NADPH oxidase-driven phagocyte recruitment controls Candida albicans filamentous growth and prevents mortality

Abstract

Candida albicans is a human commensal and clinically important fungal pathogen that grows as both yeast and hyphal forms during human, mouse and zebrafish infection. Reactive oxygen species (ROS) produced by NADPH oxidases play diverse roles in immunity, including their long-appreciated function as microbicidal oxidants. Here we demonstrate a non-traditional mechanistic role of NADPH oxidase in promoting phagocyte chemotaxis and intracellular containment of fungi to limit filamentous growth. We exploit the transparent zebrafish model to show that failed NADPH oxidase-dependent phagocyte recruitment to C. albicans in the first four hours post-infection permits fungi to germinate extracellularly and kill the host. We combine chemical and genetic tools with high-resolution time-lapse microscopy to implicate both phagocyte oxidase and dual-specific oxidase in recruitment, suggesting that both myeloid and non-myeloid cells promote chemotaxis. We show that early non-invasive imaging provides a robust tool for prognosis, strongly connecting effective early immune response with survival. Finally, we demonstrate a new role of a key regulator of the yeast-to-hyphal switching program in phagocyte-mediated containment, suggesting that there are species-specific methods for modulation of NADPH oxidase-independent immune responses. These novel links between ROS-driven chemotaxis and fungal dimorphism expand our view of a key host defense mechanism and have important implications for pathogenesis.

Figure 1. Photoswitching time-lapse shows phagocytosis block of germination is long-lasting and NADPH oxidase-independent.

(A) CAF2-dTomato-infected fish were treated with DMSO or DPI and scored at 4 hpi for fungal morphotype and internalization. Filamentous growth is only seen extracellularly, and this difference is highly significant (p<0.001) in both DMSO (n = 11) and DPI (n = 12), as measured by Fisher's exact test. Data are pooled from three independent experiments. (B–E) CAF2-dTomato-infected Tg(mpeg1:GAL4/UAS:Kaede) fish were treated with DMSO (n = 11) or DPI (n = 7) from two hours pre-infection and throughout imaging. At 4 hpi, only macrophages at the infection site in the hindbrain were photoswitched green-to-red by exposure to 405 nm laser light. Photoswitched fish were imaged every two hours for an additional 18 hours. (B) Representative images of macrophage movement during the time-lapse, with a schematic above depicting the movement of macrophages out of the hindbrain and acquisition of yellow color as fresh green-fluorescent Kaede protein is produced. Scale bar = 50 µm. (C) Representative images of DMSO- (left) and DPI-treated (right) fish at 18 hpi, showing filamentous growth in the DPI-treated but not the control fish. Scale bar = 50 µm. (D) Macrophages at the site of infection were enumerated at each time-point. Green and red bars represent green-fluorescent (native Kaede) and red or yellow-fluorescent (photoswitched Kaede). Solid bars represent averages from control fish and lightly shaded bars from DPI-treated fish. Although all the infected, DPI-treated, fish died by 18 hpi, uninfected, DPI-treated, fish do not die due to this treatment alone (data not shown and Fig. 5D). (E) Movement of photoswitched macrophages from the hindbrain was tracked for two classes of Kaede+ macrophages: those that had internalized fungi and those that did not phagocytose fungi. A best-fit line for each population shows that half of those that did not phagocytose fungi have left by approximately 18 hpi, whereas there is no appreciable emigration of macrophages with internalized yeast from the hindbrain in this time. (D–E) Shown are the averages per fish pooled from three independent experiments of at least two fish per group.

Figure 2. NADPH oxidase is required for efficient phagocyte recruitment and containment of C. albicans.

