Macrophages Respond Rapidly to Ototoxic Injury of Lateral Line Hair Cells but Are Not Required for Hair Cell Regeneration

Abstract

The sensory organs of the inner ear contain resident populations of macrophages, which are recruited to sites of cellular injury. Such macrophages are known to phagocytose the debris of dying cells but the full role of macrophages in otic pathology is not understood. Lateral line neuromasts of zebrafish contain hair cells that are nearly identical to those in the inner ear, and the optical clarity of larval zebrafish permits direct imaging of cellular interactions. In this study, we used larval zebrafish to characterize the response of macrophages to ototoxic injury of lateral line hair cells. Macrophages migrated into neuromasts within 20 min of exposure to the ototoxic antibiotic neomycin. The number of macrophages in the near vicinity of injured neuromasts was similar to that observed near uninjured neuromasts, suggesting that this early inflammatory response was mediated by "local" macrophages. Upon entering injured neuromasts, macrophages actively phagocytosed hair cell debris. The injury-evoked migration of macrophages was significantly reduced by inhibition of Src-family kinases. Using chemical-genetic ablation of macrophages before the ototoxic injury, we also examined whether macrophages were essential for the initiation of hair cell regeneration. Results revealed only minor differences in hair cell recovery in macrophage-depleted vs. control fish, suggesting that macrophages are not essential for the regeneration of lateral line hair cells.

Keywords: hair cell; inner ear; lateral line; macrophage; ototoxicity; regeneration; zebrafish.

Figure 1

Distribution of macrophages in the posterior lateral line of larval zebrafish. (A,A′) Tail regions of two Tg(mpeg1:yfp) zebrafish, fixed at 6 days post-fertilization (dpf). YFP in macrophages is shown in green, hair cells are labeled for otoferlin (red), and all nuclei are stained with DAPI (gray). Arrows show the four posterior-most neuromasts in each fish. Note that neuromasts possess 1–2 nearby macrophages. (B,B′,B”) High magnification views of neuromast L5 from three zebrafish. Arrows point to nearby macrophages.

Figure 2

Macrophage response to neomycin ototoxicity. Images are maximum-intensity projections of confocal z-stacks of 15 μm depth. Uninjured neuromasts (“Control”) usually possess 1–2 adjacent macrophages (green). Macrophages enter the neuromasts in response to treatment in 100 μM neomycin. Images in the center and right show macrophages at 10 and 20 min of exposure to neomycin. Some macrophages (arrow, 10 min) extended processes that contact hair cells (red, otoferlin), while other macrophages had internalized the debris of dead hair cells (arrowhead, 20 min).

Figure 3

Detailed time course of macrophage response to neomycin ototoxicity. Fish (6 dpf) were incubated in 100 μM neomycin. Some fish were fixed after 5, 10, 20, or 30 min of neomycin treatment, while other fish were removed from neomycin after 30 min, and rinsed and maintained in fresh embryo medium (EM) for 1 or 2 h. Fixed specimens were processed for immunolabeling of YFP-expressing macrophages (green) and hair cells (otoferlin, red), and nuclei were labeled with DAPI (blue). Images in panel (A) are maximum intensity projections, while images in panel (B) are single z-sections taken from 15 μm-depth confocal stacks. (A) Formation of pyknotic nuclei in response to neomycin treatment. Images show DAPI labeling in an untreated neuromast and in neuromasts that were exposed to neomycin for 10 and 30 min. Neomycin induces the formation of pyknotic nuclei (arrowheads), which are assumed to be dying hair cells. (B) Single z-stack sections at various time intervals after initiation of neomycin treatment. Arrows in all images indicate evidence for macrophage phagocytosis of dying hair cells. A few specimens displayed a macrophage response after only 5 min of neomycin exposure. Note that the macrophage in the 5 min image has internalized otoferlin-labeled debris (arrow). At later time points, macrophages had engulfed both otoferlin-labeled debris as well as pyknotic nuclei (arrowheads in 10, 20, and 30 min images). Also, beginning at 20 min of neomycin exposure, the number of otoferlin-labeled hair cells was significantly reduced. Neuromasts at 30 min exposure and 1 and 2 h recovery typically contained 0–3 surviving hair cells. (C) Surviving hair cells as a function of exposure/post-exposure time. Intact hair cells were identified by healthy nuclei that were completely enclosed by the otoferlin-labeled hair cell membrane. A significant decrease in hair cell numbers was observed, beginning after 10 min of neomycin exposure (***p = 0.019, **p < 0.0001, Tukey’s multiple comparisons test). (D) Changes in the numbers of pyknotic nuclei plotted as a function of neomycin exposure/post-exposure time. The number of pyknotic nuclei/neuromast became significantly elevated after 20 min of neomycin exposure (**p < 0.0001). The time course of the neomycin-induced formation of pyknotic nuclei parallels the death of hair cells. (E) Contacts between macrophages and hair cells increased after 20 min of neomycin treatment (***p = 0.0038, one-way ANOVA) and remained elevated until 60 min after treatment. (F) Macrophage phagocytosis of hair cell debris (i.e., internalization of otoferlin-labeled material and/or pyknotic nuclei) was also increased, beginning after 20 min of neomycin treatment (***p < 0.0044, one-way ANOVA).

