(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Cyclooxygenase production of PGE2 promotes phagocyte control of A. fumigatus hyphal growth in larval zebrafish

['Savini Thrikawala', 'Department Of Biological Sciences', 'Clemson University', 'Clemson', 'South Carolina', 'United States Of America', 'Mengyao Niu', 'Department Of Medical Microbiology', 'Immunology', 'University Of Wisconsin-Madison']

Date: 2022-05

Invasive aspergillosis is a common opportunistic infection, causing >50% mortality in infected immunocompromised patients. The specific molecular mechanisms of the innate immune system that prevent pathogenesis of invasive aspergillosis in immunocompetent individuals are not fully understood. Here, we used a zebrafish larva-Aspergillus infection model to identify cyclooxygenase (COX) enzyme signaling as one mechanism that promotes host survival. Larvae exposed to the pan-COX inhibitor indomethacin succumb to infection at a significantly higher rate than control larvae. COX signaling is both macrophage- and neutrophil-mediated. However, indomethacin treatment has no effect on phagocyte recruitment. Instead, COX signaling promotes phagocyte-mediated inhibition of germination and invasive hyphal growth. Increased germination and invasive hyphal growth is also observed in infected F0 crispant larvae with mutations in genes encoding for COX enzymes (ptgs2a/b). Protective COX-mediated signaling requires the receptor EP2 and exogenous prostaglandin E 2 (PGE 2 ) rescues indomethacin-induced decreased immune control of fungal growth. Collectively, we find that COX signaling activates the PGE 2 -EP2 pathway to increase control A. fumigatus hyphal growth by phagocytes in zebrafish larvae.

Invasive aspergillosis causes mortality in >50% of infected patients. It is caused by a free-living fungus Aspergillus fumigatus which releases thousands of airborne spores. While healthy individuals clear inhaled spores efficiently, in immunocompromised individuals these spores grow into filamentous hyphae and destroy lungs and other tissues causing invasive aspergillosis. The immune mechanisms that control this fungal growth in healthy people are still largely unknown. Here, we used a larval zebrafish model of A. fumigatus infection to determine that cyclooxygenase enzymes, which are the target of non-steroidal anti-inflammatory drugs such as aspirin and ibuprofen, are important to control the fungus. Innate immune cells use cyclooxygenase signaling to prevent hyphal growth and tissue destruction. Our study provides new insights into the mechanisms that immune cells deploy to stop invasive growth of A. fumigatus and inform development of future strategies to combat invasive aspergillosis.

Funding: This work was supported by the National Institute Of Allergy And Infectious Diseases of the National Institutes of Health under award number K22AI134677 to E.E.R. and by the National Institute of General Medical Sciences under a pilot project to E.E.R. from COBRE award number P20GM109094. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here we use this zebrafish larva-Aspergillus infection model to determine the role of the host COX pathway in phagocyte-mediated A. fumigatus clearance. We find inhibition of host COX signaling increases spore germination and invasive growth of hyphae in infected larvae, thereby decreasing host survival. Genetic targeting of COX enzymes with CRISPR/Cas9 confirms that these host enzymes promote control of A. fumigatus germination and growth. COX signaling does not affect macrophage or neutrophil recruitment but instead activates these cells to target the fungus. Exogenous PGE 2 injection restores control of hyphae in COX-inhibited larvae, suggesting that PGE 2 is a major driver of COX-mediated control of fungal growth by phagocytes.

Analyzing how given pathways affect specific aspects of dynamic host-pathogen interactions in vivo is challenging. The zebrafish larva-Aspergillus infection model overcomes many of these challenges, as larvae are transparent and allow for direct visualization of phagocyte-Aspergillus interactions through high-resolution repetitive imaging of the same larvae over the course of a multi-day infection [ 15 ]. Multiple steps in pathogenesis such as phagocyte recruitment, phagocytosis, spore killing, germination, and hyphal growth or clearance can be quantified using this live imaging technique [ 16 ]. Zebrafish have a well-conserved immune system with humans, but depend solely on their innate immune system for the first few weeks of their life, providing a window to study innate immune mechanisms with no interference from the adaptive system [ 17 , 18 ]. The zebrafish larva-Aspergillus model recapitulates multiple aspects of human IA: while immunocompetent larvae are resistant, immunocompromised larvae are susceptible to the infection and develop invasive hyphae [ 19 , 20 ].

Eicosanoids, such as prostaglandins, are arachidonic acid-derived lipid signaling molecules that function in both an autocrine and paracrine manner by binding to their receptors and can have a variety of effects on immune cell function [ 6 , 7 ]. Prostaglandins are produced by prostaglandin endoperoxide synthases (PTGSs), also called cyclooxygenase (COX) enzymes. COX enzymes are the target of non-steroidal anti-inflammatory drugs such as aspirin, ibuprofen and indomethacin [ 8 – 10 ]. COX-derived prostaglandins can have either pro- or anti-inflammatory effects on innate immune cells, modulating both phagocyte recruitment and phagocyte functions [ 6 , 7 , 11 ]. In response to fungal pathogens, prostaglandin signaling is known to inhibit phagocytosis of Candida albicans by macrophages, H 2 O 2 -mediated fungicidal activity against Paracoccidioides brasiliensis, and M1 polarization of alveolar macrophages and killing of Cryptococcus neoformans [ 12 – 14 ]. However, the roles of COX activation and prostaglandins during A. fumigatus infections are not known.

