Characterization of mouse melanocytes reveals ultrastructural and immunological insights into the inner ear function

Abstract

Melanocytes are important components of the inner-ear cellular architecture. However, limited morphological research hinders our understanding of inner-ear function. Here, we characterize the morphology of inner-ear melanocytes and cells often misidentified as melanocytes. Immunofluorescence, smart-seq, transmission/field emission scanning electron microscopy, and immunoelectron microscopy are used. Cells previously referred to as “perivascular macrophage-like melanocytes” are not observed in the stria vascularis and are actually macrophages; along with melanocytes, they constitute intermediate cells. Cells with a “black ball” appearance in the vestibule are identified as macrophages. We examine variation in melanocytes or macrophages with age, strain, and cisplatin injury. Kir4.1 expression and the greater noise resistance observed in pigmented mice suggest melanocyte functions. Based on melanin distribution in Pou3f4^y/- mice, we hypothesize that melanocytes migrate from the modiolus along Reissner’s membrane area to the stria vascularis, following a base-to-apex gradient. These findings provide novel ultrastructural and immunological insights into inner-ear function.

Electron microscopy and single-cell RNA sequencing analysis help distinguish melanocytes and macrophages in the inner ear, which are often misidentified.

Introduction

Pigments were first observed in the inner ear of cattle and sheep in 1851; however, they were not conclusively identified as melanin until 80 years later¹. In the 1970s, electron microscopy enabled the visualization of melanin, and by the early 2000s, researchers proposed the existence of “perivascular-resident macrophage-like melanocytes (PVM/Ms)” in the stria vascularis of the cochlea². However, the role of melanin in otology remains unclear.

Melanin is extensively expressed in the cochlea and vestibule, primarily in regions with high metabolic and secretory activities, such as the stria vascularis and dark cell areas of the vestibular labyrinth. It is produced by melanocytes derived from neural crest cells during development^3–6. Disruptions in neural crest-derived cell migration or melanin synthesis may impair inner ear function. Sensorineural hearing loss and vestibular disorders are associated with pigmentary abnormalities in Vogt–Koyanagi–Harada disease and Waardenburg and Tietz syndrome^7,8. Hearing impairment has also been observed in patients with vitiligo and melanoma^9,10.

Our observations in black C57BL/6 J mice revealed that some cells in the stria vascularis and vestibular regions were misidentified as melanocytes^2,11,12. Not all cells containing melanin granules are melanocytes; these cells, particularly macrophages, are often difficult to distinguish from one another. Previously, researchers have found that perivascular macrophages in the cochlear stria vascularis contain “melanin” and classified them as a type of composite cell, referring to them as “PVM/Ms” ². However, we questioned this designation. Cochlear macrophages are immune cells derived from the yolk sac or bone marrow, whereas stria vascularis melanocytes originate from the neural crest. Considering the different developmental lines and phenotypes of macrophages and melanocytes, a composite of these two cell types is unlikely to exist in the inner ear stria vascularis. In this study, we observed a population of cells with a characteristic “black ball-like” appearance in the vestibular region. These cells contain dense melanin granules at various stages of melanogenesis, closely resembling melanocytes in morphology. However, they were not clearly labeled with the melanocyte antibody Melan-A, suggesting that they may not be true melanocytes.

In this study, we aimed to characterize these cell types using immunofluorescence (IF), smart-seq, transmission electron microscopy (TEM), immunoelectron microscopy (IEM), and field-emission scanning electron microscopy (SEM). We also examined variations in melanocytes and macrophages according to age, strain, and cisplatin injury. Our findings revealed that “PVM/Ms” were not observed in the stria vascularis and were macrophages. With aging, melanocyte features in the stria vascularis became less distinct, while the number of vestibular melanocytes decreased. Concurrently, macrophage activation increases melanosome phagocytosis, which was marked by the presence of abundant “black ball-like” macrophages. Melanocytes expressed Kir4.1, and melanin played a protective role in the inner ear against noise-induced damage. Furthermore, we proposed hypotheses regarding the migration pathways of melanocytes in the cochlea during inner-ear development, based on the melanin distribution in Pou3f4^y/- mice.

Results

Melanocytes and Macrophages in the Cochlear Stria Vascularis

Fontana–Masson (F–M) staining revealed that the pigments were distributed in the cochlear stria vascularis (Fig. 1A), appearing black (arrow) and brown (arrowhead). TEM revealed substances with high electron density in the stria vascularis, including lysosomal remnants (Fig. 1B and Supplementary Fig. 1A), and melanosomes (Fig. 1C, D; arrows). The lysosomal remnants were amorphous, uneven, and high-density granules that could easily be mistaken for melanin.

Fig. 1. Melanocytes and macrophages in the cochlear stria vascularis of mice are independent cell types.

A F–M staining showing black granules (arrow) and brown granules (arrowhead) in the stria vascularis. B–D Pigments in the stria vascularis observed under TEM. B Lysosomal remnants are amorphous, uneven, and high-density granules. C The first type of cells containing melanosomes has an “octopus-like” appearance with multiple protrusions. Melanosomes at various stages are sparsely scattered (arrows), and mature granules are exocytosed across the cell membrane. Arrowhead indicates the mitochondria. D The second type of melanosome-containing cells exhibit clustered melanosomes associated with a gray halo, distributed in masses, which are often surrounded by a membrane. Arrowheads indicate the envelope. The low-magnification image is shown in Supplementary Fig. 1. E,F IEM showing a representative image labeled with CD68 (E) and Melan-A (F) gold particles (arrowheads). Arrows indicate melanosomes. G IF staining showing that macrophages (CD68 and Iba1) and melanocytes (Melan-A) are independent cell types. V, blood vessel; N, cell nucleus; Ly, lysosome. Macrophages are outlined in blue, and melanocytes are outlined in pink. Scale bars: A 200 μm; B–E 1 μm; F 2 μm; G 100 μm.

Based on their morphology, two types of cells containing melanosomes were identified in the stria vascularis. One cell type exhibited a relatively low cytoplasmic electron density, extended multiple slender pseudopodia, and had an octopus-like appearance. Sparse and scattered melanosomes were observed in the cytoplasm (Fig. 1C and Supplementary Fig. 1B). The cells were initially identified as melanocytes. Furthermore, IEM showed that the cells were negative for CD68 (specific for macrophages, Fig. 1E) and positive for Melan-A (specific for melanocytes, Fig. 1F), confirming their identity as melanocytes (outlined in pink). The other cell types exhibited slightly higher cytoplasmic electron density. Melanosomes were observed at all stages, distributed in clusters with a grayish halo as the background, and often surrounded by a layer of membrane (Fig. 1D, arrowheads). The cells were usually located adjacent to the blood vessels, and we speculated that they were macrophages (Supplementary Fig. 1C). Subsequent IEM revealed that the cells were CD68 positive (Fig. 1E), Melan-A negative (Fig. 1F), confirming their identity as macrophages (outlined in blue). Notably, in macrophages, CD68 gold particles were densely distributed along the unit membrane of lysosomes and did not mix with melanosomes, suggesting that melanosomes were phagocytosed. CD68, Iba1, and Melan-A IF staining (3D) clearly labeled the macrophages and melanocytes. The fluorescence signals of the two cell types did not overlap (Fig. 1G), and “Kir4.1” was labeled in the non-macrophage area (Fig. 2D, F). This suggests that macrophages and melanocytes are independent cell types, and no composite cell type such as “PVM/Ms” was observed.

