PEG-fDAO Reduces Lung Inammation in Chronic Granulomatous Disease Mice Post Challenge with Nonviable Candida Albicans

We previously reported that polyethylene glycol-conjugated recombinant porcine D-amino acid oxidase (PEG-pDAO) could supply reactive oxygen species (ROS) to defective NADPH oxidase in neutrophils of patients with chronic granulomatous disease (CGD), and neutrophils regain bactericidal activity in vitro. In the present study, we employed an in vivo nonviable Candida albicans (nCA)-induced lung inammation model using gp91-phox knockout CGD mice and novel PEG conjugates of Fusarium spp. D-amino acid oxidase (PEG-fDAO), rather than PEG-pDAO. Using three experimentation strategies with the in vivo lung inammation model, the mouse body weight, lung weight, and lung pathology were evaluated to conrm the ecacy of ROS-generating enzyme replacement therapy with PEG-fDAO. The lung weight and pathological ndings were signicantly ameliorated by the administration of PEG-fDAO followed by intraperitoneal injection of D-phenylalanine or D-proline. These data suggest that PEG- fDAO with the function of targeted delivery to the nCA-induced inammation site is applicable in the treatment of inammation in CGD in vivo.


Introduction
Chronic granulomatous disease (CGD) is a primary immunode ciency characterized by the inability of phagocytes to produce reactive oxygen species (ROS) due to a defect in the NADPH oxidase complex. Clinically, CGD patients frequently suffer from bacterial infections and excessive in ammatory disorders, such as CGD-associated bowel in ammation and sterile granulomas. ROS generation is crucial for the phagocytic killing of ingested microbes [1] and in ammasome signaling processes, such as the Caspase-1 activation platform for cytokine production [2][3][4], the Keap1-Nrf2 pathway for antioxidative stress [5], and efferocytosis through phosphatidylserine and its receptor [6].
Anti-in ammatory cytokine therapies (anti-interleukin (IL)-1 antibody, etc.) have been proven to alleviate excessive cytokine production in an in vivo CGD mouse model and clinical studies on severe colitis [4]. In addition, a PPARγ agonist (pioglitazone) has been con rmed to enhance mitochondrial ROS production and partially restore host defense in a CGD mouse model [6]. Bone marrow transplantation or gene therapy with a lentiviral vector and myeloid-speci c promoter has been proposed as a successful curative therapy [7]. Although bone marrow transplantation and gene therapy are thought to be theoretically curative treatments, residual pathogenic components, live or dead, lead to severe vital reactions, such as in ammation and/or organic damage, in CGD patients. Unprecedented antibiotic treatments are pathogen-speci c remedies for the removal of pathogens. Potent pathogen-speci c antibiotic treatments can kill pathogens, but may not prevent the excessive in ammation that can persist due to the presence of residual materials of killed pathogens, such as fungal cell wall b-glucan. Moreover, persistent chronic in ammation in patients with CGD also leads to impairment of hematopoietic stem cell functions [8].
To overcome these limitations, we aimed to develop a novel enzyme replacement therapy with D-amino acid oxidase (DAO) for supplying H2O2 in vivo. DAO is a avo-enzyme that selectively catalyzes the oxidative deamination of D-amino acids, leading to the generation of the corresponding imino acid and H2O2. We previously reported that DAO restores the bactericidal activity of ROS-de cient neutrophils by supplying H2O2 in vitro [9]. The advantages of using fungal DAO are easy availability of this enzyme and more potent selective activity compared with that of mouse or human DAO. [10]. This phenomenon is called the enhanced permeability and retention effect [11].
In this study, a novel polyethylene glycol-conjugated Fusarium spp. DAO (PEG-fDAO) was characterized and evaluated for its in vivo anti-in ammatory activity in a lung in ammation model induced with nonviable Candida albicans (nCA) using gp91-phox knockout CGD mice. Lack of ROS production has been shown to result in high levels of proin ammatory mediators (IL-1b, tumor necrosis factor-α, and keratinocyte chemoattractant) through the in ammasome activation system (dectin-1 receptor) in neutrophils and macrophages [3]. In this model, viable fungi are not always a prerequisite for the development of in ammation in CGD mice [12].
Therefore, using this model, we propose that oxidase replacement therapy with the novel PEG-fDAO could be applicable in the treatment of in ammation in vivo.
Succinimidyl-activated PEG (SUNBRIGHT MEGC-50HS-PEG), with an average molecular size of 5250 Da, was purchased from NOF Corp., Tokyo, Japan. All other reagents were of reagent grade and used without further puri cation. Animals CGD mice (gp91-phox knockout) obtained from Dr. Mary C. Dinauer [13] were backcrossed at least 12 times with C57BL/6 mice to ensure similar genetic backgrounds. Experiments on 8-12-week-old CGD C57BL/6 mice were performed according to the guidelines of the Laboratory Protocol of Animal Handling, Sojo University Faculty of Pharmaceutical Sciences. All animals were housed under speci c pathogenfree conditions. Preparation of nCA nCA was kindly gifted by Dr. Ohno [14]. Brie y, a C-limiting medium was used to grow Candida albicans and cultured at 27 °C with aeration. Viable cells were collected by centrifugation, killed with ethanol, and dried with acetone. The resulting nCA was suspended in phosphate-buffered saline (PBS).

