Zebrafish (Danio rerio) were maintained at the University of Milan, Via Celoria 26 – 20133 Milan, Italy (Aut. Prot. n. 295/2012-A – December 20, 2012). Zebrafish strains AB, and Tg(mpeg1:mcherry) (Ellett et al., 2011), were maintained according to international (EU Directive 2010/63/EU) and national guidelines (Italian decree No. 26 of the 4th of March 2014). Embryos were collected by natural spawning, staged according to Kimmel et al. (1995) and raised at 28°C in E3 fish growth medium (Instant Ocean, 0,1% Methylene Blue) in Petri dishes, according to established techniques. After 24 hours post fertilization (hpf), 0,003% 1-phenyl-2-thiourea (PTU, Sigma-Aldrich, Saint Louis, MO, United States) was added to the fish water to prevent pigmentation. Embryos were washed, dechorionated and anaesthetized with 0.016% tricaine (Ethyl 3-aminobenzoate methanesulfonate salt; Sigma-Aldrich), before observations, microinjection and picture acquisitions.

Pseudomonas aeruginosa PAO1 strain (Stover et al., 2000) and its derivatives PAO1 217-2a, namely a spontaneous mutant resistant to CKΦ, were used throughout this work. PAO1-GFP and PAO1-217-2a-GFP carry a plasmid expressing the GFP and conferring carbenicillin resistance (Forti et al., 2018). Bacterial cultures were freshly streaked out from glycerol stocks and were grown with shaking at 37°C to exponential phase [OD₆₀₀ = 0.5, corresponding to about 5 × 10⁸ colony forming units (CFU)/ml] in Lennox lysogeny broth (LB, 1% tryptone, 0,5% yeast extract and 0.5% NaCl). The culture was pelleted, resuspended in half the volume of physiological solution and stored at 4°C for up to 20 h. Before use, to avoid clumping, bacterial suspensions were vortexed, then homogenized through a 25-gauge needle and resuspended at about 5 × 10⁸ bacteria/ml in physiological solution added with 10% phenol red to aid visualization of the injection process.

The four virulent phages able to infect P. aeruginosa were isolated and characterized previously (Forti et al., 2018). The phages belong to Caudovirales order and in particular two are Schitoviridae (formerly classified as Podoviridae), PYO2 (GenBank accession numbers vB_PaeP_PYO2, MF490236) and DEV (vB_PaeP_DEV, MF490238), and two Myoviridae, E215 (vB_PaeM_E215, MF490241), and E217 (vB_PaeM_E217, MF490240). The phage preparations were grown and purified as described (Forti et al., 2018). Briefly, high-titer phage lysates of PAO1 cultures were filtrated with 1.2 μm diameter filters and incubated with DNase (1 μg/ml) and RNase (1 μg/ml); then, treated lysates were PEG-precipitated, purified by cesium chloride ultracentrifugation and dialyzed against TN buffer (10 mM Tris-HCl, 150 mM NaCl, pH 7). Finally, phage preparations were passed through endotoxin removal columns (EndoTrap HD; Hyglos, Germany). The levels of residual endotoxins in the phage preparations were below the limit value recommended for intravenous administration (5 EU/kg/h by European Pharmacopoeia, 1997) as assessed by measuring the endotoxin level with the Pierce LAL Chromogenic Endotoxin Quantification assay (Thermo Fisher Scientific, Waltham, MA, United States). The CKΦ was assembled immediately before each experiment by mixing equivalent volumes of the four phage preparations at the same titer (CKΦ titer, 5 × 10⁸ plaque forming units (PFU)/embryo/ml).

Injection of oligo-antisense morpholino were carried out on 1- to 2-cell stage embryos. Morpholinos were diluted in 1× Danieau buffer [58 mM NaCl, 0.7 mM KCl, 0.4 mM MgSO₄, 0.6 mM Ca(NO₃)₂, 5.0 mM HEPES (pH 7.6)] and the dye tracer rhodamine dextran was co-injected when necessary to allow visualization. cftr mRNA translation repression was achieved by co-injecting 0.25 pmole/embryo of each cftr-ATG-MO and cftr-splice-MO (Gene Tools LLC, Philomath, OR, United States), as previously described (Phennicie et al., 2010). A standard control morpholino oligonucleotide with no target in zebrafish (Gene Tools LLC) was injected as negative control.