Aspergillus fumigatus is a free-living saprophytic fungus which reproduces asexually by producing thousands of conidia or spores. Owing to their small size and hydrophobicity, spores become airborne causing widespread contamination both indoors and outdoors. It is estimated that an average person can inhale 100–1000 spores per day [ 1 ]. Although healthy immune systems can combat these spores, in immunocompromised individuals spores can germinate to form invasive hyphae which spread to multiple organs and tissues—a condition called invasive aspergillosis (IA) [ 1 ]. IA remains a major cause of mortality in immunocompromised patients, particularly individuals with hematological malignancies, bone marrow or solid-organ transplant recipients, HIV patients, ICU patients, and patients with altered lung conditions [ 2 ]. Despite the availability of anti-fungal drugs, the mortality rate of IA remains at ~50% [ 3 – 5 ]. Hence, it is imperative to develop novel strategies to target fungi and augment anti-fungal immune responses, but this requires a better understanding of immune cell-pathogen interactions. Innate immune cells act as the first line of defense against inhaled A. fumigatus conidia. However, the signaling and effector mechanisms that these cells use to inhibit fungal growth are not fully understood.

In these experiments, in the absence of indomethacin treatment, we found increased fungal germination and invasive hyphae upon PGE 2 injection, although these differences were not statistically significant ( Fig 7B ). To test if PGE 2 can directly act on spores to promote germination, we quantified A. fumigatus germination in vitro in the presence of exogenous PGE 2 , but did not find any effect ( Fig 7D ). PGE 2 levels must be precisely controlled and balanced and an alternative explanation is that high levels of PGE 2 without concomitant indomethacin-mediated COX inhibition promote an anti-inflammatory environment via the EP4 receptor, leading to lowered immune control of fungal growth [ 28 ]. Therefore, we tested if excess PGE 2 still leads to increased fungal germination and invasive hyphal growth when the EP4 receptor is also blocked. Infected larvae exposed to the EP4 antagonist AH23848 were injected with PGE 2 at 1 dpi. However, in these experiments, PGE 2 injection alone did not cause increased fungal growth, possibly due to small differences in injection volume, further underlining how small differences in PGE 2 levels can lead to different outcomes. Interestingly, however, exposure to AH23848 alone or in combination with PGE 2 increased invasive hyphal growth in infected larvae, although these differences were not statistically significant ( Fig 7E ). These data together with the slightly increased mortality of infected larvae in the presence of AH23848 ( Fig 1D ) suggest that EP4 may also play a minor role in mediating immune cell-mediated fungal growth control. Collectively, our findings demonstrate that COX-mediated PGE 2 production and signaling via the EP2 receptor promotes phagocyte-mediated control of A. fumigatus germination and hyphal growth ( Fig 8 ).

Larvae were injected with mCherry-expressing TBK5.1 (Af293) spores and exposed to 10 μM indomethacin or DMSO vehicle control at 2 dpf. At 1 dpi, larvae were injected with 10 μM PGE 2 or DMSO vehicle control. (A) Survival of wild-type larvae was monitored. Cox proportional hazard regression analysis was used to calculate P values and hazard ratios (HR). Data are pooled from four independent replicates, at least 23 larvae per condition, per replicate and total larval N per condition is indicated. Average injection CFUs: 40. (B-C) Larvae were imaged at 3 dpi. (B) Percentage of larvae with germination and invasive hyphae. Data are pooled from three independent replicates, at least 8 larvae per condition, per replicate, P values calculated by Fisher’s Exact Test. (C) Representative images showing hyphal growth in each condition. Scale bars = 50 μm. (D) TBK1.1 (Af293) spores were inoculated into RPMI media in the presence of 1 μM PGE 2 , 10 μM PGE 2 , or DMSO vehicle control. Every 2 hours, an aliquot of spores was removed and scored for germination. Points represent pooled means ± SEM from two independent replicates. (E) Larvae were injected with GFP-expressing TBK1.1 (Af293) spores and exposed to 5 μM AH23848 or DMSO vehicle control at 2 dpf. At 1 dpi, larvae were injected with 10 μM PGE 2 or DMSO vehicle control. Larvae were imaged at 3 dpi and the percentage of larvae with germination and invasive hyphae in each condition was calculated. Data are pooled from three independent replicates, at least 11 larvae per condition, per replicate, P values calculated by Fisher’s Exact Test.

