Journal of Molecular Histology (2018) 49:63–73

https://doi.org/10.1007/s10735-017-9747-y

ORIGINAL PAPER

Immunohistochemical characterization of brush cells in the rat larynx

Yoshio Yamamoto1 · Yuko Ozawa1 · Takuya Yokoyama2 · Nobuaki Nakamuta2

Received: 26 October 2017 / Accepted: 28 November 2017 / Published online: 1 December 2017
© Springer Science+Business Media B.V., part of Springer Nature 2017

Abstract
The immunohistochemical characteristics of brush cells in the laryngeal mucosa were examined using immunohistochemistry
for various immunohistochemical cell markers including villin at the light and electron microscopic levels. Cells that were
immunoreactive to villin were barrel-shaped with thick cytoplasmic processes extending toward the lumen of the laryngeal
cavity. Immunoelectron microscopic observations revealed thick and short microvilli with long rootlets of microfilaments.
Numerous small clear vesicles and small finger-like cytoplasmic processes were observed in the apical process and lateral
membrane, respectively. Double immunofluorescence showed villin-immunoreactive cells were not immunoreactive for the
markers of solitary chemosensory cells, GNAT3 and phospholipase C, β2-subunit (PLCβ2), or for that of neuroendocrine
cells, synaptosome-associated protein 25kD. Furthermore, immunoreactivities for cytokeratin 18 (CK18) and doublecortin
like-kinase 1 in the perinuclear cytoplasm of villin-immunoreactive cells. However, some CK18-immunoreactive cells were
immunoreactive to GNAT3 but not to villin. Regarding sensory innervation, only a few intraepithelial nerve endings with
P2X3, SP, or CGRP immunoreactivity attached to villin-immunoreactive cells. In the present study, brush cells in the rat
laryngeal mucosa were classified by immunoreactivity for villin, and were independent of other non-ciliated epithelial cells
such as solitary chemosensory cells and neuroendocrine cells.

Keywords  Brush cells · Villin · Immunohistochemistry · Electron microscopy · Larynx · Rat

Introduction
Brush cells are one of the non-ciliated epithelial cells
observed in the respiratory tract, gastrointestinal tract,
and urinary tract (Sbarbati and Osculati 2005a; Sato 2007;
Deckmann and Kummer 2016), and are characterized by
thick blunted microvilli with the elongated rootlets of
cytoplasmic filaments and vesiculotubular components in
the supranuclear cytoplasm (Reid et al. 2005; Sato 2007).
Although brush cells have been suggested to play roles in
mechanosensation, chemosensation, excretion, and absorp-
tion, their exact functions remain unknown (Sbarbati and
Osculati 2005a; Reid et al. 2005; Sato 2007). In the respira-
tory tract, electron microscopic observations identified brush
* Yoshio Yamamoto
yyoshio@iwate‑u.ac.jp
1
Laboratory of Veterinary Anatomy and Cell Biology, Faculty
of Agriculture, Iwate University, 3‑18‑8 Ueda, Morioka,
Iwate 020‑8550, Japan
2
Department of Anatomy (Cell Biology), Iwate Medical
University, 2‑1‑1 Nishitokuta, Yahaba, Japan
cells in the larynx, trachea, and lung (Luciano et al. 1968,
1969; Jeffery and Reid 1975; Taira and Shibasaki 1978;
Marin et al. 1979; Lewis and Prentice 1980; Souma 1987).
In addition to electron microscopy, immunohistochemistry
has been attempted in order to identify brush cells. Previous
studies demonstrated that brush cells are immunoreactive for
villin, fimbrin, cytokeratin 18 (CK18), and ankyrin (Höfer
and Drenckhahn 1992, 1996; Kasper et al. 1994). In the
gastrointestinal tract, other molecules such as doublecortin-
like kinase 1 (DCLK1) and prostaglandin-endoperoxide syn-
thase-1 were also used to identify brush cells (Gerbe et al.
2011; Bjerknes et al. 2012).
In the larynx, various morphological studies classified
respiratory epithelial cells into various cell types including
ciliated cells, serous cells, neuroendocrine cells (Kulchit-
sky cells), and brush cells based on electron microscopic
observations (Lewis and Prentice 1980). Although electron
microscopy and immunohistochemistry are used in the
classification of non-ciliated respiratory epithelial cells, as
described above, the characteristics of brush cells remain
unclear. Immunoelectron microscopy revealed that some
solitary chemosensory cells with GNAT3 immunoreactivity