Fig. 2. Cochlear stria vascularis intermediate cells are composed solely of macrophages and melanocytes.

A IF staining shows the distribution of Met-positive cells. B IEM shows that melanocytes (outlined in pink) are labeled with Met gold particles (arrowheads), whereas macrophages (outlined in blue) are not labeled. The arrows indicate melanosomes. C IF staining shows the distribution of Nrp2-positive cells. D IEM shows that melanocytes (outlined in pink) are labeled with Kir4.1 gold particles (arrowheads), whereas macrophages (outlined in blue) are not labeled. The arrows indicate melanosomes. E Schematic diagram of cell types in the stria vascularis and the staining characteristics of Kir4.1 photographed at three levels a, b, and c (F). V, blood vessel; Ma, macrophage; Me, melanocyte; M, marginal cell; B, basal cell; F, fibroblast. Scale bars: (A, C, and F) 100 μm; (B, D) 2 μm.

Single-cell sequencing studies have reported that, after the removal of macrophages and vascular accessory cells, the “strictly defined intermediate cells” in the mouse stria vascularis were a single type of cell expressing Met and Nrp2¹³. IF revealed that the distribution of Met- and Nrp2-positive cells was similar to that of melanocytes (Fig. 2A, C). Furthermore, IEM confirmed that the Met-positive cells were melanocytes (Fig. 2B, outlined in pink). Therefore, we concluded that intermediate-layer cells, often referred to as intermediate cells, are composed solely of macrophages and melanocytes (Fig. 2E).

Vestibular Melanocytes and “Black Ball-like” Cells

Under bright-field conditions, a noticeable difference between the vestibular regions of young and suckling mice was the presence of more black masses in areas such as the wall of the utricle (Fig. 3A, B). In young mice, F–M staining revealed a large amount of melanin in the wall of the utricle, transitional area of the ampulla, and membranous semicircular canals (Fig. 3C–F). Most melanin was present in a scattered dot-like pattern with irregular shapes and occasionally formed lines (arrowheads). Large black masses were sporadically scattered (arrows). Under light microscopy, these masses were regularly round or oval and uniformly dense, with diameters of 7–25 µm. These “black ball-like” masses were distributed at the perilymphatic interface or within the nerve and vascular plexuses of the ampulla and utricle (Fig. 3C–E, arrows). Surprisingly, these cells were not labeled with Melan-A (Fig. 3G, arrows). No melanin distribution was observed in the saccule (Fig. 3E). A small number of melanin granules were observed on the wall of the endolymphatic sac near the intracranial surface, which appeared to be a continuation of meningeal pigment (Fig. 3F, arrowhead).

Fig. 3. Vestibular melanocytes and “black ball-like” cells in mice.

A, B Photographs of the inner ear of the dissected suckling mice (A) and young mice (B). Arrows indicate the utricle wall, and black masses appear in young mice. C–F F–M staining showing melanin granules in the vestibular ampulla (C), utricle and saccule (E), semicircular canals (D), and endolymphatic sac (F). One type of granule is distributed in a scattered dot-like pattern (arrowheads), whereas the other is visible as a large black mass (arrows). G IF shows that the black masses in the ampulla are not labeled by Melan-A (arrows). The arrowhead indicates a melanocyte. H–N Vestibular melanocytes (outlined in pink) and “black ball-like” cells under TEM. On the endolymphatic side of the basement membrane, cells containing melanin granules can be seen (I, outlined in brown, arrow), and melanosomes can be seen in the endolymphatic and perilymphatic cavities (K and L, arrows). J Epithelial cells extend across the basement membrane to the perilymphatic side (outlined in brown). O, P SEM shows that the “black ball-like” cells are distributed independently, and melanosomes are in a “cocoon” -like shape. V, blood vessel; N, cell nucleus; BM, basement membrane; S, saccule; U, utricle; **, Otolith. Scale bars: (C–G) 100 μm; (H) 1 μm; (I–P) 2 μm.

Melanin was detected using TEM, and a cell type located far from the basement membrane and close to the perilymphatic interface was identified. The cell body was round or oval with few protrusions on the cell membrane. It contains many dense melanosomes at different stages¹⁴, which occupy most of the cell body volume. The nucleus and a small number of organelles were often located on one side, and the cytoplasmic matrix exhibited relatively high electron density. Its spatial distribution was consistent with the “black ball-like” masses observed in the F–M experiments. Therefore, we named it a “black ball-like” cell and considered it a macrophage (Fig. 3M, N, arrows). Another type of cells observed was classic melanocytes. These cells were numerous and spatially close to the basement membrane and differed from the above-mentioned “black ball-like” cells. Specifically, most cells were flat or fusiform, with protrusions distributed along the basement membrane and a relatively low cytoplasmic electron density (Fig. 3H, M, N, outlined in pink). Scattered melanosomes were distributed in the cytoplasm; some immature melanin granules (stage II/III) were often close to the periphery, and some cells exocytosed melanosomes (Fig. 3K and Supplementary Fig. 2, arrows). On the endolymphatic interface of the basement membrane, a small number of cells contained melanosomes (Fig. 3I, outlined in brown). Epithelial cells on the endolymphatic side extended pseudopodia with numerous villus-like protrusions, penetrated the basement membrane, and extended to melanocytes on the perilymphatic side. They appeared to assist in the transport of melanin granules to the endolymphatic cavity surface (Fig. 3J, outlined in brown). Melanosomes were occasionally observed at the endolymphatic interface (Fig. 3L, arrow). Unlike the melanosomes in the eye¹⁵, which exist in various sizes and shapes, those in the inner ear are uniform and consistent.

“Black ball-like” cells were particularly conspicuous under SEM. They were distributed independently in a spherical shape with numerous thin sheet-like protrusions densely covering their surface. The melanosomes were densely arranged and tightly covered by the cell membrane, outlining an oval contour, like a “cocoon.” Slightly flattened cells were often observed at the bottom of the “black ball-like” structure. The cell membrane covered the loosely arranged “cocoons” inside them. These cells were melanocytes that were often connected to the sheets. On their surfaces, amorphous cells with thin sheet or strip-like protrusions were scattered (Fig. 3O, P).