Apoptotic body like liposomes (ABLs) were produced as previously described (Poerio et al., 2017). Briefly, the inner monolayer lipids composed by: L-α-phosphatidic acid (PA, Avanti Polar Lipids) were suspended in anhydrous dodecane (Sigma) at the concentration of 0.05 mg/ml. L-α-phosphatidylserine (PS) (Avanti Polar Lipids, Alabaster, AL, United States) was used as outer monolayer lipid and was added to a 99:1 dodecane:silicone solution to obtain a final concentration of 0.05 mg/ml. Asymmetric liposomes were prepared by adding 2 ml of outer monolayer lipid suspension over 3 ml of saline. Finally, 100 μl of the inner monolayer lipid suspensions were added over 2 ml lipid phase and the samples were centrifuged at 120 g for 10 mins. After the centrifugation, ABLs were collected in the aqueous phase using a 5 ml syringe with a 16-gauge stainless steel needle, in order to produce PS outside/PA inside liposomes (PS/PA). Liposomes were then quantified by a flow cytometer FACSCalibur (Becton Dickinson, Franklin Lakes, NJ, United States).

For prophylactic and therapeutic treatments, 1–2 dpf embryos were microinjected with 2 nl of PAO1 strains (PAO1, PAO1-GFP or PAO1 217-2a-GFP; approximatively 100–300 CFU/embryo) systemically, and/or 2 nl of PS/PA liposome suspension and/or 2 nl of CKΦ (approximately 500 PFU/embryo; 5 × 10⁸ PFU/ml) into the duct of Cuvier, as described in Clatworthy et al. (2009), or locally into hindbrain ventricle (Benard et al., 2012). To titre the injected phages, drops of 2 nl of phage suspension were diluted in TN buffer and measured by agar overlay method (Blair, 1959) to determine the PFU number. To titre the injected bacteria (CFU/embryo), drops of 2 nl of PAO1 suspension were diluted in physiological solution and plated. The titre of the injected phages/embryo was extrapolated from the average of five independent measures. The evaluation of bacterial infection was performed following the guidelines of Takaki et al. (2013). PAO1-infected embryos were raised at 32°C for the entire duration of the experiment.

To measure bacterial burden related to PAO1 infection, embryos at 8 or 20 hpi (hours post-infection) were thoroughly washed in sterile physiological solution and then anesthetized in sterile tricaine solution. At least two groups of 5–10 infected embryos for each treatment were mechanically homogenized in 1% Triton X-100 (Thermo Fisher Scientific, Waltham, MA, United States) in PBS by means of an insulin syringe (with a 25-gauge needle). The resulting homogenates were serially diluted and plated on LB agar. Ampicillin (100 μg/ml) was added to LB medium to select for the amp-resistant PAO1 strain. This procedure was necessary to avoid the alteration in CFU counterselection, by limiting the eventual growth of endogenous bacteria of embryos (or in embryo medium) selecting only amp-PAO1 resistant strains. Plates were incubated at 37°C for 16–20 h. Then, colonies were counted (under fluorescence microscope when needed) and corresponding bacterial titers were calculated as the mean of the titers obtained for the two groups of homogenized embryos. The average CFU per embryo was extrapolated by dividing the obtained bacterial titer by the number of embryos in one group. Corresponding bright field (BF) and fluorescence images of colonies on plate were acquired. The mean and SEM of at least three independent experiments were reported on graphs and for each experiment values were normalized on ctrl embryos.

Infected embryos were anesthetized in Tricaine and bright-field and fluorescence images were sequentially acquired using an epifluorescence stereomicroscope (M205FA, Leica, Wetzlar, Germany) equipped with fluorescent lamp and a digital camera, and mounting GFP-filter (excitation of 488 nm).