Preparation of PEG-fDAO
PEGylation of fDAO was conducted as described previously [15]. In brief, to the fDAO solution (2.0 mg/mL protein in 0.1 M sodium bicarbonate), succinimide-activated PEG was added at a 3-fold molar excess of PEG to free the amino groups in fDAO and allowed to react for 1 hour at 4 °C. The reaction mixture containing PEG-fDAO was puri ed to remove free PEG and other low-molecular-weight reactants by ultra ltration with a YM-10 membrane (Merck, Darmstadt, Germany) using 0.1 M sodium bicarbonate aqueous solution. PEG-fDAO was stored at -80 °C until further use.

Evaluation of the enzymatic activities of fDAO and PEG-fDAO
The enzymatic activity of fDAO and PEG-fDAO were determined by a horseradish peroxidase-coupled colorimetric assay with o-dianisidine as the substrate. In this assay, the substrate was reduced, and a color developed with a maximal absorption of 460 nm. D-alanine was used as the substrate for fDAO at a nal concentration of 10 mM. The enzymatic reaction was performed at 25 °C in 0.1 M Tris-HCl buffer (pH 8.2), where 1 U of fDAO activity is de ned as the rate of formation of 1 µmol of H2O2 per minute. The maximal rate of activity (Vmax) and Km for each D-amino acid were calculated by curve tting of the non-linear plot of reaction rate versus substrate concentration using the gnu plot software.
Analysis of in vivo pharmacokinetics of fDAO and PEG-fDAO using the plasma For the in vivo pharmacokinetics analysis, fDAO or PEG-fDAO (20 U/mL, 0.1 mL/mouse, n = 3) was injected intravenously into C57BL/6 (BALB/cAjcl) mice. Blood was withdrawn from the medial canthus of the eye using a microhematocrit at 0.5, 4, 24, 48, and 72 hour after fDAO or PEG-fDAO administration. Each blood sample was centrifuged and the plasma was obtained in an ice-cold buffer (100 mM Tris-HCl, pH 8.0) containing a mixture of protease inhibitors (1mM PMSF, 10 µg/mL leupeptin, and 2.5 mM EDTA).
The DAO activity in the plasma was determined based on the formation of pyruvic acid during the reaction between D-Alanine and DAO, as described previously [15].
Induction of lung in ammation by nCA aspiration Mice were anesthetized by intraperitoneal injection with 200 mg/kg of 2,2,2-tribromoethanol (Sigma-Aldrich), as described by Dr. Aratani. CGD mice were challenged with intranasal administration of 10 7 nCA cells in a volume of 30 μL of PBS, and control mice were administered 30 μL of PBS alone [12]. Since we had no information on the time course of in vivo production of H2O2 induced by PEG-fDAO with Damino acids in the nCA-induced lung in ammation model, a short, medium, and long treatment schedule was explored.