(A–F) CAF2-dTomato C. albicans or were microinjected into the hindbrain ventricle of (A) Tg(mpx:GFP)ⁱ¹¹⁴ (n = 11) or (B) Tg(mpeg1:GAL4/UAS:Kaede) (n = 15) prim25 stage larvae. Fish were treated with DMSO (vehicle) or DPI from two hours prior to infection to 4 hpi and imaged from 0 to 4 hpi by confocal microscopy. (A–B) Images represent at least three independent experiments with three fish of each group per experiment. Scale bars = 10 µm. (C,D) Tg(mpx:GFP)ⁱ¹¹⁴ fish with GFP-expressing neutrophils were used. (E–F) Tg(mpeg1:GAL4/UAS:Kaede) fish with Kaede-expressing macrophages were used. (C, E) Phagocytes were counted at the site of infection at 4 hpi. Total phagocytes includes all EGFP⁺ phagocytes (both with and without yeast) and EGFP⁻ phagocytes that internalized fungi. Phagocytes with Candida includes only phagocytes with internalized fungi. Data from at least three independent experiments was pooled and the average and standard error of all fish are shown. (D, F) At 4 hpi, fungi were scored as intracellular or extracellular, and the % internal was calculated per fish. Average and standard error are shown. *p<0.05 **p<0.01.

Figure 3. Phox morphants have impaired neutrophil migration.

(A–B) CAF2-dTomato C. albicans were microinjected into the hindbrain ventricle of Tg(mpx:GFP)ⁱ¹¹⁴ control (n = 7) or p47^phox morphants (n = 8) and imaged for 4 hours. (A) Representative time-lapse images of control and p47^phox morphants. Results are representative of at least three experiments with at least two fish per group per experiment. Scale bars represent 10 µm. (B) Total phagocytes and phagocytes with internalized fungi were counted at 4 hpi. Numbers from fish over three experiments were pooled and mean and standard error per fish are shown. *p<0.05 **p<0.01 as calculated by two-tailed Student's T-test.

Figure 4. Duox knockdown phenocopies p47^phox knockdown and DPI treatment.

(A–F) Control and Duox morpholinos were co-injected with p53 morpholino into 1-cell stage Tg(mpx:GFP)ⁱ¹¹⁴ zebrafish embryos to create morphants. (A) Samples were collected and prepared for RT-PCR verification of morpholino knockdown. A 39 base pair deletion in the duox message is observed in morphants. (B) Basal level of neutrophils in head at time of infection. Control and duox morphant Tg(mpx:GFP)ⁱ¹¹⁴ fish were injected at the prim25 stage with PBS to simulate infection. At 1 hpi and 4 hpi neutrophils in the head were counted at 1 hpi; n = 68 controls and 77 duox morphants, at 4 hpi n = 70 controls and 71 duox morphants. Data pooled from 5 independent experiments. (C) Basal levels of neutrophils in caudal hematopoetic tissue (CHT). Control and duox morphant Tg(mpx:GFP)ⁱ¹¹⁴ fish were PBS-injected at the prim25 stage and imaged at 1 hpi and 4 hpi. Data shown are representative of three independent experiments; n = 16 control and n = 13 duox morphants. (D–F) Phagocyte migration. Control (n = 15) and duox (n = 15) morphant Tg(mpx:GFP)ⁱ¹¹⁴ fish were injected at the prim25 stage with CAF2-dTomato and imaged until 4 hpi. (D) Representative images of infection site show severe reduction in phagocytosis and extensive extracellular filamentous growth in duox morphants (top) compared with controls (bottom). Scale bar = 10 µm. Representative of three independent experiments. (E) Phagocytes were counted at the site of infection at 4 hpi. Total phagocytes includes all EGFP+ neutrophils and EGFP- phagocytes that internalized fungi. Phagocytes with Candida includes only phagocytes with internalized fungi. Data from all experiments was pooled and the average and standard error of all fish are shown. (F) At 4 hpi, fungi were scored as intracellular or extracellular, and the percent internal was calculated per fish. Data from three independent experiments were pooled and the average per fish and standard error are shown. *p<0.05 **p<0.01 ***p<0.001.

Figure 5. Weak early immune response permits filamentous growth and promotes pathogenesis.