13
Vol.:(0123456789)

64
showed the brush cell morphology (Sbarbati and Osculati
2005b). These findings appear to be consistent with some
CK18 and GNAT3 co-existing with gastrointestinal brush
cells (Höfer et al. 1996; Hass et al. 2007). Thus, several stud-
ies concluded that chemosensory cells are a subpopulation
of brush cells.
On the other hand, electron microscopy showed that nerve
endings attached to brush cells in the trachea and excretory
ducts of the submandibular gland (Luciano et al. 1968; Sato
and Miyoshi 1997). Since solitary chemosensory cells and
diffuse endocrine cells are densely innervated by P2X3-,
SP-, and CGRP-immunoreactive nerve endings (Takahashi
et al. 2016), the innervation pattern of brush cells needs to
be compared with those of other cell types.
In the present study, we examined brush cells in the rat
larynx using immunohistochemistry for villin at the light
and electron microscopic levels. Immunohistochemical and
morphological differences were then investigated in order to
clarify differences among brush cells, chemosensory cells,
and neuroendocrine cells in the laryngeal epithelium using
various immunohistochemical markers. Furthermore, the
innervation of villin-immunoreactive cells was examined
using immunohistochemistry for P2X3, SP, and CGRP. We
attempted to re-consider the morphological and immunohis-
tochemical characteristics of brush cells.
Materials and methods
Animal procedure
All procedures for animal handling were performed in
accordance with the guidelines of the local Animal Eth-
ics Committee of Iwate University (accession number:
A201610). Male Wistar rats (8 weeks old, totally n = 32)
were purchased from Japan SLC, Inc. (Slc: Wistar, Japan
SLC, Hamamatsu, Japan).
Immunohistochemistry
For cryostat sections, rats (n = 25) were anesthetized by an
intraperitoneal pentobarbital injection (15 mg/Kg), and tran-
scardially perfused with Ringer’s solution (300 ml) followed
by the 4% paraformaldehyde in 0.1M phosphate buffer (PB;
pH 7.4; 300 ml). The larynges were dissected out and further
fixed at 4 °C for 12–18 h. Tissues were soaked in phosphate-
buffered saline (PBS; pH 7.4) containing 30% sucrose and
frozen at -80 °C with compound medium (Tissue-Tek O.C.T.
compound, Sakura Finetech, Tokyo), then they were serially
sectioned in transversal planes from the cricoid cartilage to
the tip of epiglottis at 20 μm. These sections were mounted
on glass slides coated with chrome-alum gelatin. After
being incubated with normal donkey serum (1:50) at room
13
Journal of Molecular Histology (2018) 49:63–73
temperature for 30 min, sections were then incubated with
the primary antibody at 4 °C for 12 h. Sections were sub-
sequently incubated with the secondary antibodies at room
temperature for 90 min after rinsing with PBS. Preparations
were then incubated with DAPI solution (1 μg/ml, Dojindo,
Kumamoto, Japan) for nuclear staining. Sections were cov-
erslipped with aqueous mounting medium (Fluoromount,
Diagnostic Biosystems, Pleasanton, CA, USA). The details
of primary and secondary antibodies are shown in Tables 1
and 2, respectively. Combinations of antibody for multila-
beling staining were listed in Table 3.
Regarding whole-mount preparations, rats (n = 4) were
euthanized by exsanguination from the abdominal aorta
under deep anesthesia by an intraperitoneal pentobarbital
injection (15 mg/kg). Then, larynges were opened at dorsal
midline with fine scissors, pinned on the silicone board and
fixed with 4% paraformaldehyde in PB at 4 °C for 12–18 h.
After washing with PBS, laryngeal cartilages and muscles
were removed by use of fine tweezer with binocular dissect-
ing microscope, and obtained the preparation of laryngeal
mucosa. Then, the specimens were incubated for 60 min
with non-immune donkey serum (1:100), followed by an
incubation with mixture of antibodies for villin, GNAT3 and
SNAP25 at 4 °C for at least 48 h. Preparations were then
incubated with secondary antibodies at room temperature
for 3 h, followed by an incubation with DAPI solution. The
specimens were mounted on the glass slide with the mount-
ing medium.
Observations
Preparations were examined with a confocal scanning laser
microscope (C1, Nikon, Tokyo, Japan). Images of Alexa488,
Cy3, Alexa647, and DAPI were captured and colored using
computer software (NIS-Elements, Nikon). Some sections
were simultaneously captured using differential interference
contrast images. Projection images were made from z-stacks
of confocal images (10–30 series at 0.5–1-μm intervals)
using the same software. Other images were reconstructed
to a three-dimensional view from intact confocal images of
z-stack series or binary images converted from the original.
Some photographs were merged into one figure with com-
puter software (Photoshop CS5, Adobe systems, San José,
CA, USA).
Electron microscopy
For scanning electron microscopy, rats (n = 2) were eutha-
nized by exsanguination from the abdominal aorta under
deep anesthesia by an intraperitoneal pentobarbital injection
(15 mg/Kg). Then, larynges were dissected out and fixed
with 3% glutaraldehyde in PB for 24 h. The small pieces
were cut from the larynges and tissues were treated with 1%
Journal of Molecular Histology (2018) 49:63–73 65