Experiment-1 (Exp-1)
Mice were divided into two groups. The rst group was administered nCA intranasally as the positive control (n = 5). The second group was administered nCA intranasally on day 0 of experiment and PEG-fDAO was injected through the tail vein on the second day, followed by intraperitoneal administration of D-phenylalanine (0.1 M, 0.5 mL/mouse) on the third, fourth, and fth days (n = 5). Mice were sacri ced on the eighth day.

Experiment-2 (Exp-2)
Mice were divided into three groups. The rst group was administered PBS intranasally as the negative control (n = 3). The second group was administered nCA intranasally as the positive control (n = 3). The third group was administered nCA on day 0 and injected with PEG-fDAO (10 U/mL, 0.1 mL/mouse) (n = 3) through the tail vein on the fourth day, followed by intraperitoneal injection of D-phenylalanine (0.1 M, 0.5 mL/mouse) on the fth, sixth, and seventh days instead of D-proline. Rhodamine-labeled bovine serum albumin (rhodamine-BSA) (10 mg/kg) was injected one day before sacri ce on the 14 th day. Fluorescent imaging of the excised lungs was conducted using the IVIS Lumina XR (excitation: 555-585 emission: 695-770 nm; PerkinElmer Japan Co., Ltd., kanagawa, Japan).

Experiment-3 (Exp-3)
Mice were divided into three groups. The rst group was administered PBS intranasally as the negative control (n = 3). The second group was administered nCA intranasally as the positive control (n = 3). The third group was administered nCA on day 0 and injected with PEG-fDAO (10 U/mL, 0.1 mL/mouse) (n = 3) on the ninth day, followed by intraperitoneal injection of D-proline (1 M, 0.5 mL/mouse) on the 10 th , 11 th , and 12 th days. Rhodamine-BSA (10 mg/kg) was injected one day before sacri ce on the 21 st day. Fluorescent imaging of the excised lungs was performed using the IVIS Lumina XR (PerkinElmer Japan Co., Ltd.).

Analysis of the lung pathology
Mice were sacri ced on the eighth day in Exp-1, 14 th day in Exp-2, or 21 st day in Exp-3 after nCA administration. The lungs were removed and xed in 10% buffered formalin. For light microscopy, tissues were xed overnight, dehydrated in graded ethanol solutions, embedded in para n, sectioned at 2-μm thickness, and stained with hematoxylin and eosin using standard protocols.

Statistical analysis
Statistical analysis of the obtained data was performed using a two-tailed unpaired Student's t-test. Differences were considered statistically signi cant at p < 0.05.

Physicochemical characterization of fDAO and PEG-fDAO
Characterization of PEG-fDAO is summarized as follows. The reaction between fDAO and succinimide-PEG resulted in ve chain numbers of PEG conjugates on fDAO, which was determined by quanti cation of the primary amine. The fDAO formed a tetramer (165 kDa) in the physiological solution, and the molecular weight of PEG-fDAO was calculated to be 206 kDa. An increase in the molecular size of fDAO by PEGylation was also con rmed by SDS-PAGE and size exclusion chromatography (data not shown).
Enzyme-speci c activity of fDAO was 26 ± 1.3 units/mg protein, and its enzyme activity was not altered by PEGylation (PEG-fDAO: 26 ± 0.53 units/mg protein). The Vmax and Km values of fDAO for each Dtype amino acid were almost similar with those of PEG-fDAO. Vmax and Km values of fDAO and PEG-fDAO towards neutral D-amino acids also showed similar trends at the optimal pH of 8.2, and an in ammatory environmental pH of 6.5 (Supplemental Table 1). The enzymatic activity of PEG-fDAO at pH 6.5 was lower than that at pH 8.2, for all substrates tested. Similar substrate speci city was observed at pH 6.5 and 8.2.
D-phenylalanine showed the highest Vmax (Vmax = 18.6 µmol H2O2/min at pH 6.5) and was active even at low substrate concentrations (Km = 0.22 mM at pH 6.5) among the D-amino acids tested. We employed D-proline in Exp-3 to compare the in vitro bactericidal activity 1 with the in vivo antiin ammatory activity (Vmax = 6.4, Km = 5.2 at pH 6.5). In Exp-1 and Exp-2, D-phenylalanine was selected as the substrate to con rm PEG-fDAO activity in vivo.