So far we have established that COX signaling promotes macrophage- and neutrophil-mediated control of germination and invasive hyphal growth in an A. fumigatus infection. It is likely that COX signaling acts via a PGE 2 -EP2 signaling axis, as EP2 receptor antagonist-treated infected larvae have increased fungal growth ( S8 Fig ) and succumb to infection at the same rate as indomethacin-treated larvae ( Fig 1C ). We therefore tested if exogenous PGE 2 can rescue the effects of indomethacin treatment in infected larvae. Since PGE 2 is short-lived and elicits short-range effects, we injected A. fumigatus-infected, indomethacin- or DMSO-treated larvae with PGE 2 or DMSO vehicle control into the hindbrain at 1 dpi. PGE 2 injection partially rescues survival of indomethacin-treated larvae, although the effect is not statistically significant ( Fig 7A ). To determine if PGE 2 can rescue indomethacin-inhibited functions of phagocytes against invasive fungal growth, we imaged the larvae at 3 dpi. As seen previously, indomethacin treatment increases the percentage of larvae harboring both germination and invasive hyphae at 3 dpi ( Fig 7B ). PGE 2 supplementation rescues these phenotypes, decreasing germination and development of invasive hyphae ( Fig 7B and 7C ), without affecting phagocyte recruitment ( S12A and S12B Fig). However, PGE 2 injection cannot increase the survival of either wild-type or neutrophil-defective larvae not treated with indomethacin (Figs 7A and S12C ).

Embryos were injected with gRNAs targeting ptgs2a and ptgs2b or control gRNAs for gfp and Cas9 protein. F0 crispant larvae were then injected with GFP-expressing TBK1.1 (Af293) spores at 2 dpf and imaged at 3 dpi. (A) Percentage of larvae with germination or invasive hyphae. Data are pooled from two independent replicates, at least 23 larvae per condition, per replicate. Average injection CFUs: control = 38, ptgs2a/2b = 39. (B) Representative images showing hyphal growth in each condition. Scale bar = 50 μm.

One explanation for the observed effects of indomethacin specifically on control of later stages in fungal pathogenesis is a delay in drug absorption and accumulation at the infection site. To test this, we tried pre-treating larvae and treating larvae with a higher indomethacin concentration, but both conditions caused toxicity as evidenced by high mortality in mock-infected larvae ( S10 Fig ). To avoid these issues of drug absorption and to confirm the role of host COX enzymes in controlling A. fumigatus spore germination and invasive fungal growth, we used CRISPR/Cas9 to target both zebrafish COX2 genes: ptgs2a and ptgs2b. F0 larvae injected with gRNAs targeting both genes or control gRNAs were infected with A. fumigatus spores expressing GFP and imaged at 3 dpi. PCR of genomic DNA flanking the gRNA target sites confirmed efficient DNA alteration ( S11 Fig ). In ptgs-targeted larvae, we observed an increased occurrence of both spore germination and invasive hyphal growth ( Fig 6 ), confirming that host COX enzymes increase the ability of immune cells to control these later stages of fungal pathogenesis.

Neutrophils are thought to be the major phagocyte that targets and kills invasive hyphae. In irf8 -/- larvae which lack macrophages and have an abundance of neutrophils, neutrophils destroy fast-germinating strains of A. fumigatus such as CEA10 within a few days [ 20 ]. We therefore decided to use this infection scenario to specifically test the requirement for COX signaling in neutrophil-mediated hyphal killing. We infected irf8 -/- larvae with CEA10 spores and isolated larvae for CFU enumeration at 0, 1, and 2 dpi. CFUs from irf8 -/- were normalized to CFUs of irf8 +/+ /irf8 +/- at each dpi for each condition. At 2 dpi in DMSO-treated larvae, ~26% of the fungal burden remains in irf8 -/- larvae compared to macrophage-sufficient larvae (irf8 +/+ or irf8 +/- ), demonstrating the neutrophil-mediated clearance of fungus that occurs in these larvae ( S9A Fig ). Fungal clearance is slightly alleviated but not significantly different in indomethacin-exposed irf8 -/- larvae, suggesting that COX signaling may promote but is not required for neutrophil-mediated killing of hyphae ( S9A Fig ). Similar to infection with Af293-derived strains, CEA10-infected larvae also succumb to the infection at a higher rate in the presence of indomethacin both in irf8 +/+ /irf8 +/- and irf8 -/- backgrounds ( S9B Fig ).

Next, we rated the severity of fungal growth on a scale of 0 to 4, from no germination to severe invasive growth of hyphae, and a lethal score of 5 ( S7 Fig ), focusing only on larvae that had germination within them at some point in the experiment. Severe growth of invasive hyphae is prominent in larvae exposed to indomethacin, eventually causing mortality ( Fig 5D ). Although germination occurred in the vehicle control group, these larvae are able to delay invasive growth compared to indomethacin-treated larvae ( Fig 5D ), suggesting that the major defect in these larvae is a failure to control hyphal growth post-germination. To quantify this time of delay between appearance of germination and invasive hyphae, we analyzed the timeline of first appearance of germlings and invasive hyphae in larvae more closely. We quantified the day germination was first observed, the day invasive hyphae was first observed, and the time between these two occurrences. The day to first observe germination was similar between the two groups ( Fig 5E ). However, invasive hyphae appear significantly earlier after both initial infection ( Fig 5F ) and after germination ( Fig 5G ) in indomethacin-exposed larvae. Once spores are germinated, invasive hyphae appear on average ~1 day later in control larvae, while in indomethacin-treated larvae, this growth only takes an average of ~0.5 days ( Fig 5G ). We also used microscopy to determine fungal growth in larvae exposed to the EP2 antagonist AH6809. Blockage of this receptor also leads to increased spore germination and invasive hyphal growth at 3 dpi at a level similar to indomethacin ( S8 Fig ). Together, these results suggest that COX-PGE 2 signaling promotes phagocyte-mediated control of invasive hyphal growth.

