days, and LAMB was given intravenously at 10 mg/kg per day for 4 days. Survival and tissue fungal burden served as endpoints as above. The MPO assay was conducted as we previously described(11).

Pharmacokinetics study

For the PK studies, infected neutropenic mice were intravenously injected with VX-01 or anti-CotH3 C2 antibodies at 10 μg, 30 μg, or 100 μg 24 hours after infection. Mouse serum samples were collected at 0.5 hours, 2 hours, 4 hours, 8 hours, 24 hours, 72 hours, 120 hours, and 168 hours after the antibody injection. The antibody concentration in the mouse serum samples was detected by ELISA using plates coated with 2 μg/ml recombinant CotH3, and anti-human-IgG-HRP or anti-mouse-IgG-HRP as the secondary antibody. The mAb concentration was determined from a standard curve obtained from serum spiked with known concentrations of the antibody. The mean concentration-time data for each antibody was evaluated, and Cmax, the area under the curve (AUC), and T_1/2 were calculated as was previously described(52).

Flow cytometry analysis

DKA or neutropenic mice were euthanized at day 4 post infection. After disrupting the excised lungs or spleen with a gentleMACs tissue dissociator (Miltenyi Biotec) to yield single cells, red blood cells were removed by treating with ACK lysis buffer (Thermo Fisher). The resulting single-cell suspension was counted and approximately 2×10⁶ cells were used for antibody staining. The antibodies used were brilliant ultraviolet (BUV) 395-anti-CD45 (BDB564279, Invitrogen), FITC-Ly-6G Monoclonal Antibody (1A8-Ly6g) (50-112-2179, Invitrogen), phycoerythrin (PE)-cyanine (Cy) 7-anti-Ly6C (BDB560593, Invitrogen), fluorescein isothiocyanate (FITC)-anti-CD103 (11-1031-82, Invitrogen), Pacific Blue-anti-CD11b (RM2828, Invitrogen), PE-Cy7-anti-CD11c (558079, BD Biosciences), and Alexa Fluor (AF) 700-anti-major histocompatibility complex class II (MHCII)- (56-5321-82, Invitrogen). The staining protocols were followed as previously described(55). Neutrophils were identified as CD45⁺Ly6G⁺Ly6C^low, and MPO expression in neutrophils was further stained with Alexa Fluor 647 Mouse Anti-Mouse Myeloperoxidase (BDB570233, Fisher Scientific) after cell fixation using intracellular (IC) fixation buffer (501129058, Fisher Scientific) followed by permeabilization using permeabilization buffer (501129059, Fisher Scientific). Macrophages were identified as CD45⁺CD11c⁺MHCII^var, and dendritic cells were identified as CD45⁺CD11c⁺MHCII^varCD103⁻CD11b⁺. Data were acquired using a five-laser BD FACSymphony A5 and analyzed with FlowJo Software V10. The gating strategy is diagrammed in fig. S11.

Immunohistochemistry

To evaluate the effect of VX-01 on the in vivo treatment of angioinvasion and vascular permeability, we conducted immunohistochemistry on lung sections of neutropenic mice infected with R. delemar and treated with a single dose of 30 μg of VX-01 or human isotype-matched IgG1 given 24 hours post infection. Lungs were harvested on day 4 post infection and snap frozen in Optimal Cutting Temperature (OCT) tissue embedding medium (4585, Fisher Scientific), after which 10-μm thick sections were cut with a cryostat, dried for 1 hour, and fixed with ice-cold methanol(56). The cryosections were rinsed with PBS, blocked with 1% goat serum, and then stained with VEGF rabbit mAb at 10μg/mL (MA5-32038, Fisher Scientific), followed by goat anti-rabbit IgG (H+L) Alexa Fluor 568-labeled secondary antibody (A-11011, Fisher Scientific). The stained slides were mounted by Prolong gold antifade reagent with DAPI and imaged by All-in-One Fluorescence microscopy (BZ-X Series, Keyence). The same image acquisition settings were used to enable comparison of fluorescence intensity among the different samples. The quantification of VEGF was taken from 20 random individual fields from two mice per group through bz-X810 software.

Statistical analysis

Individual-level data are presented in data file S1. The data was collected, graphed, and statistically analyzed using Microsoft Office 360 and Graph Pad 8.0 (GraphPad Software). All results presented in the graphs are the median ± interquartile range. N indicates the number of biological replicates. Data are normally distributed; survival data are consistent with proportional hazards in a Cox regression model. The unpaired, two-sided Student’s t test was used to compare two datasets. The one-way analysis of variance (ANOVA) with Dunnett’s multiple comparisons test was used to analyze the opsonophagocytic killing assay. The one-way ANOVA with Tukey’s multiple comparisons tests was used for tissue fungal burden analyses. The non-parametric Log-rank (Mantel-Cox) test was used to determine differences in mouse survival times. P < 0.05 was considered significant.