In vivo pharmacokinetics of fDAO and PEG-fDAO using the plasma
We employed a methodology to exclude denatured DAO proteins with lost enzymatic activity for measuring the enzymatic activity of DAO in vivo as in Figure 1. The apparent half-life (t1/2) of fDAO (n=3) and PEG-fDAO (n=7) are statistically 7.69±1.36, hour, 33.68 ±3.10 hour, respectively, and T-Test with un-paired samples shows that p-value is 0.000000415. Since the t1/2 of PEG-fDAO appeared to be long enough, we employed intraperitoneal injection of D-amino acid on three consecutive days post PEG-fDAO administration.

In vivo experimental ndings
In Exp-1 with an 8-day schedule, there was a marked mean body weight loss in the nCA group (-6.18 ± 2.28 g) compared with that in the PEG-fDAO experimental group (-4.14 ± 2.74 g), but the difference was not signi cant (p = 0.236, n = 5) (Figure 2A and 2B). The mean lung weight in the nCA group (0.562 ± 0.04 g) was signi cantly higher than that in the PEG-fDAO experimental group (0.314 ± 0.095 g) (p = 0.001, n = 5) ( Figure 3A). Images of the excised lungs are shown in Supplemental Figure 1. Rhodaminelabeled uorescent imaging was not performed in this experiment.
In Exp-2 with a 14-day schedule and D-phenylalanine substrate, the mean body weight loss in the nCA group (-1.6 ± 0.7 g) was lower than that in the PEG-fDAO group (-2.5 ± 2.8 g) (p = 0.618, n = 3) ( Figure 2D and 2E). The mean lung weight in the nCA group (0.283 ± 0.025 g) was signi cantly higher than that in the PEG-fDAO group (0.236 ± 0.005 g) (p = 0.035, n = 3) ( Figure 3B 2/3). Furthermore, the mean lung weight in the control group was 0.24 ± 0.01 g, which was almost the same as that in the PEG-fDAO group (p = 0.64, n = 3) ( Figure 3B 1/3). To determine the in ammatory region in the lung, rhodamine-BSA was intravenously injected, and accumulated rhodamine-BSA in the lung was quanti ed by measuring the uorescence intensity. Rhodamine-labeled uorescence intensity in the excised lungs of mice from the nCA group was brighter and the tissue was more swollen than that observed for the mice in the PEG-fDAO and control groups ( Figure 4A). The pathological ndings were different in the nCA group; pneumonia and focal accumulation of neutrophils reacting to and phagocytizing eosinophilic laments were observed in the nCA group ( Figure 5C/D), but these were milder in the lung tissues from mice in the PEG-fDAO group ( Figure 5E/F).
In Exp-3 with a 21-day schedule, there was no marked difference in the mean body weight loss in the nCA group (-0.33 ± 0.68 g) compared with that in the PEG-fDAO group (-0.17 ± 0.87 g) (p = 0.66, n = 3) ( Figure  2G and 2F). The mean lung weight in the nCA group (0.334 ± 0.065 g) was higher than that in the PEG-fDAO group (0.230 ± 0.021 g), but the difference was not signi cant (p = 0.058, n = 3) ( Figure 3C). In the control group, the mean lung weight was 0.156 ± 0.020 g. Rhodamine-labeled uorescence intensity in the excised lungs from mice in the nCA group was brighter and the tissue was more swollen than that noted for mice in the PEG-fDAO and control groups ( Figure 4B). The pathological ndings were particularly different in the nCA group, showing pneumonia with accumulation of eosinophilic lamentphagocytizing neutrophils, accompanied by granulomatous accumulation of histiocytic cells and lymphoid cells. These pathological changes were apparently mild, with the eosinophilic laments fragmented and smaller in the tissues from mice in the PEG-fDAO group ( Figure 5K/L) compared with that found in mice from the nCA group ( Figure 5I/J). Unexpectedly, a small amount of granuloma formation with giant cells was observed in the PEG-fDAO group.