After germination, A. fumigatus hyphae branch and grow into a network disrupting the host tissue. The cumulative percentage of larvae with invasive hyphae (as defined by branched hyphal growth ( S7 Fig )) is also significantly higher with indomethacin treatment ( Fig 5B ). We also quantified the hyphal burden by measuring the fungal area, finding more extensive hyphal growth in indomethacin-treated larvae at both 3 and 5 dpi, although this difference is not statistically significantly ( Fig 5C ), likely due to high variability between larvae and the large number of indomethacin-treated larvae that succumbed to infection before 5 dpi ( Fig 1A ).

Since germination is increased upon indomethacin treatment, we wanted to confirm again that the effects of indomethacin are on the host and that indomethacin does not directly alter A. fumigatus germination. To test this, A. fumigatus spores were inoculated in vitro in liquid RPMI medium in the presence of indomethacin or DMSO and the percentage of germinated spores was scored at 2-hour intervals. We find no difference in the rate of germination between indomethacin- and DMSO-treated spores ( S6 Fig ). Therefore, our data demonstrate that indomethacin decreases host immune cell-mediated control of A. fumigatus germination.

Zebrafish larvae were injected with mCherry-expressing TBK5.1 (Af293) spores at 2 dpf, exposed to 10 μM indomethacin or DMSO vehicle control and imaged at 1, 2, 3, and 5 dpi. (A) Representative images showing spore germination (inset white arrow) and invasive hyphae (branched hyphae, inset open white arrow). Scale bar = 50 μm (10 μm in insets). (B) Cumulative percentage of larvae with germination (dotted line) and invasive hyphae (solid line) through 5 dpi. Cox proportional hazard regression analysis was used to calculate P values and hazard ratios (HR). (C) 2D fungal area was quantified from image z projections. Each line represents an individual larva and bars represent pooled emmeans ± SEM from 8 independent replicates, at least 12 larvae per condition, per replicate. (D) Severity of fungal growth was scored for all larvae and displayed as a heatmap. Representative images for each score can be found in S7 Fig . (E-G) In larvae in which (E) germination (indomethacin n = 61, DMSO n = 45) and (F) invasive hyphae occurred (indomethacin n = 42, DMSO n = 27), the day on which each was first observed is plotted. (G) The number of days between germination and invasive hyphae was also calculated. Bars represent pooled emmeans ± SEM from eight individual replicates, P values calculated by ANOVA. Each data point represents an individual larvae, color-coded by replicate.

As spore killing is not affected by COX inhibition, we next hypothesized that immune control of the next stages in fungal pathogenesis—spore germination and invasive hyphal growth—is modulated by COX signaling. To monitor spore germination and hyphal growth in larvae, we infected larvae with A. fumigatus spores expressing mCherry and imaged at 1, 2, 3 and 5 dpi with confocal microscopy. We find spore germination in both indomethacin- and DMSO-exposed larvae ( Fig 5A ). However, both the rate at which larvae are observed to have germination inside of them and the total percentage of larvae that harbor germinated spores is significantly higher in the presence of indomethacin ( Fig 5B ).

(A, B) Macrophage-labeled larvae Tg(mfap4:mTurquoise2) were injected with YFP-expressing A. fumigatus TBK1.1 (Af293) spores coated with AlexaFluor 546 at 2 dpf, exposed to 10 μM indomethacin or DMSO vehicle control, and imaged at 2 dpi. (A) Representative images showing live (white arrow) and dead (open arrow) spores within a macrophage. Scale bar = 10 μm. (B) The percentage of live spores in the hindbrain, and specifically within macrophages, per larvae. Each data point represents an individual larvae, color-coded by replicate (indomethacin n = 31, DMSO n = 30). Bars represent pooled emmeans ± SEM from three independent replicates, P values calculated by ANOVA. (C) Wild-type larvae were injected with TBK1.1 (Af293) spores at 2 dpf, exposed to 10 μM indomethacin or DMSO vehicle control, and fungal burden was quantified by homogenizing and plating individual larvae for CFUs at multiple days post injection. Eight larvae per condition, per dpi, per replicate were quantified, and the number of CFUs at each dpi is represented as a percentage of the initial spore burden. Bars represent pooled emmeans ± SEM from three individual replicates, P values calculated by ANOVA. Average injection CFUs: 27. Non-normalized raw CFU data is presented in S5 Fig .

We next hypothesized that the functions of these phagocytes are modulated by COX signaling. The initial response of macrophages is to phagocytose injected spores and activate spore killing mechanisms [ 19 , 20 ]. To determine if COX inhibition affects spore killing, we used a live-dead staining method in which A. fumigatus spores expressing YFP are coated with AlexaFluor546 and injected into zebrafish larvae expressing mTurquoise in macrophages [ 20 , 33 ]. Larvae were imaged with confocal microscopy at 2 dpi, and we enumerated the number of live versus dead spores. Live spores are visualized as YFP signal surrounded by AlexaFluor signal, while dead spores only have AlexaFluor signal ( Fig 4A ). The percentage of live spores is similar in indomethacin and DMSO groups both within macrophages and in the whole imaged hindbrain area ( Fig 4B ). To confirm these results, we also measured the overall fungal burden in indomethacin- or DMSO-treated larvae over the 7-day infection period with CFU counts. Consistent with live-dead staining, the fungal burden is similar between DMSO- and indomethacin-exposed larvae throughout the infection (Figs 4C and S5 ), indicating that COX signaling does not drive spore clearance.