For quantification of bacterial load by Fluorescent Pixel Counts (FPC), fluorescence signal was measured as described (Phan et al., 2018), by computation using Fiji (ImageJ software, developer: Wayne Rasband) as following: (1) background of image was subtracted, (2)“make binary” function was run, (3) “measure area” function was used to determine the number of fluorescent pixels, with avoiding the auto-fluorescence of the yolk. The mean values of two independent experiments with 3–10 embryos/treatment were used.

For this essays zebrafish Tg(mpeg1:mcherry) line was used (Ellett et al., 2011). To assess macrophage activation, prophylactic treatment with PS/PA liposome and infection with PAO1-GFP were performed locally, through microinjection in the close cavity of the hindbrain ventricle, respectively, at 28 and 48 hpf, as previously described (Davis et al., 2002). Injected embryos were incubated in fresh PTU at 32°C, and at 6 hpi BF and fluorescence images were sequentially acquired using a epifluorescence stereomicroscope (M205FA, Leica, Wetzlar, Germany) equipped with fluorescent lamp and a digital camera, and mounting GFP-filter (excitation of 488 nm) and mcherry-filter (excitation of 587 nm). For phagocytosis observation, live confocal images were acquired. Embryos were mounted in 35 mm glass-bottom dishes in 1,5% low-melting point agarose (Sigma) in fish water containing Tricaine, and immediately imaged on an SP2 confocal inverted microscope (Leica, Wetzlar, Germany) with an HC PL APO 10× objective and 488 nm argon laser for GFP and 587 nm for mcherry acquisition. Series of typically spanning 100–120 μm z-stacks at 1–2 μm intervals were acquired. Guidelines to visualize host-bacterial interaction were followed (Milligan-Myhre et al., 2011). Images were processed using the Adobe software. Macrophage recruitment in brain ventricle were measured by counting mcherry-positive cells in a defined region of interest by computation using of Fiji, using “Find maxima” function, as described in Ellett and Lieschke (2012). Phagocytic activity were quantified by measuring FPC related to PAO-GFP and mpeg1:mcherry signals co-localization by computation using of Fiji as following: (1) different channels of images were merged and brightness/contrast were adjusted for better visualization; (2) a “color threshold” was set; (3) “measure area” function was used to determine the number of the overlapped fluorescent pixels of the image.

To induce macrophage ablation, 2 nl of Lipo-Clodronate (clodronateliposomes.com) (Van Rooijen et al., 1996) were systemically injected into the duct of Cuvier of 1 dpf Tg(mpeg1:mcherry) embryos as previously described (Bernut et al., 2014). Lipo-PBS was injected as control. 24 h post-treatment, macrophage-depleted embryos were selected based on the reduction of red macrophages.

The human CF bronchial epithelial cell line CuFi-1 (ATCC CRL-4013™) were grown as monolayer cultures as previously described (Loberto et al., 2014). Cells were seeded in 96 well plates pre-coated with human placental collagen (8 × 10³ cells/well) 24 h before PS/PA liposomes and CKΦ administration.

The human leukemia monocytic cell line THP-1 (ATCC TIB-202™) were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 1 mM sodium pyruvate, 0.05 mM 2-mercaptoethanol and differentiated into a macrophage-like phenotype using phorbol-12-myristate-13-acetate (PMA) according to Daigneault et al. (2010) with minor modifications. Briefly, THP-1 cells (1 × 10⁵ cells/well in 96 well plates) were exposed to 50 nM PMA for 72 h followed by PMA-free medium replacement for 24 h before PS/PA liposome and CKΦ administration.

Both cell lines were incubated for 48 h with PS/PA liposomes at a 1:1 liposome/cell ratio and CKΦ (5 × 10⁸ PFU/ml) either as single treatments or in combination.

In order to inhibit the CFTR channel, differentiated THP-1 cells were incubated in presence of 10 μM CFTR_inh-172 [(3-trifluoromethyl)phenyl]-5-[(4-carboxyphenyl)methylene]-2-thioxo-4-thiazolidinone) 30 min before and during PS/PA liposomes and CKΦ administration. THP-1 not treated with CFTR inhibitor were used as control.