Figure 4

Rapid externalization of phosphatidylserine (PtS) in response to neomycin treatment. Larval zebrafish (6 dpf) were incubated in Alexa 555-conjugated annexin V and neomycin was added to the water, for a final concentration of 100 μM. Fish were euthanized and fixed after 90 s, 2, 5, or 10 min of exposure to neomycin. Annexin V labeling was observed on the stereocilia bundles of most neomycin-treated fish (e.g., arrow, 90-s image, each red structure represents a single stereocilia bundle), and indicates the presence of PtS on the outer membrane surface. We observed annexin V binding as early as 90 s after the initiation of neomycin treatment (**p = 0.0045, one-way ANOVA), and the number of annexin V-labeled cells increased with longer exposure times (***p < 0.0001, one-way ANOVA). Annexin V labeling was limited to the apical (external) surface of hair cells, as confirmed by side view reconstruction of confocal stacks (arrow, “side view”). We observed no labeling on the basal hair cell surface, which is located within the epidermal barrier of the fish. We also did not observe annexin V binding on any other cells within the body of the fish, suggesting that annexin V does not become internalized. Plotted data were collected from neuromasts L4/5, for a total of 17–40 neuromasts/time point.

Figure 5

Inhibition of Src-family kinases reduced macrophage entry into injured neuromasts. Larval zebrafish were treated in 20 μM PP2, an inhibitor of several Src kinases. Control fish were treated in parallel with 0.1% DMSO. After 1 h pretreatment, neomycin was added to the water of both treatment groups (for a final concentration of 100 μM), and all specimens were euthanized and fixed after 30 min of neomycin exposure. (A–C) Single z-section images taken from 15 μm-depth confocal stacks. Arrows in panel (B) indicate macrophage phagocytosis of dying hair cells. (D–F) Quantification of macrophage activity. Normal numbers of macrophages were present near neuromasts L4/5 in all fish (D). Fish that were treated with 0.1% DMSO displayed a normal macrophage response to neomycin injury, but pretreatment with PP2 resulted in fewer macrophage contacts with dying hair cells (E) and reduced numbers of phagocytic events (F). Data were obtained from 34 fish for each treatment group/time point.

Figure 6

Selective depletion of macrophages in Tg(mpeg1:Gal4FF/UAS:NTR-mCherry) fish. Experiments utilized a transgenic fish line that expresses the bacterial enzyme NTR under the regulation of the Mpeg1 promoter (i.e., selectively in macrophages and microglia). Treatment of these fish with 0.1% DMSO did not affect the numbers of macrophages distributed within the posterior-most 500 μm of the spinal column (A,C). However, treatment for 24 h in 10 mM MTZ resulted in the nearly complete elimination of macrophages from this same region (B,C). Macrophage numbers remained low after 48 h recovery from MTZ treatment (C). Data obtained from 10 to 15 fish/treatment group.

Figure 7

Depletion of macrophages had minimal impact on hair cell regeneration. Tg(mpeg1:Gal4FF/UAS:NTR-mCherry) fish were treated for 24 h in either 10 mM MTZ or 0.1% DMSO. All fish were then rinsed and incubated for 30 min in 100 μM neomycin. Fish were rinsed again and maintained for an additional 2 h or 48 h. Hair cells and macrophages were immunolabeled and hair cell numbers were quantified from neuromast L5 and the two terminal neuromasts. Images from MTZ-treated fish (A) show the absence of macrophages, but otherwise normal levels of neomycin injury and subsequent regeneration. (B) Normal ototoxicity, macrophage recruitment (arrows), and regeneration in DMSO-treated fish. Quantitative data (C,D) show that hair cell injury and regeneration were nearly-identical in normal and macrophage-depleted fish, although the number of regenerated hair cells in the two terminal neuromasts of MTZ-treated was slightly reduced when compared to DMSO-treated controls (*p = 0.0012; n.s.=p > 0.05).