“Black Ball-like” Cells were Macrophages

Vestibular macrophages co-expressed F4/80, CD68, and Iba1, with strong CD68 signals in “black ball-like” cells, suggesting their identity as macrophages. Melanosomes in these “black ball-like” cells were significantly dense, obscuring some fluorescence (F4/80 and Iba1) at the visual resolution level, resulting in weak detection of both markers (Fig. 4A). Co-staining for CD68 and Melan-A revealed that the “black ball-like” cells were CD68-positive and Melan-A-negative (Fig. 4B–D, arrows).

Fig. 4. “Black ball-like” cells in the inner ear of mice labeled with CD68.

A IF shows confocal images of “black ball-like” cells (arrows) expressing a strong fluorescent signal of CD68. B–D IF staining shows CD68 (red, arrows) and Melan-A (green, arrowheads)-positive cells in ampulla (B), utricle (C), and semicircular canals (D). E Representative image labeled with CD68 gold particles (arrowhead). The arrows indicate various stages of melanosomes. F Representative image labeled with Melan-A gold particles (arrowheads). Melanocyte is outlined in pink, and “black ball-like” cell is outlined in yellow. G,H “Black ball-like” macrophages are also distributed on the Reissner’s membrane (G, arrow) of the cochlea. H Corresponding TEM image. **, Otolith; N, cell nucleus. Scale bars: (A–D, and G) 100 μm; (E) 2 μm; (F) 5 μm; (H) 1 μm.

IEM confirmed that the “black ball-like” cells were CD68-positive, with weak Melan-A expression (Fig. 4E, F, and Supplementary Fig. 3, outlined in yellow). Unlike in the stria vascularis, CD68 gold particles were cytoplasmically dispersed (arrowhead). Notably, early melanin granules have also been detected (arrows). Furthermore, examination of the cochlear structure revealed that “black ball-like” macrophages were also distributed at the scala vestibuli interface of Reissner’s membrane (Fig. 4G, H).

Next, we used smart-seq, an RNA-sequencing technique, to analyze the whole transcriptome of the “black ball-like” cells. A marker gene heatmap of melanocytes/macrophages revealed high expression of genes such as CD68, C1qb, C1qa, C1qc, IL18, DCT, and Mlana (Melan-A) (Fig. 5B). We then focused on the key words “Tyrosine,” “Melanocyte,” “Melanin,” “Immune”, “Lysosome,” “Macrophage,” and “Vesicle fusion to plasma membrane.” Gene expression profile-related functions were evaluated using GO and KEGG enrichment analyses. Among the top 14 GO terms, “macrophage activation” and “melanin metabolic process” were the most significantly expressed (Fig. 5C). Among the top 10 KEGG terms, “lysosome,” “efferocytosis,” and “synaptic vesicle cycle” were significantly expressed (Fig. 5D). These findings suggest that the “black ball-like” cells may be macrophages.

Fig. 5. Smart-seq and scRNA-seq combined analyses revealed that the “black ball-like” cells in the inner ear of mice are macrophages.

A Cell sampling of smart-seq. B Marker gene heatmap (log2 (fpkm+1)) of melanocytes and macrophages from “black ball-like” cells by smart-seq. C GO enrichment analysis of expressed genes using smart-seq. D KEGG enrichment analysis of signaling pathways using smart-seq. E UMAP of scRNA-seq datasets (GSE168901) from P3 and P7 mouse ampullae; clustering identified macrophages, melanocytes, and “hybrid” cells. F Distribution of the corresponding cells from the GSE dataset and the three groups of cells measured in this study are visualized in the UMAP plot. G Dotplot of macrophage-related and melanin-related gene expression in scRNA-seq datasets (GSE168901). The cells labeled as “hybrid” cells expressed a combination of macrophage markers and melanocyte markers. H Expression of known marker genes (CD68 and Mlana) in each major cell group detected by scRNA-seq (GSE168901).

To further clarify the nature of “black ball-like” cells, we used both our smart-seq results and the published single-cell RNA-seq (scRNA-seq) database data (GSE168901)¹⁶ for a combined analysis and clustered these data to detect the cell types. In the same-strain mouse ampullae (P3 and P7) from the published database, we labeled a “hybrid” cell type, which expressed a combination of macrophage markers (CD68, C1qa, C1qb, C1qc, and Aif1), melanocyte markers (Dct and Mlana), and the inflammatory cytokine IL-18 (Fig. 5E, G, H and Supplementary Fig. 4), and whose characteristics were consistent with those of our “black ball-like” cells (Fig. 5B, F). Macrophage marker genes were expressed more predominantly; thus, we believe that the “black ball-like” cells are more likely to be macrophages and that components of phagocytosed melanocytes may have influenced these results.

Overall, inner ear macrophages exhibited a unique “black ball-like” appearance at the perilymphatic interface (Fig. 6).

Fig. 6. Schematic illustration of the distribution of macrophages and melanocytes in the inner ear of mice.

Macrophages are located at the perilymphatic interface of the inner ear cavities and within the parenchyma of inner ear tissues. Melanocytes are located in the stria vascularis of the cochlea (Co), utricle (U), and ampulla (A). No melanocytes are observed in the saccule (S). Macrophages and melanocytes are two independent cell types in the cochlear stria vascularis. Therefore, it is important to avoid misidentifying macrophages as melanocytes. The “black ball-like” macrophages in the vestibular region are easily mistaken for melanocytes and are mainly distributed at the perilymphatic interface or within the nerve and vascular plexuses of the A, U, and semicircular canals.

Variations in Patterns of Macrophages and Melanocytes in the Cochlear Stria Vascularis with Age, Strain, and after Cisplatin Injury

F–M staining revealed a significant increase in the number of pigment granules in the old-age and cisplatin-treated old-age groups (Fig. 7B, F), and TEM analysis confirmed that the granules were lysosomes (Fig. 7D, E). IF showed that macrophages in old and cisplatin-treated mice adopted amoeboid shapes, that is, larger bodies with short, thick protrusions (Fig. 7C). In contrast, suckling mice had fewer macrophages with long, branched protrusions and sparse lysosomes (Fig. 7C). Macrophage numbers increased with age and stabilized after the young stage, with minimal strain differences (Fig. 7C, G and Supplementary Fig. 5E).

Fig. 7. Changes in melanocytes and macrophages in the cochlear stria vascularis among different groups.

A Timeline of the experimental protocol. B F–M staining shows changes in pigment granules in the stria vascularis of the basal turn across various groups. C Representative images of IF staining, showing melanocytes (green) and macrophages (red) in the stria vascularis of the basal turn. The number of melanocytes decreases with age, whereas the number of macrophages gradually increases and remains stable after the young stage. Cisplatin injury significantly reduced the number of melanocytes. D,E Characteristics of melanocytes (D) and macrophages (E) in the stria vascularis as observed under TEM. The number of lysosomes increased significantly with age. Melanosomes in melanocytes present granules at all stages of infancy, whereas only stage 4 granules are present as age increases. The number of melanosomes phagocytosed by macrophages significantly increases after cisplatin injury. The low-magnification image is shown in Supplementary Fig. 6. F Counts of black and brown granules in F–M staining images. n = 3–20 sections from 3 mice. G Counts of melanocytes and macrophages in the three turns of the cochlea in IF. n = 5–11 slides from 5 mice. Data were mean ± SEM. *, #, &, % p < 0.05, one-way analysis of variance (ANOVA) compared with P3 (*), 2 M (#), 12 M (&), and Cis (0 M) (%). Symbols 1, 2, and 3 correspond to p < 0.05, 0.01, and 0.001 respectively. Scale bars: (A, B) 100 μm; (C, D) 2 μm.