At the end of incubations, cell viability was evaluated by MTT assay (Sigma M2128) according to the manufacturer’s protocol. As controls, cells were treated with the vehicle alone in the same experimental conditions (i.e., PBS for PS/PA liposomes and TN buffer for CKΦ).

Reverse transcription-PCR and real-time quantitative-PCR (RT-qPCR) assays were carried out to detect the mRNA expression levels of inflammatory genes il-10, il-13, il-1β and TNF-α (Cafora et al., 2019; Ferrari et al., 2019; Bottiglione et al., 2020). Total RNA was extracted from zebrafish embryos using Trizol reagent (Life Technologies, Carlsbad, CA, United States) according to the producer’s instructions. Concentration and purity of RNA were measured using the Nanodrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States). To avoid possible genomic contamination, RNA was treated with DNase I RNase-free (Roche Diagnostics, Basel, Switzerland). 1 μg of DNase-treated RNA was reverse-transcribed by means of the “ImProm-II™ Reverse Transcription System” (Promega, Madison, WI, United States), using a mixture of random primers and oligo(dT), following the manufacturer’s protocol. qPCRs were performed in a total volume of 20 μl containing 1X iQ SYBR Green Super Mix (Promega), using proper amount of synthesized cDNA. qPCRs were performed using the BioRad iQ5 Real Time Detection System (Biorad, Hercules, CA, United States) following the manufacturer’s guidelines. Thermocycling conditions were: 95°C for 10 min, 95°C for 10 s, and 55°C for 30 s. All reactions were performed at least in triplicate for 40 cycles. Primers used for mRNA expression analysis are listed in Supplementary Table 1. The relative expression level of each gene was calculated according to the 2^–ΔΔCt method (Livak and Schmittgen, 2001). For normalization purposes, rpl8 and beta-actin were used as internal reference genes. The mean and SEM of at least three independent experiments were reported on graphs.

Statistical analyses were generated using GraphPad Prism software version 8.0.2 for Windows (La Jolla, CA, United States). The Gaussian data distribution of all datasets was guaranteed by Shapiro-Wilk normality test or Kolmogorov-Smirnov normality test. Data resulted as outliers were excluded from analysis. Specific statistical tests were used to evaluate the significance of differences between groups, as indicated in the relative figure legend: unpaired two-tailed Student’s t-test when comparing two groups or ordinary one-way ANOVA followed by post hoc Tukey’s correction for multiple comparisons. Data represent results of at least two/three independent experiments for zebrafish and five for cells, and mean ± SEM or mean with min to max values were reported in graphs. P-value < 0.05 was considered to indicate statistically significant differences.

The possible toxic effect of PS/PA liposomes was firstly evaluated by systemic injection. 4 nl of PS/PA liposomes suspension was injected in the duct of Cuvier at the 28 hpf stage, when blood circulation begins. As a control, the embryos were injected with the same amount of physiological solution. The microinjected embryos were grown at 28.5°C in fish water. At 24 hpi, we did not observe lethality and morphological alterations, confirming that PS/PA liposomes are not toxic when systemically injected in zebrafish model (data not shown).

Then, we analyzed whether a prophylactic treatment with asymmetric PS/PA liposomes could elicit antimicrobial activity following an infection with PAO1. Zebrafish embryos, both wild-type (WT) and cftr-LOF, were systemically injected with PS/PA liposomes at 28 hpf and, 20 h later, with PAO1-GFP strain to generate a systemic infection. To evaluate the effectiveness of the prophylactic treatment with PS/PA liposomes in counteracting bacterial infection, the bacterial burden (CFU/embryo) of PAO1 infected embryos were tested at 8 hpi (Figure 1A, same results were obtained with analysis at 20 hpi, data not shown). The prophylactic treatment with PS/PA liposomes systemically injected, significantly reduced the bacterial burden both in the WT and cftr-LOF embryos, although to a lesser extent in these latter (Figure 1B), in line with the reduced antimicrobial response of the cftr-LOF embryos (Phennicie et al., 2010). To directly follow the infection in vivo (Figure 1C and Supplementary Figure 1A) and to distinguish and count the Pa colonies from the non-GFP colonies formed by zebrafish endogenous bacteria, PAO1-GFP bacterial were used (Supplementary Figure 1B). The GFP plasmid is maintained and does not interfere with the growth and survival of PAO1 in zebrafish, at least at 8 hpi (Supplementary Figure 1C).