As described previously, macrophages are recruited starting at 1 day post infection (dpi) and form clusters around spores starting at 2–3 dpi ( Fig 3B ), with neutrophils primarily responding after spores germinate ( Fig 3C ) [ 20 ]. The number of recruited macrophages ( Fig 3D ), macrophage cluster area ( Fig 3E ), and the number of recruited neutrophils ( Fig 3F ) are not significantly different between indomethacin- and DMSO-exposed groups at days 1, 2, and 3 post infection. At 5 dpi, however, more macrophages ( Fig 3D ) and neutrophils ( Fig 3F ) are found at the infection site in indomethacin-treated larvae. Fungal germination occurs at these later time stages and attracts more immune cells. To control for this variable, we analyzed the number of macrophages and neutrophils in each larva relative to the day germination and invasive hyphae were first observed. Using this normalization, we find that macrophage numbers ( Fig 3G ) and neutrophil numbers ( Fig 3H ) are similar between the two conditions at each stage of fungal pathogenesis. Overall, our results indicate that phagocyte recruitment is not dependent on COX activation.

Larvae were injected with mCherry-expressing A. fumigatus TBK5.1 (Af293) spores at 2 dpf. After injection larvae were exposed to 10 μM indomethacin or DMSO vehicle control and were imaged through 5 dpi. (A) Schematic showing the infection and imaging area of zebrafish larvae (created with BioRender.com ). (B, D, E, G) Macrophage nuclear-labeled Tg(mpeg1:H2B-GFP) larvae were imaged at 1, 2, 3, and 5 dpi. (C, F, H) Neutrophil-labeled Tg(lyz:BFP) larvae were imaged at 2, 3, and 5 dpi. (B, C) Representative z-projection images showing macrophage and neutrophil recruitment. Scale bars = 50 μm. (D) Number of macrophages recruited, (E) 2D macrophage cluster area, and (F) number of neutrophils recruited were quantified. Each line represents an individual larva followed for the entire course of infection and bars represent pooled emmeans ± SEM from four independent replicates, at least 12 larvae per condition, per replicate. P values were calculated by ANOVA. (G) Number of macrophages and (H) neutrophils one day before germination occurred, on the day of germination, and on the day invasive hyphae occurred were quantified. Bars represent pooled emmeans ± SEM from all larvae with germination from four independent replicates. Data points represent individual larva and are color coded by replicate. P values were calculated by ANOVA.

We next sought to define how the innate immune response is altered by COX inhibition. COX-synthesized prostaglandins are chemical messengers that can function to recruit immune cells to infection sites, and we wondered whether COX inhibition affects macrophage or neutrophil recruitment to A. fumigatus infection [ 7 , 20 ]. Zebrafish larvae expressing GFP in macrophages (Tg(mpeg1:H2B-GFP)) or BFP in neutrophils (Tg(lyz:BFP)) were infected with A. fumigatus spores expressing mCherry, and treated with indomethacin or DMSO vehicle control and we enumerated the number of macrophages and neutrophils at the infection site through daily confocal imaging ( Fig 3A ).

To further ascertain the roles of each cell type in COX signaling, we determined whether macrophages and neutrophils sorted from 3 dpf larvae express genes in the COX/PGE 2 pathway. Zebrafish possess two copies of COX2: prostaglandin-endoperoxide synthase-2a and -2b (ptgs2a and ptgs2b) [ 32 ], both of which we find expressed in macrophages and neutrophils, while neither cell type expresses the zebrafish COX1 gene (ptgs1) ( S4 Fig ). COX enzymes synthesize PGH 2 , an intermediate that is converted to PGE 2 by prostaglandin E synthases, which are either membrane-bound (mPges) or cytosolic (cPges) [ 6 ]. All zebrafish pges genes are expressed by macrophages, while neutrophils express only the cytosolic versions ( S4 Fig ). At least one EP2 receptor gene (ep2a or ep2b) is expressed in both phagocytes ( S4 Fig ). These data demonstrate that both macrophages and neutrophils express the machinery needed to generate and respond to COX/PGE 2 signaling.

We next tested the survival of neutrophil-defective (mpx:rac2D57N) infected larvae. In these larvae neutrophils are unable to migrate to the infection site [ 30 ]. As found previously, neutrophil-defective larvae are more susceptible to A. fumigatus infection than wild-type controls ( Fig 2C ) [ 20 , 31 ]. Indomethacin exposure further decreases survival of neutrophil-defective larvae ( Fig 2C ). Compared to wild-type larvae which are 3.9 times more likely to succumb to infection, neutrophil-defective larvae are only 1.8 times more likely to succumb to infection, suggesting that neutrophils also partially mediate the host-protective effects of COX signaling, but that other cell types can be involved. Lack of neutrophils has no effect on survival of mock-infected larvae treated with indomethacin ( S3C Fig ). Together, these data demonstrate that both macrophages and neutrophils participate in COX-mediated responses to promote survival of A. fumigatus-infected zebrafish larvae.