Table 1  Primary antibodies used in the present study

No. Antibody against Function Immunohistochemical marker Catalog number Host Dilution Source
for

1 Villin Actin-binding protein Brush cells sc-58897 Mouse 1:1000 A

2 Villin Actin-binding protein Brush cells E16640 Rabbit 1:200 B
3 GNAT3 Taste signaling protein Chemosensory cells ab1133664 Goat 1:2000 C
4 Phospholipase C β2 (PLCβ2) Taste signaling protein Chemosensory cells sc-206 Rabbit 1:500 A
5 Synaptosomal-associated protein SNARE protein Neuroendocrine cells MCA1308 Mouse 1:2000 D
25 kD (SNAP25)
6 Cytokeratin 18 (CK18) Cytoskeletal protein Brush cells ab133263 Rabbit 1:1000 C
7 Doublecortin-like kinase 1 Cytoplasmic enzyme Brush cells AP7219b Rabbit 1:200 E
(DCLK1)
8 Espin Actin-binding protein Certain microvilli NBP1-90588 Rabbit 1:100 F
9 P2X3 ATP receptor (P2X3) Ionotropic ATP receptor Certain sensory nerves RA10109 Rabbit 1:4000 G
10 Substance P (SP) Neuropeptide C-fiber sensory nerve AB1556 Rabbit 1:4000 H
11 Calcitonin gene-related peptide Neuropeptide C-fiber sensory nerve 1720-9007 Goat 1:500 D
(CGRP)

The antibody numbers in this table were used in Table 3

A Santa Cruz Biotechnology, Dallas, TX, USA; B Spring Bioscience, Pleasanton, CA, USA; C Abcam, Cambridge, UK; D Bio-Rad, Hercules,
CA, USA; E Abgent, San Diego, CA, USA; F Novus Biologicals, Littleton, CO, USA; G Neuromics, Edina, MN, USA; H, Merck Millipore,
Billerica, MA, USA

Table 2  Secondary antibodies used in the present study

Catalog number Host Dilution for cry- Dilution for whole- Dilution for
ostat sections mount preparations immunoelectron
microscopy

a Cy3-labeled anti-mouse IgG 715-165-151 Donkey 1:50 1:100

b Cy3-labeled anti-rabbit IgG 711-165-152 Donkey 1:50
c Alexa Fluor 488-labeled anti-goat IgG 705-545-147 Donkey 1:100 1:200
d Alexa Fluor 488-labeled anti-rabbit IgG 711-545-152 Donkey 1:100
e Alexa Fluor 488-labeled anti-mouse IgG 715-545-150 Donkey 1:100
f Alexa Fluor 647-labeled anti-rabbit IgG 711-175-152 Donkey 1:500
g Alexa Fluor 647-labeled anti-goat IgG 705-175-147 Donkey 1:500
Biotinylated anti-mouse IgG 715-065-151 Donkey 1:500