Generous melanocytes were observed in suckling mice (Fig. 7C). The number gradually decreased with age, particularly in basal turns. Few melanocytes were observed in the old-age and cisplatin groups; however, the number of melanocytes in the cisplatin old-age group was slightly higher than that in the control group, suggesting that melanocytes may play a role in cisplatin injury (Fig. 7C, G and Supplementary Fig. 5E). TEM revealed that suckling mouse melanocytes had abundant organelles and melanosomes, whereas old-age “octopus-like” cells retained their number but showed reduced melanosomes and increased lysosomes (Fig. 7D and Supplementary Fig. 6A). After cisplatin injury, the nuclei of melanocytes enlarged and became round, the cytoplasmic electron density decreased, and fewer melanosomes were observed (Fig. 7D and Supplementary Fig. 6A). The significant increase in the number of melanin granules phagocytosed by macrophages indirectly suggested that melanocytes were activated in the early stage of injury and exocytosed more melanin granules into the intracavitary space (Supplementary Fig. 6B). Numerous melanocytes were observed under TEM in the albino mice, and their distribution was similar to that in the pigmented mice (Fig. 7D and Supplementary Fig. 6A). Melanosomes in albino mice were only present in stages 1 and 2¹⁷. However, our study found that they were also present in stages 3 and 4, albeit in small numbers. This may explain why only a few cells were positive for Melan-A staining. Macrophages in albino mice phagocytosed fewer melanosomes than those in pigmented mice (Fig. 7E and Supplementary Fig. 6B). In contrast to pigmented mice, the patterns of high and low electron densities in melanocytes and macrophages were reversed in the albino mice (Fig. 7D, E and Supplementary Fig. 6).

Variations in Patterns of Vestibular Macrophages and Melanocytes with Age, Strain, and after Cisplatin Injury

F–M and IF staining revealed that vestibular pigments were present since the suckling stage, and gradually decreased with age, along with “black ball-like” macrophages significantly increased and stabilized at 2 months after birth, suggesting that the main function of the “black ball-like” macrophages is to phagocytose decreasing melanocytes and their melanosomes (Fig. 8A–C). In suckling mice, the proportion of melanosomes in macrophages was relatively small. With increasing age, melanosomes gradually filled the cells, suggesting increased phagocytic accumulation. In suckling and young mice, large masses of melanosomes were occasionally surrounded by a single- or double-membrane layer in macrophages (Supplementary Fig. 7A, D).

Fig. 8. Variation patterns of vestibular macrophages and melanocytes of mice.

A F–M staining shows the changes in pigment granules in the A and U. B Representative images of IF staining. The number of melanocytes in the U, A, and SC gradually decreases from suckling mice to young mice with increasing age, whereas the number of “black ball-like” macrophages significantly increases and then remains stable. After cisplatin injury, the number of “black ball-like” cells in the A and SC increases significantly. C Statistics of the number and size of “black ball-like” macrophages in each group under IF. n = 3–27 slides from 5 mice. *, #, &, % p < 0.05 by one-way ANOVA compared with P3 (*), 2 M (#), 12 M (&) and Cis (0 M) (%). D ABR thresholds at all tested frequencies (n = 8 mice per group). Data were mean ± SEM. * indicates p-values for 2 days post versus pre-exposure; # indicates p-values for 2 weeks post versus pre-exposure; & indicates p-values for 2 days post versus 2 weeks post; Symbols 1, 2, and 3 correspond to p < 0.05, 0.01, and 0.001, respectively. A, ampulla; U, utricle; SC, semicircular canals. Scale bars: (A, B) 100 μm.

Following cisplatin injury, the number of “black ball-like” macrophages increased in the ampulla and semicircular canals, indicating rapid activation (Fig. 8B, C). In the old-age and cisplatin-treated old-age groups, the number of lysosomes increased (Supplementary Fig. 7A).

In albino mice, macrophages contained small “steel wool ball” structures, which were likely aggregated stage II melanosomes (Supplementary Fig. 7B); a few early-stage granule-containing cells were observed (Supplementary Fig. 7C). Vestibular melanocytes and macrophages exhibited reversed electron density patterns similar to those observed in the stria vascularis (Supplementary Fig. 7B, C). ABR results showed hearing recovery in pigmented mice at 2 weeks post-noise, but not in albino mice (Fig. 8D).

Distribution Characteristics of Melanin and Macrophages in Pou3f4^y/- Mice

Most researchers have focused on mouse models with reduced pigmentation⁸. We found that, in a mouse model wherein Brn4/Pou3f4 (an important transcription factor that affects mesenchymal formation) was knocked out, there was excessive pigmentation in the vestibular tissue and abnormal pigment distribution in the cochlea. The expression of Brn4 in the inner ear occurs from E10 to 13.5, which overlaps with the time when melanoblasts migrate into the otic vesicle. Therefore, we conducted further studies using this mouse model. In Pou3f4^y/- mice, pigment retention was observed in the modiolus (Fig. 9B, C, arrows), and the stria vascularis was thinner, which may be explained by reduced melanocytes, as revealed by IF and TEM (Fig. 9 G–I). Specifically, almost no melanocytes were found in the apical turn; in the basal turn, some melanocytes were noted to be lying on the Reissner’s membrane. Therefore, we speculated that melanocytes reached the stria vascularis via the Reissner’s membrane area, supporting a base-to-apex migration pattern (Fig. 9J).

Fig. 9. Distribution characteristics of melanin and macrophages in the cochlea of Pou3f4^y/- mice.

A Administration and testing time points. B Bright field microscopy showing pigment retention (arrow) in the modiolus. C F–M staining revealed that pigment retention in the upper part of the modiolus (arrows) and on the basal turn of Reissner’s membrane (arrowheads) in all groups of Pou3f4^y/- mice. D,E Representative images of melanocytes (D) and macrophages (E) in the basal turn of the stria vascularis under TEM. The low-magnification image is shown in Supplementary Fig. 8. F Melanocytes lying on the basal turn of Reissner’s membrane of Pou3f4^y/- mice. G TEM shows that the basal turn of the stria vascularis in Pou3f4^y/- mice is thinner than that in the control group. H IF staining indicates that the number of melanocytes in the stria vascularis of Pou3f4^y/- mice is lower than that in the normal control group. I Statistics on the thickness of the stria vascularis (n = 4–16 sections from 3 mice) and melanocyte count across various groups (n = 6–10 slides from 4 mice). J Possible migration route of melanocytes in the cochlea during inner ear development: melanocytes migrate from the modiolus along Reissner’s membrane area to the stria vascularis following a gradient from the base to the apex. The melanocytes retained in the modiolus or on Reissner’s membrane reduce the number of melanocytes reaching the stria vascularis, resulting in thinning of the stria vascularis. Data were mean ± SEM. * p < 0.05 by one-way ANOVA compared with control; Symbols 1, 2, and 3 correspond to p < 0.05, 0.01, and 0.001, respectively. Scale bars: (C, H) 100 μm; (D–F) 2 μm; (G) 5 μm.