Then we interfered with macrophage and neutrophil function individually to determine if each cell type is required for COX-mediated host protection. We injected 1 dpf larvae with clodronate liposomes to deplete macrophages or PBS liposomes as a control. As observed previously, macrophage-depleted larvae rapidly succumb to the infection ( Fig 2B ) [ 20 ]. However, indomethacin exposure further decreases the survival of macrophage-depleted larvae ( Fig 2B ). While indomethacin treatment makes control larvae 8.4 times more likely to succumb to infection, clodronate liposome-injected larvae are only 1.7 times more likely to succumb upon indomethacin treatment ( Fig 2B ), suggesting that macrophages partially mediate the host-protective effects of COX signaling, but that even in the absence of macrophages COX signaling increases host survival. Larvae injected with clodronate liposomes and then given a PBS mock-infection also have lower survival upon indomethacin treatment, suggesting that some of this death may be due to the effects of the clodronate alone, although this difference in PBS mock-infected larvae is not statistically significant ( S3B Fig ).

Survival of larvae injected with TBK1.1 (Af293) spores at 2 dpf. (A) Development of phagocytes was inhibited by pu.1 morpholino (MO). Control larvae received standard control MO. (B) Macrophages were depleted via clodronate liposome i.v. injection at 1 dpf. Control larvae received PBS liposomes. (C) Neutrophil-defective larvae (mpx:rac2D57N) were compared to wild-type larvae. Data are pooled from three independent replicates, at least 10 larvae per condition, per replicate and the total larval N per condition is indicated in each figure. Cox proportional hazard regression analysis was used to calculate P values and hazard ratios (HR). Average injection CFUs: (A) control MO = 25, pu.1 MO = 24, (B) PBS liposomes = 20, clodronate liposomes = 23, (C) wild-type = 26, mpx:rac2D57N = 22.

We next sought to determine which innate immune cells utilize COX signaling to fight A. fumigatus infection. Macrophages and neutrophils are the primary immune cells that combat A. fumigatus infection in zebrafish larvae [ 20 ]. To determine if these cell types play a role in COX-mediated host responses, we inhibited development of both phagocytes by knocking down pu.1 (spi1b) via morpholino injection [ 29 ]. If COX signaling activates phagocytes to clear the infection, we expect that indomethacin treatment of larvae that are already depleted of phagocytes would have no effect on larval survival. Larvae were injected with A. fumigatus spores or PBS and exposed to indomethacin or DMSO. Indomethacin exposure significantly decreases survival of larvae injected with a control morpholino but has no effect on pu.1 morphants (Figs 2A and S3A ), suggesting that COX-mediated host protection is phagocyte-dependent.

Among COX-biosynthesized prostaglandins, prostaglandin E 2 (PGE 2 ) is a major product synthesized by phagocytes that moderates a range of inflammatory processes and has pro- or anti-inflammatory functions, depending on the receptor to which it binds [ 26 ]. PGE 2 elicits its actions via four different E type prostanoid receptors, EP1-4, with most immunomodulatory effects mediated via EP2 and EP4 [ 26 ]. Therefore, we tested if EP2 and 4 receptor antagonists affect the disease outcome of A. fumigatus-infected larvae. We used antagonists of EP2: AH6809 [ 27 ] and EP4: AH23848 [ 22 , 27 , 28 ] previously used in zebrafish larvae. Larvae exposed to AH6809 succumb to the infection at a similar rate as indomethacin-exposed larvae, both with a hazard ratio of 3 compared to control larvae, while AH23848-exposed larvae show no significant difference in survival compared to control ( Fig 1C ). These data suggest that COX signaling promotes A. fumigatus-infected larval survival via a PGE 2 -EP2 signaling pathway.

A. fumigatus also harbors three Ppo enzymes (PpoA, PpoB and PpoC) with high identity to vertebrate COX [ 24 ]. While a previous study reported that indomethacin does not affect the function of these enzymes [ 25 ], we wanted to confirm that the observed effects of indomethacin on infected larval survival are due to inhibition of host enzymes, not fungal enzymes. First, we confirmed that during the larval stage, zebrafish have detectable levels of COX activity ( S1C Fig ). Next, to determine the role of A. fumigatus Ppo enzymes, we infected zebrafish larvae with A. fumigatus spores lacking all three ppo genes (ΔppoA, ΔppoB, ΔppoC) ( S2 Fig ). Deletion of ppo enzymes had no effect on fungal virulence, as survival of larvae infected with triple-ppo-mutant A. fumigatus spores is similar to larvae infected with wild-type spores ( Fig 1B ). Additionally, indomethacin treatment decreased survival equally in larvae infected with triple-ppo-mutant and wild-type spores ( Fig 1B ). These data demonstrate that indomethacin inhibits host enzymes to compromise survival of A. fumigatus-infected larvae. Since no survival difference was observed in larvae infected with a triple-ppo-mutant, we focused on wild-type A. fumigatus for the remainder of the study.