The antibody symbol letters in this table were used in Table 3

Antibodies were supplied from Jackson ImmunoResearch, West Grove, PA, USA

tannic acid in PB for 1 h followed by exposure to 1% O­ sO4
in PB for 1 h. Specimens were dehydrated with a graded
series of ethanol, immersed in 2-methyl-2-propanol (t-butyl
alcohol), and freeze-dried with a dryer (JFC-1200, JEOL,
Tokyo). Specimens were mounted on a metal stab, coated
with gold by ion sputtering (JFC-1200, JEOL), and exam-
ined with a scanning electron microscope (JSM-5800LV,
JEOL) at an accelerating voltage of 10 kV.
Regarding immunoelectron microscopy, larynx was
obtained from a rat by same procedure as mentioned in the
section ‘Immunohistochemistry’. The tissues were trans-
versely cut at a thickness of 50 μm with a cryostat. Free-
floating sections were immunohistochemically stained for
villin using the avidin biotin peroxidase complex method.
Sections were incubated with a mouse monoclonal anti-
villin antibody for at least 36 h. Preparations were incu-
bated with a biotinylated donkey antibody against mouse
IgG (Jackson Laboratories, West Grove, PA, USA) at room
temperature for 60 min. They were then incubated with the
reagent from the avidin–biotin peroxidase complex kit (Elite
ABC kit, PK-6100, Vector Laboratories, Burlingame, CA,
USA). Sections were treated with 0.02% 3,3′-diaminobenzi-
dine tetrahydrochloride (DAB) and 0.006% H ­ 2O2 in 0.05 M
Tris–HCl buffer (pH 7.4) to visualize immunoreaction
sites. Tissues were post-fixed with 1% ­OsO4 in PB (pH 7.4)
at 4 °C for 1 h. They were then dehydrated with a graded

13

66 Journal of Molecular Histology (2018) 49:63–73

Table 3  Combinations of Combination Primary Secondary Primary Secondary Primary Secondary

antibodies for multilabeling antibody 1 antibody 1 antibody 2 antibody 2 antibody 3 antibody 3
immunofluorescence
Villin/GNAT3 1 a 3 c
Villin/PLCβ 1 a 4 d
Villin/SNAP25 2 b 5 e
Villin/GNAT3/SNAP25 2 b 3 e 5 f
Villin/CK18 1 a 6 d
Villin/DCLK1 1 a 7 d
Villin/CK18/GNAT3 1 a 6 d 3 g
Villin/DCLK1/GNAT3 1 a 7 d 3 g
Villin/Espin 1 a 8 d
Villin/P2X3 1 a 9 d
Villin/SP 1 a 10 d
Villin/CGRP 1 a 11 c

Numbers and letters are shown in Tables 1 and 2

Fig. 1  Villin-immunoreactive cells in the epithelium of the epiglot- Villin-immunoreactive cells were barrel-shaped with thick apical
tis (a), glottis (b), and subglottis (c). Villin-immunoreactive cells are cytoplasmic processes and thick microvilli (arrows in d). Nuclei were
distributed in most areas of the laryngeal mucosa regardless of the stained with DAPI (a–d). A fluorescent image was merged with a dif-
epithelial type. d higher magnification view of the glottic epithelium. ferential interference contrast image

series of ethanol, immersed in propylene oxide, and embed- Leica, Wetzler, Germany). Sections were examined under a
ded with Epon812 (TAAB, Aldermaston, U.K.). Ultrathin transmission electron microscope (H-7650, Hitachi, Tokyo,
sections were made with an ultramicrotome (EM-UC6, Japan) after staining with lead citrate.

13
Journal of Molecular Histology (2018) 49:63–73 67
Results
Distribution and morphology
of villin‑immunoreactive cells
Villin-immunoreactive epithelial cells were widely dis-
tributed in the rat laryngeal epithelium in sections stained
with the mouse antibody against villin (Fig. 1). Cells that
were immunoreactive to villin were barrel-shaped with
thick cytoplasmic processes extending toward the lumen
Fig. 2  a Scanning electron micrograph of the surface structure of the
mucosa overlying arytenoid cartilages. Arrows indicate the dome-
like accumulation of the microvilli of brush cells among non-cil-
iated cells. b Electron micrograph of the ciliated epithelium immu-
nostained for villin. Barrel-shaped villin-immunoreactive cells are
shown in the epithelial lining (arrowheads). CC Ciliated cell, SC
of the laryngeal cavity. A large number of blunt-tipped
thick microvilli with intense villin immunoreactivity were
densely arranged in the apical membrane (Fig. 1d). The
perinuclear cytoplasm was weakly immunoreactive. The
stainability of the mouse antibody for villin was equiva-
lent to that of the rabbit antibody.
In the epiglottis, cells were sparsely distributed in the
stratified cuboidal epithelium (Fig. 1a). In the mucosa over-
lying the arytenoid cartilage of the glottis, villin-immuno-
reactive cells were densely arranged in the pseudostratified
serous cell. c, d. Higher magnification view of villin-immunoreactive
cells in the epiglottic epithelium. Thick and short microvilli (Mv) and
cytoplasmic vesicles (V) are shown in apical processes (c). Bunches
of cytoplasmic filaments (arrows) and lateral processes (arrowheads)
are shown in the perinuclear region (d)
13