Most melanosomes in the melanocytes of Pou3f4^y/- mice were early-stage granules, and the pigment color was light, accompanied by a relatively large number of lysosomes, particularly in the 2-month group (Fig. 9D, E and Supplementary Fig. 8). Macrophages in the stria vascularis of Pou3f4^y/- mice were transformed into amoeba-like shapes by the 2-month stage (Fig. 9H), similar to the morphology of the old-age group of normal mice, suggesting a premature aging state.

Pigmentation on the vestibular end organs of Pou3f4^y/- mice was significantly darker than that of the control group (Fig. 10A). As the mice aged, there was a significant decrease in the number of melanocytes, along with a significant increase in “black ball-like” macrophages (Fig. 10A, E, and F). Transcriptome sequencing analysis was conducted to determine the characteristics of the increased number of melanocytes in 10-day-old Pou3f4^y/- mice compared with those in the control group. Gene ontology enrichment analysis showed that Pou3f4^y/- mice were mainly enriched in terms of structural development, such as cell adhesion, extracellular matrix, ion transport, and nerve synapses (Supplementary Fig. 9B, C). KEGG enrichment analysis indicated that Pou3f4^y/- mice were enriched in ECM-receptor interaction, focal adhesion, the “PI3K/AKT pathway,” and the “PPAR signaling pathway” (Fig. 10C). We focused on molecules related to melanin synthesis that had changed. The volcano plot shows Adamts9 and Itgb3 upregulation and Adamts20 downregulation (Supplementary Fig. 9A). A quantitative polymerase chain reaction (qPCR) was performed to confirm the mRNA expression levels of the three genes of interest (Fig. 10D).

Fig. 10. Distribution characteristics of melanocytes and macrophages in the vestibules of Pou3f4^y/- mice.

A Changes in vestibular pigment granules under a bright field. There was excessive pigmentation and the number of “black ball-like” macrophages increases significantly with age (arrows). B–D Transcriptomic analysis of Pou3f4^y/- mice compared to normal mice. B Heatmap. C KEGG enrichment analysis of differentially expressed genes (DEGs). Genes with p-adjust < 0.05 and fold change (FC) ≥ 1.5 or FC ≤ 0.6666 are considered as DEGs. The bar chart shows the 15 representative enriched pathways. D mRNA expression levels of DEGs of interest confirmed using qPCR. Data were mean ± SEM. (n = 3, unpaired t-tests). E Representative images of IF staining for A, U, and SC, and representative images of “black ball-like” macrophages under TEM. F Statistics of the number and size of “black ball-like” macrophages in each group under IF. The number of macrophages increased significantly and exceeded that in the control group. N = 4–11 slides from 4 mice. Data were mean ± SEM. *, #, & p < 0.05 by one-way analysis of variance (ANOVA) compared with P10 (*), 4 W (#), 8 W (&); Symbols 1, 2, and 3 correspond to p < 0.05, 0.01, and 0.001, respectively. A, ampulla; U, utricle; SC, semicircular canals. Scale bars: (E) 100 μm for IF images and 2 μm for TEM images.

Discussion

Within the stria vascularis, a group of cells near the blood vessels rich in dark material was observed, appearing as shadows under light microscopy and as brown or black granules after F-M staining. These cells could be easily mistaken for melanocytes. The dark matter was accurately identified using TEM as secondary lysosomes/lysosome residual bodies, which were distinct from melanin. Melanosomes were also observed in these cells, but were often wrapped in lysosomes, indicating that they underwent lysosomal digestion after being engulfed. The combination of positive labeling with the CD68 antibody and negative labeling with the Melan-A antibody on the IEM confirmed that the cells were macrophages. CD68 gold particles were localized to the cytoplasm or lysosomal membranes and did not mix with melanin granules, indicating that melanosomes were phagocytosed. This was particularly evident in stress injuries, such as those caused by cisplatin, in the stria vascularis.

“Octopus-like” cells with multiple protrusions were frequently observed near the basal cells and at the upper and lower ends of the stria vascularis. IEM confirmed that the cells were melanocytes. The internal melanosomes were mostly scattered and less aggregated, making it difficult and incidental to identify melanocytes under light microscopy.

Cable and Steel distinguished two subclasses of stria vascularis intermediate cells: bright and dark¹⁸. We found that bright intermediate cells corresponded to melanocytes and that dark cells corresponded to macrophages. In normal rodents, not every clear intermediate cell contains observable melanosomes; however, in chinchilla, clear intermediate cells have been shown to contain tyrosinase and can therefore be considered melanocytes capable of melanin synthesis¹⁹. Hilding and Ginzberg demonstrated that pigment cells in the stria vascularis are genuine melanocytes rather than melanin host cells that obtain melanosomes from melanocytes²⁰. Using single-cell sequencing, Milon discovered that when macrophages were removed, the intermediate cells were a single-cell type that expressed Met and Nrp2¹³. Our IEM analysis revealed that Met-positive cells were melanocytes. Therefore, we conclude that intermediate cells in the stria vascularis are composed solely of macrophages and melanocytes.

Although Zhang et al.² detected “GSTα4, GST” in the macrophages of the stria vascularis, there is no direct evidence suggesting that GSTα4 and GST are melanocyte biomarkers²¹. Zhang et al. detected Kir4.1 in macrophages of the stria vascularis²; however, we detected Kir4.1 more frequently outside macrophages, suggesting that macrophages differ from Kir4.1-positive melanocytes.

Recently, Renauld et al. used genetic lineage tracing to show that pigment cells that form intermediate cells can differentiate from either the neural crest of the dorsolateral migratory pathway or neuron-related Schwann cell precursors (SCPs)⁶. Adameyko et al. confirmed that cranial melanocytes have a dual embryonic origin⁵. However, the unclear classification of intermediate cells in the stria vascularis hampered the understanding of the dual origin of embryos. Therefore, research on the classification of stria vascularis cells is of fundamental significance.

We concluded that macrophages and melanocytes are two independent cell types that jointly constitute intermediate cells in the stria vascularis. We believe that no composite “PVM/Ms” cell type exists. Accordingly, it is more appropriate to refer to macrophages adjacent to blood vessels as “perivascular-resident macrophages (PVMs).” Macrophages can engulf melanin granules or contain lysosomal remnants, a common reason why they are mistaken for melanocytes.