Discussion
In the present study, our pharmacokinetics analysis revealed that fDAO (165 kDa) was rapidly cleared from the circulation after intravenous infusion (t1/2 = 7.67 ± 1.36 hour (n = 3)), but PEG-fDAO (206 kDa) was rather stable (t1/2 = 33.68 ± 3.10 hour (n = 7)) ( Fig. 1). Based on these ndings, we employed PEG-fDAO administration, followed by administration of D-phenylalanine or D-proline for three days in the nCA-induced lung in ammation model to evaluate the anti-in ammatory effect of PEG-fDAO. According to studies on Aspergillus hyphae- [16] and sterile nCA- [12] induced lung in ammation models, sterile fungal cell wall (branched (β-1,3) (β-1,6) glucan)-induced in ammation model with gp91 −/− CGD mice [17] and zymosan (the same fungal cell wall-derived product)-induced lung in ammation model with p47 −/− [18] and p91 −/− [19] CGD mice, the MAPK signaling pathway (ERK1/2 and NF-κB) involving toll-like receptor 2, dectin-1 receptor and cytokines (IL-1b, tumor necrosis factor-α, IL-17, keratinocyte chemoattractant, and granulocyte colony-stimulating factor) and leukotriene B4 [19] are activated on the second day or earlier post stimulation. The net balance of cytokine and leukotriene B4 expression in the lungs of CGD mice in response to the fungal cell wall component is prone to be in a proin ammatory state, which could promote a continued in ammatory response [16,19]. Based on these previous ndings, we administered PEG-fDAO on the second day, followed by an intraperitoneal injection of D-phenylalanine for three days in Exp-1. Endo et al [12] reported that nCA-treated CGD mice show prominent airway accumulation of in ammatory cells at 3 and 6days post nCA administration.
Similarly, the nCA-treated CGD mice in Exp-1 showed a prominent loss of body weight and gain of the lung weight. In contrast, nCA-treated CGD mice injected with PEG-fDAO and D-phenylalanine showed a signi cantly lower loss of body weight and gain of the lung weight (Figs. 2A/B and Fig. 3A) compared with that observed in the control group. These differences were also apparent from a physiological point of view.
Next, to determine the effect of PEG-fDAO over a longer period, as in the Aspergillus hyphae-induced lung in ammation model [16], PEG-fDAO was administered on the fourth day in Exp-2 with a 14-day schedule, and the ninth day in Exp-3 with a 21-day schedule. We found the body and lung weight to tend to recover over the experimental time course. In Exp-2, the loss of body weight in the nCA group was approximately the same as that in the PEG-fDAO group (p = 0.62) (Fig. 2D/E), but the lung weight gain in the nCA group was signi cantly higher (p = 0.035) than that in the PEG-fDAO group (Fig. 3B). Furthermore, the lung pathology in the nCA group appeared to be stronger than that in the PEG-fDAO group (Fig. 5C/D vs 5E/F). In Exp-3, the loss of body weight in the nCA group was higher than that in the PEG-fDAO group, but the difference was not signi cant (p = 0.66) (Fig. 2G/H). The lung weight gain in the nCA group was not signi cantly high (p = 0.058), but it was higher than that in the PEG-fDAO group ( Fig. 3C; 2 vs 3). Although PEG-fDAO seemed effective in ameliorating the physiological changes induced by nCA stimulation, we could not con rm if this effect was signi cant through these evaluations.