68
epithelium (Fig. 1b). In the subglottis, villin-immunoreac-
tive cells were also densely distributed in the pseudostrati-
fied epithelium as well as the glottis (Fig. 1c).
Electron microscopy
Brush cells were observed among ciliated or non-ciliated
cells by scanning electron microscopy (Fig. 2a). The micro-
villi of brush cells were densely arranged and gathered as a
dome-like accumulation of microvilli.
Immunoelectron microscopic observations revealed vil-
lin-immunoreactive cells within ciliated or other types of
non-ciliated cells, and they contained round nuclei (Fig. 2b).
In apical cytoplasmic processes, thick and short microvilli
with the long rootlets of microfilaments were detected in
villin-immunoreactive cells (Fig. 2c). Numerous small clear
vesicles were observed in the apical process. Bunches of
cytoplasmic filaments were also found in the cytoplasm. In
the lateral membrane, small finger-like cytoplasmic pro-
cesses projected toward neighboring cells (Fig. 2d).
Fig. 3  Double immunofluorescence for villin with various cell mark-
ers in the laryngeal epithelium. Villin-immunoreactive cells (a, d, g)
are not immunoreactive for GNAT3 (b), PLCβ2 (e), or SNAP25 (h).
c, f and i merged figures with DAPI staining. Arrows in a–c indicate
13
Journal of Molecular Histology (2018) 49:63–73
Double immunofluorescence
On the other hand, immunoreactivities for GNAT3 and
PLCβ2, immunohistochemical markers for solitary che-
mosensory cells, were not observed in villin-immuno-
reactive cells (Fig.  3a–f). GNAT3 and PLCβ2 immu-
noreactivities were observed in pleomorphic cells with
elongated basal cytoplasmic processes with one apical
process. In some cells with GNAT3 immunoreactivity,
the tip of the apical process was weakly immunoreactive
for villin (Fig. 3a–c). In double immunofluorescence for
villin and SNAP25, the marker for neuroendocrine cells,
SNAP25 immunoreactivity was not observed in villin-
immunoreactive cells (Fig. 3g–i). SNAP25 immunoreac-
tivity was detected in cells with a thin apical process and
some intraepithelial nerve fibers. Whole mount prepara-
tions confirmed the results of immunofluorescence for
cryostat sections (Fig. 4). Immunoreactivity for villin,
GNAT3, or SNAP25 was not colocalized, and labeled dif-
ferent morphological cell types. Villin-immunoreactive
weak villin-immunoreactive microvilli in GNAT3-immunoreactive
cells. Arrowheads in h and i indicate SNAP-immunoreactive nerve
fibers in the laryngeal epithelium
Journal of Molecular Histology (2018) 49:63–73 69
Fig. 4  Whole mount preparation stained with triple immunofluo-
rescence for villin (labeled with Alexa Fluor 647, cyan), GNAT3
(labeled with Alexa Fluor 488, green), and SNAP25 (labeled with
Cy3, red). a Maximum intensity projection view from the z-stack of
cells were barrel-like in shape with thick apical processes.
Most GNAT3-immunoreactive cells had an oval perinu-
clear region with elongated basal processes. SNAP25-
immunoreactive cells were flask-like with slender apical
processes.
To compare the immunoreactivity for villin with other
markers for brush cells, multilabeling immunofluores-
cence were performed. Double immunofluorescence
showed immunoreactivities for CK18 and DCLK1 in the
perinuclear cytoplasm of villin-immunoreactive cells
(Fig. 5). Some CK18-immunoreactivity was observed in
GNAT3-immunoreactive cells but not in the villin-immu-
noreactive cells (Fig. 5a–f). DCLK1-immunoreactivity
was not observed in the GNAT3-immunoreactive cells
(Fig. 5g–l).
Espin immunoreactivity in microvilli
Espin immunoreactivity was observed in thick microvilli
in the apical processes of villin-immunoreactive cells
(Fig. 6a–c). In GNAT3-immunoreactive cells, espin immu-
noreactivity was also observed in microvilli that had a thin
arrangement and were slender in shape (Fig. 6d–f). Immu-
noreactivity for espin was not detected in SNAP25-immu-
noreactive cells (Fig. 6g–i).
confocal images (36 images at 1-μm intervals), and b three-dimen-
sional reconstitution view of a. Immunoreactivities to villin, GNAT3,
and SNAP25 are not colocalized in the same epithelial cell. (Color
figure online)
Distribution of P2X3‑, SP‑,
and CGRP‑immunoreactive nerves
P2X3-immunoreactive intraepithelial nerve fibers were
relatively thick with flat axon terminals on epithelial cells
(Fig. 7a). P2X3-immunoreactive nerve fibers attached to
some villin-immunoreactive cells.
Varicose nerve fibers with SP and CGRP immunoreac-
tivities were observed in the interface between the epithelium
and lamina propria, and branched to intraepithelial endings
(Fig. 7b, c). Some intraepithelial nerve endings with SP or
CGRP immunoreactivity were accompanied by villin-immu-
noreactive cells.
Discussion
Brush cells have been categorized by morphological
characteristics in various organs including the trachea
and bronchus. These cells have also been characterized
in detail by electron microscopy in the respiratory tract,
larynx (Lewis and Prentice 1980), trachea (Luciano et al.
1968; Marin et al. 1979), and alveolar mucosa (Meyrick
and Reid 1968; Hijiya et  al. 1977; Lewis and Prentice
1980). Their electron microscopic characteristics are as
follows: thick microvilli with the long rootlets of micro-
filaments, the accumulation of cytoplasmic filaments, a
13