Vestibular melanocytes form a large network, and epithelial cells on the endolymph side extend pseudopodia, penetrating the basement membrane, seemingly to help transport melanin granules to the luminal surface of the endolymph. Other studies have found that vestibular melanocytes in the dark cell area extend protrusions that penetrate the basement membrane and are inserted between the basal folds of epithelial cells²². In the late fetal stage of human development, interruptions in the basement membrane have also been observed²³. These results suggest a close connection between melanocytes and epithelial cells. The presence of melanosomes in the endolymphatic cavity may be normal secretory products of activated melanocytes or waste disposal products of dead melanocytes that enter the endolymphatic cavity²⁴.

Although the mechanism by which melanosomes are transferred to adjacent cells in the inner ear remains unclear, it may be similar to the four different models of melanosome transfer in the skin²⁵. Notably, melanin granules are transported from the perinuclear area to the periphery of melanocytes in the skin²⁶; however, we found that in vestibular melanocytes, immature melanin granules (stages II/III) are often close to the periphery, and it is uncertain whether they are expelled (Supplementary Fig. 2). This phenomenon may partially explain the appearance of immature granules in “black ball-like” macrophages.

Highly distinctive “black ball-like” cells were observed in the vestibule. Dense and multi-stage melanin granules have often caused these cells to be mistaken for melanocytes in previous studies^11,12. We preliminarily identified these cells as macrophages, indicating that their primary function was to perform phagocytosis. Unlike macrophages in the stria vascularis, in this type of cell, the CD68 gold particles and melanosomes were relatively evenly distributed in the cytoplasm, intertwined, and mixed. Furthermore, membrane structures separating them were less frequently observed, and some melanin granules were occasionally clustered and wrapped in a single- or double-layer membrane, suggesting that the cells engulfed the melanin granules discharged into the tissue gaps or part of the cytoplasm of melanocytes, with a strong digestive ability to quickly eliminate the original envelope structures. As mice age and their digestive functions decline, residual envelopes and lysosomes can be observed around clustered melanin granules in “black ball-like” macrophages. Macrophages at different locations specialize in distinct forms, each with different phagocytic and digestive abilities.

Another explanation for the presence of melanin granules at various stages in the cytoplasm of “black ball-like” cells is that these macrophages can produce melanin. Over the past 50 years, melanin macrophages have been reported in fish, amphibians, and reptiles²⁷. These cells are scavenger cells, similar to melanophages in the eye¹⁵, and melanin is either present owing to phagocytosis or is endogenously synthesized^28,29. Strong evidence is available for melanin production by macrophages in poikilothermic vertebrates³⁰. However, the ability of hematopoietic cells to synthesize melanin remains unclear³¹. Our smart-seq analysis showed that “lysosome” and “efferocytosis” were dominant. Our scRNA-seq combined analysis revealed that cells with “hybrid” characteristics exist in the vestibule, and macrophage-related marker genes were expressed more predominantly. Although melanin-synthesizing genes such as Dct and Tryp1 were expressed, the expression of Mitf was very low. Mitf is considered a master regulator of melanocyte fate³². Our observations do not exclude the possibility that macrophages can phagocytose melanocyte cell bodies or their components. As the decrease in melanocytes coincided with an increase in the number of “black ball-like” cells, we consider that these “black ball-like” cells were macrophages.

We speculate that such macrophages are widely present in the eyes or meninges^4,33. Macrophages filled with melanin granules are present between the retinal pigment epithelium and Bruch’s membrane³⁴ or on the surface of the choroid and iris¹⁵, which are similar to the “black ball-like” cells in the inner ear. However, more detailed characterization has been overlooked and requires further investigation.

Although melanin has long been observed in the inner ear, its function remains unknown. Our results clearly show Kir4.1 expression in inner ear melanocytes, indicating their involvement in K⁺ transport. Melanocytes are crucial for the development and function of the stria vascularis³⁵. Noise exposure, ototoxic drug administration, and aging are external factors that affect cochlear pigmentation^36,37.

Relatively large numbers of melanocytes were observed in the stria vascularis of young mice, suggesting that they may play a role in the development and maintenance of inner ear function. In aged mice, melanocytes in the stria vascularis appear similar to those in the uvea, are similar to pluripotent stem cells, and remain dormant^38,39.

Melanin has a considerable affinity for ototoxic agents. We observed an increase in melanin granules in the cisplatin-treated group. Pigments may protect tissues by binding to potentially harmful substances and slowly releasing substances at low, non-toxic concentrations^36,40–42.

Albino animal models have shown that the inner ear is more susceptible to damage from noise, aging, or gentamicin^43–45. Melanin can bind calcium and “buffer” the influx of calcium during noise exposure. When melanin is lacking, calcium levels in the inner ear may remain high, thereby increasing hair cell damage⁴⁶. This may explain why our pigmented mice recovered quickly after noise exposure. These findings suggest that melanin exerts protective effects in the inner ear.

Brn4/Pou3f4 plays a crucial role in inner ear development. Mutant animals exhibit defects in auditory and vestibular functions^47,48. The EP in Pou3f4^y/- mice is significantly decreased, which is one of the reasons for deafness⁴⁹. In Pou3f4^y/- mice, the stria vascularis was thinner, accompanied by an abnormal distribution of melanin. We speculated that melanocytes migrated from the modiolus along Reissner’s membrane area to the stria vascularis, following a gradient from the base to apex, similar to the differentiation of the organ of Corti and the formation of the scala vestibuli and tympani^6,50,51. Shibata et al. described the integration of neural crest cells into the lateral wall of the stria vascularis as a three-step process⁵². However, the migration path of melanocytes from the otic vesicle to the stria vascularis has not yet been identified. However, our observations require confirmation through lineage tracing.

During evolution and development, the vestibular system emerges earlier than the cochlea^53,54. In Pou3f4^y/- mice, the development of vestibular melanocytes has been completed and their number has rapidly decreased, along with a significant increase in the number of “black ball-like” macrophages; this pattern was the same as that in normal mice. Although both cochlear and vestibular melanocytes originate from neural crest cells, there are some differences: most cochlear melanocytes originate from SCPs expressing Dhh, whereas vestibular melanocytes originate from the neural crest, migrating dorsolaterally (non-SCP-fated)⁵⁵. This finding explains the pigment differences observed in Pou3f4^y/- mice. We speculate that the enhanced directional recruitment of non-SCP-fated neural crest cells and obstruction of melanocyte migration (SCPs origin) from abnormal mesenchymal communication signals caused the two opposite pigment-deposition states of the vestibule and stria vascularis in Pou3f4^y/- mice (Supplementary Fig. 10), which needs to be further confirmed.