In parallel, asymmetric PS/PA liposomes were delivered through the injection in the close cavity of the hindbrain ventricle at 28 hpf, followed by local PAO1-GFP infection 20 h later (Figures 1D–F and Supplementary Figure 1A). Bacterial burden at 8 hpi, measured as fluorescent pixel count (FPC) of PAO1-GFP in a determined area, was reduced both in wild-type and cftr-LOF embryos pre-treated with PS/PA liposomes in comparison to control embryos injected with physiological solution (Figures 1D,E). As for systemic infection, also with local infection the cftr-LOF embryos showed a reduced effect in antimicrobial activity exerted by PS/PA liposomes treatment in comparison to WT embryos.

To test whether the antimicrobial effects of PS/PA liposomes administration are due to an increased activation of the host immune system, we analyzed macrophage activation. Indeed, it has been already described that PS/PA liposomes are able to enhance intracellular bacterial killing in macrophages by inducing both phagosomal acidification and reactive oxygen species (ROS) production (Greco et al., 2012; Poerio et al., 2017). Local injection of PS/PA liposomes in the hindbrain ventricle followed by PAO1-GFP injection was performed (Figure 2A) and macrophage migration toward the inflamed site was assessed in the macrophage reporter line Tg(mpeg1:mcherry). As control, we included also uninjected embryos to assess if the mechanical stimulus of the microinjection needle might activate macrophages (data not shown). Comparison was performed between them and embryos pre-treated with physiological solution (ctrl) or PS/PA liposomes, infected with PAO1-GFP. Macrophage recruitment was not significantly increased in the ventricle of WT embryos treated with PS/PA liposomes in comparison to control, while significant macrophage recruitment was observed in the cftr-LOF embryos stimulated with PS/PA liposomes (Figures 2B–D).

In parallel with the increased recruitment toward the infection site, PS/PA liposome administration led to an improved phagocytic activity of macrophages. Indeed, analysis by confocal images of the hindbrain ventricle of both WT and cftr-LOF Tg(mpeg1:mcherry) embryos, showed that green PAO1-GFP bacteria signal that co-localize with red macrophages signal was significantly increased in PS/PA pre-treated embryos in comparison to controls (Figures 2E–G). Further evidence of the enhanced phagocytic activity of macrophages stimulated by PS/PA liposomes, was the decreased levels of pro-inflammatory cytokines IL-1 beta and TNF-alpha and increased expression of anti-inflammatory cytokines IL-10 and IL-13 in comparison to controls (Figures 2H,I). Interestingly, this increase was higher in cftr-LOF than in wild-type embryos, suggesting that the PS/PA liposomes-mediated macrophages activation is more efficient when the host immune system is impaired as previously demonstrated (Phennicie et al., 2010). To demonstrate that the antimicrobial activity of PS/PA liposomes is elicited specifically through macrophages activation, we chemically depleted macrophages by means of Lipo-clodronate injection (Bernut et al., 2014; Supplementary Figure 2). Pa bacterial burden did not vary significantly in macrophages-depleted embryos treated with PS/PA liposomes and control embryos injected with physiological solution (Figure 2J).