Consistent with a previous study by Dinauer [16], which had a 21-day experimental schedule with intratracheal administration of Aspergillus fumigatus hyphae to gp91 −/− mice, many granulomatous formations in our Exp-3 were observed in the pathological lung specimens from mice in the nCA group ( Fig. 5I/J) but a few in the those from mice in the PEG-fDAO group (Fig. 5K/L). Granulomatous formations were not apparent in Exp-2, but a few macrophage accumulations around the eosinophilic Candida body were found in the pathological lung specimens from mice in the nCA group (Fig. 5C/D and not in those from mice in the PEG-fDAO group (Fig. 5E/F). In addition, the lung pathology in the nCA group ( Fig. 5I/J) was stronger than that observed in the PEG-fDAO group in Exp-3( Fig. 5K/L). These pathological ndings also support the idea that PEG-fDAO administration, followed by D-amino acid treatment, would ameliorate the pathological changes induced by nCA stimulation.
According to the characterization of PEG-fDAO, the Km value of D-phenylalanine was approximately 24-40 times lower than that of D-proline, but the blood concentration of D-phenylalanine or D-proline was supposed to be 2-3 times higher than the Km value of both D-amino acids. In addition, the Vmax of PEG-fDAO for D-phenylalanine was approximately three times higher than that for D-proline (Supplemental Table 1). Since D-phenylalanine has lower solubility than D-proline in aqueous media, bolus injectable Dphenylalanine was 10-fold lower than D-proline (D-phenylalanine: 0.1 M, 0.5 mL/mouse (Exp-1, 2) and Dproline: 1 M, 0.5 mL/mouse (Exp-3). Therefore, although PEG-fDAO activity was higher towards Dphenyalanine than D-proline, a low dose of D-phenylalanine might result in a therapeutic e cacy comparable to that of D-proline in alleviating lung in ammation. However, more stable conditions and protocols are necessary to con rm the apparent effects of PEG-fDAO observed in this model and develop clinical applications.
Therefore, in these series of experiments, PEG-fDAO administration followed by D-amino acid injection was apparently useful for the treatment of nCA-induced lung in ammation, especially in Exp-1, but also in Exp-2 and Exp-3. Considering the mechanism of the nCA-induced pneumonia model, H2O2 seems to be involved in many in ammatory processes. As Endo et al [12] reported in their 3-day pneumonia model, a low amount of MAPKs in CGD mice may lead to prolonged phosphorylation of ERK1/2. Segal et al [18] showed impaired Nrf2 activity and increased NF-κB activation in zymosan-treated mononuclear cells from X-linked CGD patients. Furthermore, many cytokines are thought to be involved in granuloma formation with giant cells [20] wherein H2O2 is involved. As neutrophil-derived ROS, the targeted delivery of PEG-fDAO could also restore the anti-in ammatory responses by supplying H2O2 to the site of in ammation. Setoguchi et al [21] also reported the therapeutic value of pulmonary granuloma by controlling nitric oxide and ROS.
In conclusion, we prepared PEG-fDAO, characterized its enzymatic properties, and evaluated its antiin ammatory effect on gp91-phox knockout CGD mice. The features of PEG-fDAO were comparable to those of pDAO in terms of Vmax and substrate speci city, but PEG-fDAO was more stable than PEG-pDAO. Moreover, PEG-fDAO exhibited anti-in ammatory effects in CGD mice with nCA-induced lung in ammation. Nonetheless, a protocol with a more optimized regimen will be necessary to con rm the anti-in ammatory effects of PEG-fDAO.
Declarations evaluation. Hiroyuki Nunoi, Taiki Nisimura and Makoto Matsukura analyzed the data and wrote the rst draft of the manuscript. All authors commented on the manuscript and approved it.

Ethics approval
Experiments on 8-12-week-old CGD C57BL/6 mice were performed according to the guidelines of the Laboratory Protocol of Animal Handling, Sojo University Faculty of Pharmaceutical Sciences. All animals were housed under speci c pathogen-free conditions.