Transcriptome sequencing analysis revealed significant enrichment in regulating melanin production and release pathways, such as the PI3K/AKT and PPAR pathways^56–58. Adamts9 regulates the survival of melanoblasts in the epidermis and is essential for normal skin pigmentation⁵⁹. We observed the upregulation of this gene in Pou3f4^y/- mice, which may explain the vestibular melanin deposition. Adamts20 from the same family regulates the migration of neural crest-derived melanoblasts, and their mutations can lead to white patches⁶⁰. The downregulation of Adamts20 in Pou3f4^y/- mice appeared to affect melanocyte migration, although this requires further confirmation. Additionally, we observed increased expression of the efferocytosis-related molecule Itgb3, which aligns with the phenomenon of the increase of “black ball-like” cells in Pou3f4^y/- mice.

In conclusion, our findings enhance the understanding of the ultrastructural anatomy and immunology of the inner ear, laying a foundation for future research on its physiological functions.

Methods

Animals and Grouping

We studied three mouse strains with different coat colors (Supplementary Table 1): black (C57BL/6 J), albino (BALB/c), and black (Pou3f4^y/-). The inner ear of the Pou3f4^y/- mice exhibited an abnormal distribution of pigments. Wild-type (WT) C57BL/6 J (3 d/2 months/12 months) and BALB/c (2 months) mice were obtained from the Animal Center of Shandong University (Jinan, China). Pou3f4 knocked-out mice (Pou3f4^y/-, 2 months) were purchased from Cyagen Biosciences. Mice were housed in temperature–controlled (20 °C–22 °C) rooms with humidity maintained at 40–70%. They were subjected to a 12/12-h light/dark cycle and had ad libitum access to food and water. All research protocols were approved by the Animal Care Committee of Shandong University (No. ECAESDUSM 20,123,011) and conformed with the Guideline for the Care and Use of Laboratory Animal of the National institutes of Health. We have complied with all relevant ethical regulations for animal use.

Cisplatin Administration

We evaluated the changes in melanocytes under damage caused by the ototoxic drug cisplatin. Forty 2-month-old C57BL/6 J mice matched by age and sex were selected for the experiment and randomly assigned to either the Cis-treated after 0 months group (Cis (0 M) group) or the Cis-treated after 10 months group (Cis (10 M) group), with 20 mice in each group. The mice were administered cisplatin (Jiangsu Hansoh Pharmaceutical Group Co., Ltd., Jiangsu, China) at a dose of 3 mg/kg (in 0.9% normal saline solution) once daily for five consecutive days. On the sixth day, the mice were dissected, and samples were collected to form the Cis (0 M) group, which was defined as the cisplatin group. After continuous feeding for 10 months, the mice were dissected, and samples were collected to form the Cis (10 M) group, defined as the cisplatin old-age group.

For C57BL/6 J mice, postnatal day 3 (P3), 2 months (2 M), and 12 months (12 M) were used to define the juvenile, young, and old age groups, respectively, which also served as controls for the cisplatin old-age group. BALB/c mice were used as albino mice.

Sample Collection

The mice were anesthetized using isoflurane and decapitated. The temporal bones were removed, inner-ear tissues were collected, and part of the inner ear was prepared for electron microscopy. Another part of the inner ear was fixed overnight at 4 °C with 4% paraformaldehyde (PFA) (Sigma-Aldrich, St. Louis, MO, USA) and rinsed with phosphate-buffered saline (PBS; pH 7.4) to remove any residual PFA. The samples were prepared for frozen sectioning or subjected to microdissection. The lateral walls and vestibular tissues (utricle, ampulla, and common crus of the semicircular canals) were adhered to small glass slides precoated with CellTack for IF staining.

Frozen Sections

The inner-ear tissues were fixed overnight at 4 °C with 4% PFA, decalcified in PBS containing 10% ethylenediaminetetraacetic acid (Solarbio) at 4 °C for 24 h, dehydrated by sequential incubation in PBS containing 10, 20, and 30% sucrose and embedded in an optimal cutting temperature compound (Tissue-Tek, Sakura Fine-Tek, USA). A cryostat (Leica CM1850; Leica, Nussloch, Germany) was used to prepare 7 μm sections.

IF Staining

Inner-ear tissues and frozen sections were permeabilized with 0.5% Triton X-100 (Sigma-Aldrich) for 0.5 h and immunologically blocked with 1% fish gelatin solution (G7765; Sigma-Aldrich) for 1 h. The specimens were selected with corresponding primary antibodies: CD68 (1:300; ab53444, Abcam or 1:500; 97778, CST), F4/80 (1:500; Sc-52664, Santa Cruz), IBA1 (1:500; Ab5076, Abcam), Melan-A (1:200; nbp-30151, Novus), Met (1:50; Ab216330, Abcam), Nrp2 (1:50; 3366, Cell Signaling), and Kir4.1 (1:100; 312844, Abcam). The antibodies were diluted and incubated overnight at 4 °C. After washing with PBS, cells were incubated with FITC- or TRITC-conjugated donkey anti-rabbit, anti-rat, or anti-goat secondary antibodies (Life Technologies) and DAPI (D9542, Sigma-Aldrich, USA) diluted 1:1000 in 1% BSA-PBS for 1 h at room temperature in the dark. Capillaries were labeled with lectin Griffonia simplicifolia IB4 (GS-IB4) conjugated with Alexa Fluor 647 (I32450; Life Technologies). Finally, after washing with PBS, the coverslips were fixed, and images were obtained using a laser scanning confocal microscope (Leica DMi8; Leica).

Transmission Electron Microscopy

The inner-ear tissues was collected; the stria vascularis and vestibular tissues were removed, rinsed with PBS, and immediately placed in a fixing solution composed of 2% PFA and 2.5% glutaraldehyde (Sigma-Aldrich, G5882, pH 7.4). The sample blocks were prepared according to the conventional sample preparation method for TEM. The samples were fixed, dehydrated and embedded in Epon 812 (45345; Sigma-Aldrich). Semi-thin sections were stained with toluidine blue, and ultrathin sections were stained with lead citrate and uranyl acetate. Finally, the samples were observed using a TEM (JEOL-1200EX, JEOL-1010) at Shandong Weiya Laboratory, and images were captured and recorded using a Morada-G2 CCD.

Emission Scanning Electron Microscopy

The inner ear tissue samples were obtained and immersed in a fixing solution composed of 2% PFA and 2.5% glutaraldehyde at 4 °C for 6 h. After rinsing with PBS, the tissues were prepared, coated with 10 nm platinum and observed and recorded using a field emission scanning electron microscope at 5 kV (Sigma-300, ZEISS, Germany).