We then verify if a combination of PS/PA liposomes and phages might improve the effects of a single treatment. Embryos were pre-treated with PS/PA liposomes, after 20 h systemically infected with PAO1 and, 3 h later, injected with a phage cocktail (CKΦ) able to counteract Pa infection (Forti et al., 2018; Figure 3A). To assess the efficacy of combination therapy, the bacterial burden of the embryos (CFU/embryo) was measured at 8 hpi. In WT embryos, presenting a natural immune response to bacteria, surprisingly PS/PA liposomes-activated macrophages resulted more efficient than CKΦ–treatment in counteracting PAO1 infection, and the combination therapy did not enhance the result obtained with single PS/PA liposome administration (Figure 3B). On the contrary, in the cftr-LOF embryos the antimicrobial activity of the single treatments (i.e., PS/PA lipo or CKΦ) was comparable but they act synergistically when combined, significantly reducing the bacterial load (Figure 3C). Similar results of bacterial burden were observed at 20 hpi (data not shown).

To assess whether PS/PA liposomes elicited an efficient anti-microbial activity also in a therapeutic setting, a situation that is more plausible in a clinical condition, embryos were firstly infected with PAO1 and then injected with PS/PA liposomes and CKΦ (Figure 4A). The progression of the infection was evaluated directly through PAO1-GFP bacteria imaging (Supplementary Figure 3) and by CFU/embryo count at 8 hpi (Figures 4B,C and Supplementary Figure 3). Although still reducing the bacterial burden of infected embryos, the therapeutic treatment with PS/PA liposomes was less efficient than the prophylactic one. This effect was expected, probably due to the reduced time needed for a full antimicrobial macrophage activation following Pa infection. Accordingly, the antimicrobial action of PS/PA liposomes is even less efficient in cftr-LOF embryos compared to WT. On the contrary, CKΦ administration significantly reduced to the same extent the bacterial burden of both WT and cftr-LOF embryos. When combined, PS/PA liposomes and CKΦ therapeutic administrations further reduced the bacterial load only in WT embryos, showing no differences compared to single CKΦ treatment in cftr-LOF embryos (Figures 4B,C).

We then tested the effect of the combination therapy on CKΦ-resistant PAO1. We infected WT and cftr-LOF embryos with a mixed culture composed 1:1 by PAO1 and a spontaneous PAO1 mutant resistant to CKΦ (strain PAO1-217-2a; Supplementary Figure 4). To distinguish between CKΦ-sensitive and CKΦ-resistant PAO1, the resistant strain carried the GFP plasmid (PAO1-217-2a-GFP).

Single CKΦ or PS/PA administration had similar effect and reduced the bacterial burden to between 40 and 50% in WT and cftr-LOF embryos. In both cases, the combination therapy was more effective than the single treatments (Figures 4D,F). We also found that the percentage of resistant bacteria (forming green fluorescent colonies) recovered from infected embryos 8 hpi was around the 10%, much lower than the 50% proportion in the cultures inoculated at time 0. This suggests that the mutant strain may have reduced survival in the infected host compared to its parental PAO1 strain. After single CKΦ treatment, significantly higher percentages of resistant-CFU (23% in WT, 33% in cftr-LOF) were found, as expected since the CKΦ treatment is not effective against resistant bacteria (Supplementary Figure 4). On the contrary, PS/PA liposomes and combination treatment reduced the proportion of CKΦ-resistant bacteria, both in WT and cftr-LOF embryos, to control level (Figures 4E,G), consistent with PS/PA liposomes efficacy in treating the infection by the CKΦ-resistant strain (Supplementary Figure 4).

To test the translational potential of these treatments, we verified the toxicity of PS/PA liposomes and CKΦ single and combined administrations to human cell cultures.

In particular, to assess the effects on bronchial epithelia and immune system of CF patients, we used immortalized CuFi-1 cells expressing F508del CFTR, and the macrophage-like differentiated THP-1 cells treated with the pharmacological inhibitor of CFTR, CFTR_inh-172 (Poerio et al., 2022). As a control, differentiated THP-1 cells not treated with CFTR inhibitor were used. Both cell lines were treated with PS/PA liposomes and CKΦ in single and combined administrations for 48 h and the cell viability was evaluated by MTT assay and compared with control cells treated with the vehicle alone. No cytotoxic effects were observed up to 48 h of treatment for all the conditions tested (Figures 5A,B).