(A) Survival of wild-type larvae injected at 2 dpf with TBK1.1 (Af293) A. fumigatus spores or PBS mock-infection in the presence of 10 μM indomethacin or DMSO vehicle control. (B) Survival of larvae injected at 2 dpf with A. fumigatus Af293 or Af293 triple-ppo-mutant (Δppo) spores and exposed to 10 μM indomethacin or DMSO vehicle control. (C) Survival of larvae injected at 2 dpf with TBK1.1 (Af293) spores and exposed to 10 μM indomethacin, 5 μM AH6809, 10 μM AH23848, or DMSO vehicle control. Data are pooled from at least three independent replicates, at least 24 larvae per condition, per replicate and the total larval N per condition is indicated in each figure. Cox proportional hazard regression analysis was used to calculate P values and hazard ratios (HR). Average injection CFUs: (A) 50, (B) Af293 = 40 and Δppo = 58, (C) 28.

Prostaglandins are lipid signaling molecules whose production is induced during inflammation via cyclooxygenase (COX) enzymes. We used the zebrafish larva-Aspergillus infection model to test the hypothesis that host COX signaling promotes larval survival and fungal clearance in an A. fumigatus infection. Wild-type A. fumigatus spores were microinjected into the hindbrain ventricle of 2 days post fertilization (dpf) larvae. Infected larvae were then exposed to the pan-COX inhibitor indomethacin or DMSO vehicle control immediately after injection and larval survival was monitored for 7 days. Indomethacin is a well-established non-steroidal anti-inflammatory drug that inhibits COX enzyme activation and prostaglandin synthesis [ 9 , 10 ] and is widely used in a variety of animal models including zebrafish [ 21 , 22 ]. Indomethacin-treated larvae succumb to infection at a significantly greater rate than control larvae ( Fig 1A ). Treatment with the COX1 inhibitor SC560 [ 23 ] or COX2 inhibitor meloxicam [ 21 ] also significantly increases infected larval mortality ( S1A and S1B Fig ). With each of these inhibitors, no significant decrease in survival was observed in PBS-injected mock-infected larvae (Figs 1A and S1A and S1B ).

Discussion

Healthy immune systems can contain and kill A. fumigatus spores despite the fact that hundreds of spores can be inhaled per day. While the physiological role of macrophages and neutrophils in this context is well-appreciated, the molecular mechanisms that each of these cell types use to combat each stage of fungal pathogenesis are not fully understood. The critical step in A. fumigatus pathogenesis is the transition from dormant spore to hyphal growth, causing tissue destruction. Here, we used a zebrafish larva-A. fumigatus infection model to identify COX-PGE 2 signaling as one mechanism that promotes control of this transition to invasive hyphae by both macrophages and neutrophils (Fig 8).

To investigate the role of COX signaling in phagocyte responses we used indomethacin, a pan-COX inhibitor, as well as COX1- and COX2-specific inhibitors. We find that activity of both enzymes promotes survival of A. fumigatus-infected larvae. Broadly, COX1 activity is involved in tissue homeostasis while COX2 is inducible and is involved in responses to inflammatory stimuli [6]. However, evidence suggests that both isoforms are activated during inflammation. Mice lacking COX1 have impaired inflammatory responses [34], and COX1 is activated in response to LPS-induced inflammation in humans [35]. Zebrafish have one functional isoform of COX1 and two functional orthologues of COX2: COX2a and COX2b [32]. COX enzymes can produce prostaglandins in multiple cell types including epithelial cells, endothelial cells and fibroblasts, but infiltrating innate immune cells are the major source of these lipid signals during inflammation, including both macrophages [36] and neutrophils [11]. We find that both copies of COX2, but not COX1, are expressed in macrophages or neutrophils, suggesting that phagocytes mainly deploy COX2-mediated responses in zebrafish larvae. In line with this, F0 crispant larvae in which ptgs2a and 2b have been targeted but in which ptgs1 is intact have impaired immune control of A. fumigatus infection. However, we cannot rule out an additional role of other cell types, such as epithelial cells, in initiating COX signaling in response to A. fumigatus, which could be mediated via COX1, COX2, or both.

COX enzymes catalyze the main regulatory step of prostaglandin synthesis: conversion of arachidonic acid to prostaglandin H 2 (PGH 2 ). PGH 2 is then converted to one of the four major types of prostaglandins, prostaglandin E 2 (PGE 2 ), prostaglandin D 2 (PGD 2 ), prostaglandin I 2 (PGI 2 ) and prostaglandin F 2α (PGF 2α ) via different synthases. Perhaps the most studied prostaglandin is PGE 2 , due to its paradoxical immunomodulatory effects [37]. Membrane-bound (mPges) or cytosolic (cPges) PGE 2 synthases convert PGH 2 to PGE 2 . cPges is functionally coupled with COX1 and responsible for constitutive PGE 2 , while COX2-mPges is inducible in response to inflammatory stimuli [38,39]. In the current study, mpges was only detected in macrophages, suggesting macrophages may initiate PGE 2 signaling. Once released, PGE 2 signals via four different receptors, EP1-EP4 which have different affinities for PGE 2 , and activate different downstream effects [26]. We find that pharmacological inhibition of EP2 leads to impaired control of fungal growth and increased susceptibility of the larval host at a rate similar to COX inhibition. At least one EP2 receptor gene (ep2a or ep2b) is expressed in both macrophages and neutrophils in larvae. Determining the specific role of each COX enzyme, PGE 2 synthase, and EP receptor in each cell type in response to A. fumigatus is complicated in zebrafish due to the presence of a teleost genome duplication that led to multiple copies of these genes [40], but should be the subject of future studies.