Immunoelectron Microscopy

Tissues were immunolabeled with nanogold particles for each protein using a pre-embedding protocol. The stria vascularis and vestibular tissues were fixed and incubated with primary antibodies against CD68 (1:200), Melan-A (1:100), Met (1:50), and Kir4.1 (1:500) overnight at 4 °C. After rinsing, goat anti-rabbit secondary antibodies (7204; Nanoprobes, 1:200) were added followed by fixation. After rinsing, the samples were incubated with a gold enhancement kit (2114, Nanoprobes, NY, US) for 1 min and rinsed with 1% sodium thiosulfate and water. After fixation with osmium tetroxide, dehydration, embedding, and sectioning, the samples were observed using a transmission electron microscope (JEOL-1200EX, JEOL-1010), and images were captured and recorded using a Morada-G2 CCD.

Fontana–Masson (F–M) Staining

The frozen inner-ear sections were washed with distilled water. The sections were stained with the F–M Stain Kit (G2032, Solarbio) to detect melanin at 55 °C in the dark for 1 h, washed with distilled water, and counterstained with eosin. After dehydration with 95% ethanol and absolute ethanol, clearing, and mounting, the presence of pigments was observed under a microscope.

RNA Extraction and qPCR

Total RNA was extracted and purified using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). cDNA was synthesized by reverse transcription using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. qPCR was performed using SYBR Premix Ex Taq (TaKaRa Bio). Gapdh was used as the reference gene. All data were analyzed using Eppendorf Realplex 2. The PCR primers used to amplify these genes are listed in Supplementary Table 2.

Noise Exposure

The mice were exposed to a 100 dB (8-16 kHz) SPL octave-band noise stimulus for 2 h using a sound generator (Crown CDi1000, AVIS). Noise calibration was performed before each exposure session.

Auditory Function Tests

Auditory function in mice was evaluated using ABR. Mice were anesthetized with xylazine (6 mg kg⁻¹) and ketamine (50 mg kg⁻¹); this anesthesia provided stable ABR thresholds. Needle electrodes were inserted into the subcutaneous tissue at the vertex (recording electrode), the ipsilateral ear (reference electrode), and the back (ground electrode). The TDT System III (Tucker-Davis Technologies, USA) was used for stimulus production and response recording. ABR was elicited using tone bursts at frequencies of 4, 8, 12, 16, 24, and 32 kHz, with 1024 stimulus repetitions and a scanning time of 10 ms. The initial stimulation intensity was set at 90 dB SPL and decreased by 5 dB steps until the response was no longer detectable. The lowest intensity eliciting a response was recorded as the hearing threshold.

RNA-seq Analysis

Total RNA from the inner ear of three 10-day-old mice (control or Pou3f4^y/-) was extracted using TRIzol reagent (Thermo Fisher Scientific), according to the manufacturer’s instructions. The quantity and purity of the total RNA were analyzed using a Bioanalyzer 2100 and RNA 6000 Nano LabChip Kit (Agilent, CA, USA). High-quality RNA samples (RIN > 7.0) were used to construct sequencing libraries. The transcriptomes of the Pou3f4^y/- and control mice were analyzed using the Illumina NovaSeq 6000 platform (Shandong Xiuyue Biotechnology Co., Ltd.); three mice from each group were sequenced. Differential gene expression analysis between the two groups was performed using DESeq2 software. Genes with a false discovery rate (FDR) parameter below 0.05 and an absolute fold change ≥ 1.5 were considered differentially expressed genes. Enrichment analyses of GO functions and KEGG pathways were performed for differentially expressed genes.

Smart-Seq Analysis

Single-cell amplification was performed using a SMART-Seq® HT Kit (Takara, USA) according to the manufacturer’s instructions. A total of 18 “black ball-like” cells from 9 C57BL/6 J mice (2-month-old, 3 mice/group) were lysed in a buffer containing free dNTP and oligo(dT) oligonucleotides. RNA fragments were copied into first-strand cDNA using reverse transcriptase and several template-free C nucleotides. The oligonucleotide primers were hybridized with poly(C) to form second-strand cDNA. The cDNA was purified and enriched using PCR. cDNA libraries were constructed using the KAPA Hyper Prep Kit (KAPA, USA) according to the manufacturer’s instructions. Purified libraries were quantified and validated using a Qubit® 3.0 Fluorometer and an Agilent 2100 bioanalyzer to confirm the insert size and calculate the mole concentration. The cluster was generated using cBot after the library was diluted to 10 pM and sequenced on the Illumina NovaSeq 6000 platform (Illumina, USA). Library construction and sequencing were performed by the Sinotech Genomics Co. Ltd. (Shanghai, China).

scRNA-seq combined analysis

We collected scRNA-seq data of the vestibular dataset from GSE168901¹⁶, processed it using the Seurat⁶¹ R package to remove low-quality cells, and performed dimensionality reduction and cell clustering. Next, we identified the corresponding macrophages and melanocytes using their canonical markers (“Cd68”, “C1qa”, and “Adgre1” for macrophages; “Dct”, “Mlana”, and “Mitf” for melanocytes). The GSE168901 data were integrated with the in-house smart-seq data (our data) using the Harmony algorithm⁶², while simultaneously subtyping cells. Melanocytes and macrophages were annotated using the corresponding markers. Cells expressing markers of both cell types are annotated as “hybrid” cells.

Image Analysis

F–M stained images were captured using a digital camera (Olympus, Tokyo, Japan) to count the number of pigment particles with a diameter of > 1 μm in the stria vascularis or to measure the width of the stria vascularis at the midpoint of its entire length. Confocal images of the stria vascularis and vestibular tissues were obtained using a Leica SPE confocal microscope and analyzed using the ImageJ software to evaluate the number of cells and cell diameters, respectively.

Statistics and Reproducibility

Statistical analyses were performed using SPSS version 25.0. One-way analysis of variance (ANOVA) was employed to measure the significance of differences among multiple groups, and unpaired t-tests were utilized between two groups. The results are expressed as the mean ± SEM. Sample sizes and number of replicates are reported in the Figure Legends. Values with p < 0.05 were considered significant. Mice were randomly divided to different groups. No samples or animals were excluded from the analysis. No statistical method was used to predetermine the sample size. All samples were processed in blind fashion. For each experiment, at least three samples were used to obtain average values for each group.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

H.B.W. and B.Q.L. designed the project. J.C., L.X., and Y.D.S. performed the experiments. J.C. wrote the manuscript. L.X., Y.D.S., Z.L., A.Z.Z., L.G.K., and J.Z.L. performed most of the experiments and data analysis. S.Y.W., Y.J., Y.X., K.F.X., and D.G.Z. provided some technical methods and revised the manuscript.

Peer review

Peer review information

Communications Biology thanks Igor Adameyko, Erling Olaf Koppang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Simona Chera and Dario Ummarino. A peer review file is available.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request. Additionally, Smart-seq (single-cell RNA-seq) and RNA-seq data have been deposited to the NCBI GEO database under accession numbers GSE306457 (Smart-seq) and GSE307045 (RNA-seq). The source data behind the graphs in the main manuscript can be found in “Supplementary Data file”. All data generated or analyzed during this study are included in this article and its supplementary information files. The previously published dataset (GSE168901) was used.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-026-09616-9.