(A–F) Tg(mpx:GFP)ⁱ¹¹⁴ fish with green fluorescent neutrophils were infected with CAF2-dTomato at the prim25 stage, imaged at 4 hpi by microscopy to quantify phagocytosis efficiency, then sorted into individual wells in a 24-well dish and scored for survival at 24 hpi. (A–C) Low and high responders were quantified at 4 hpi by microscopy. Low responders had five or greater extracellular fungi at 4 hpi. (A,C) Data shown are means from three independent experiments that were analyzed by two-tailed Student's T-test. (B, E) Data shown are pooled from three independent experiments, as the number of total fish was too small (n = 6) for similar statistical analysis. (D–F) The percent of high or low responders that died by 24 hpi was quantified. Data shown are either the averages and standard errors (D and F) or were pooled (E) from three independent experiments. Statistical analysis was performed by Student's T-test (D and F). (A, D) Infected fish were treated with DMSO (n = 131) or DPI (n = 118) from 2 hours pre-infection to 4 hpi. (B, E) Control (n = 7) or p47^phox (n = 8) morphants were infected and followed. (C,F) Control (n = 129) or duox (n = 139) morphants were infected and followed. (G) High responders (left) contain the infection by 24 hpi, in contrast to low responders (right) which permit filamentous growth and more frequently die by 24 hpi. Images are representative of three independent experiments. Scale bar = 50 µm. N/A = not applicable because no fish in this category. *p<0.05, **p<0.01.

Figure 6. The EDT1-dependent morphogenetic switching pathway plays roles in NADPH oxidase-independent phagocyte migration and virulence.

(A–E) prim25 stage Tg(mpx:GFP)ⁱ¹¹⁴ larvae were injected with CAF2-dTomato, edt1Δ/Δ-dTomato, or edt1Δ/EDT1-dTomato C. albicans. Fish were incubated with DMSO or DPI from two hours pre-infection until 4 hpi, and imaged by confocal microscopy. At 4 hpi phagocyte migration and phagocytosis were quantified and fish were sorted into high- and low-responder categories, and at 24 hpi fish were scored for survival. (A) Images are representative of three independent experiments; n = 6 for each fungal genotype and treatment group. Time-lapse images from edt1Δ/EDT1 infections are indistinguishable from wildtype infections but are not shown due to space considerations. Scale bar = 50 µm. (B) Quantitation of total phagocyte response and number of phagocytes with internalized fungi shows the average and standard error from fish pooled from three independent experiments; n = 6 for each fungal genotype and treatment group. (C) Phagocytosis efficiency was measured in fish pooled from three independent experiments; n = 6 for each fungal genotype and treatment group. (D) Survival percentage of low- and high-responders was measured for each of three independent experiments, and the means are shown; n = 70–74 for each group. (E) Low mortality of low-responders infected with yeast-locked edt1Δ/Δ mutant. Means and standard errors of mortality percentages of low- and high-responders at 24 hpi are shown. (B–E) Data represent three independent experiments. Means and standard errors are shown. Differences between groups were assessed by two-tailed Student's T-test (B, C, and E) or Fisher's exact test (D). *p<0.05, **p<0.01, ***p<0.001, n.s. no significant difference.

Figure 7. A revised model of NADPH oxidase activity in the context of C. albicans infection.

(A) Naturally, the majority of larvae mobilize phagocytes rapidly to the site of infection and phagocytose nearly all fungi by 4 hpi, as represented by the upper pathway. A minority of larvae fail to mobilize enough phagocytes to internalize most fungi by 4 hpi, as represented by the lower pathway. NADPH oxidase activity blockade by chemical treatment or gene knockdown (green) shifts this balance towards the lower pathway, in which germination of C. albicans extracellularly leads to tissue invasion and death by 24 hpi. NADPH oxidase production of ROS (blue) is required early for phagocyte recruitment and late for attacking fungi and causing oxidative stress. (B) Phagocyte mobilization in response to infection with the yeast-locked edt1Δ/Δ mutant does not depend on NADPH oxidase-produced ROS (blue). In the mutant infections there may be additional chemoattractants produced and/or a failure to block ROS-independent recruitment. This mutant is yeast-locked (red), and therefore does not germinate extracellularly, invade or cause high levels of mortality.