The decreased larval host survival that we observe when COX-PGE 2 signaling is inhibited is due to decreased phagocyte control of A. fumigatus spore germination and hyphal growth. In invasive aspergillosis, it is this hyphal growth that can penetrate into host tissues and cause organ damage and mortality. However, spore germination is a ‘double-edged sword’—while it is required for fungal growth and pathogenesis, it also causes unmasking of fungal PAMPs and activation of immune responses [41]. As a result, in some infection scenarios, increased germination actually leads to better host survival. For example, irf8-/- larval zebrafish that lack macrophages and have an excess of neutrophils clear infections with a fast-germinating CEA10 strain of A. fumigatus at a higher rate than wild-type larvae [20]. This specific host-pathogen interaction does not represent a "normal" A. fumigatus infection in which only a percentage of the spores present germinate in the presence of macrophages [20]. In these "normal" scenarios neutrophils take longer to respond and a slight decrease in immune control of fungal germination rate or hyphal growth rate will allow the infection to cause tissue destruction and death. However, we did use this irf8-/- host infected with CEA10 to specifically measure the requirement for COX signaling in neutrophil killing of A. fumigatus hyphae. We find that when COX enzymes are inhibited, neutrophil-mediated CFU clearance is not significantly affected. These experiments also demonstrated that COX signaling is important for immune responses to the CEA10 strain as well as the Af293 strain used in the majority of our study.

We report here that PGE 2 signaling, specifically through the EP2 receptor, promotes phagocyte control of A. fumigatus invasive hyphal growth, but the cellular mechanisms upstream and downstream of this signaling in larval zebrafish are not yet determined. Treatment with a COX2 inhibitor was previously found to cause higher fungal burdens in A. fumigatus-infected mice [42]. Garth et al. found that signaling through the IL-33 receptor inhibits PGE 2 production and inhibits fungal clearance in the infected lung [42]. The cytokines IL-17A and IL-22 were also decreased in mice with lower levels of PGE 2 , suggesting that these signals may act downstream of PGE 2 to coordinate the immune response [42,43]. Zebrafish have homologs for both il22 and il17a, and il22 mRNA expression is induced upon fungal infection in adult zebrafish [44], but their roles in fungal infections in larval zebrafish are unknown. It is worth noting that these cytokines are thought to be primarily expressed by T cells, which are not yet developed in larval zebrafish, and our results demonstrate that COX-PGE 2 signaling also promotes immune responses to A. fumigatus infection in the absence of functional adaptive immunity.

We do not yet know what downstream effector mechanisms are activated in neutrophils and macrophages by COX signaling to target fungal growth. Prostaglandins can mediate endothelial cell permeability and facilitate immune cell infiltration, but we observe no difference in phagocyte recruitment when COX enzymes are inhibited [6,45,46]. Prostaglandins can also regulate extracellular killing mechanisms in phagocytes such as reactive oxygen species (ROS) production, neutrophil degranulation, and extracellular trap (ET) formation [47–49]. Further testing is required to determine if these mechanisms are enhanced by COX signaling during infection with A. fumigatus.

In contrast, it was previously reported that PGE 2 suppresses phagocytosis and microbial killing of fungal pathogens such as P. brasiliensis [12], C. albicans [13] and C. neoformans [14]. These differences underline the idea that PGE 2 can have both pro- or anti-inflammatory functions, can differentially impact diverse fungi, and that PGE 2 levels must be tightly controlled during infection to promote infection clearance. PGE 2 binds to EP2 with low affinity and generally evokes pro-inflammatory responses while it binds to EP4 with high affinity and activates anti-inflammatory responses [26]. In line with this idea, PGE 2 can drive resolution of inflammatory phenotypes through EP4 in zebrafish larvae during injury [28]. However, our data demonstrate that PGE 2 -EP4 may also partially contribute to immune cell-mediated fungal growth control during A. fumigatus infection rather than inhibiting immune activation. PGE 2 levels, and eicosanoid levels in general, must be tightly controlled during infection, and an eicosanoid imbalance in either direction is known to increase pathogenesis of mycobacterial infections in zebrafish, mice, and humans [50–52].

Fungal species can also synthesize their own lipid signaling molecules to modulate pathogenesis and immune responses [53,54]. For instance, Cryptococcus neoformans strains deficient in eicosanoid production have intracellular growth defects, which can be reversed by addition of exogenous PGE 2 in a zebrafish larvae model [27]. A. fumigatus PpoA and PpoC enzyme activity can produce prostaglandins and similar bioactive oxylipins that can affect Aspergillus virulence and development [24] and phagocytosis of conidia [55]. However, we find that A. fumigatus Ppo enzymes do not affect fungal virulence in this larval zebrafish infection model, and that indomethacin treatment does not alter A. fumigatus spore germination in vitro. Additionally, while a previous study showed that exogenous PGE 2 inhibited pigment formation in A. fumigatus hyphae which could affect invasive hyphal growth [24], we do not find any direct effect of PGE 2 on A. fumigatus germination in vitro.

[END]

[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010040

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.