70
Fig. 5  Triple immunofluorescence for villin (labeled with Cy3, red)
and GNAT3 (labeled with Alexa Fluor 647, magenta) with CK18
(labeled with Alexa Fluor 488, green; a–f) and DCLK1 (labeled with
Alexa Fluor 488, green; g–l). Merged figures (f, l) are shown with
nuclear staining with DAPI. a–f, CK18 immunoreactive cells are also
tubulovesicular system, and lateral cytoplasmic branches
(Sato 2007). In the present study, villin immunoreactivity
was identified in specific cell types characterized by the
tuft-like accumulation of microvilli at the luminal surface,
numerous vesicles beneath microvilli, and lateral cytoplas-
mic processes, similar to brush cells reported in previous
studies. Thus, villin antibodies may be used for detection
of the brush cells in the rat laryngeal mucosa.
In addition to morphology, immunohistochemical mark-
ers have been used in the classification of microvillous epi-
thelial cells. CK18 is the most classical and frequently used
marker for brush cells in addition to villin (Kasper et al.
1994; Höfer and Drenckhahn 1996, 1998; Höfer et al. 1996;
Hass et al. 2007; Bjerknes et al. 2012). Although all villin-
immunoreactive cells were positive for CK18 in the present
study, some CK18-immunoreactive cells were not immu-
noreactive for villin but for GNAT3. It has been reported
that GNAT3 was also expressed in CK18-immunoreactive
cells in the stomach, duodenum, and pancreatic duct (Höfer
et al. 1996; Höfer and Drenckhahn 1998; Hass et al. 2007).
13
Journal of Molecular Histology (2018) 49:63–73
immunoreactive to either villin (arrowheads) or GNAT3 (arrows). g–l
DCLK1-immunoreactive cells are corresponded with villin-immu-
noreactive cells (arrowheads), but not with GNAT3-immunoreactive
cells (arrows)
Because the GNAT3-immunoreactive cells in the laryngeal
mucosa have different morphological characteristics from
villin-immunoreactive cells, CK18 may not be suitable as an
immunohistochemical marker for brush cells in the epithe-
lium of the respiratory tract. On the other hand, DCLK1 has
been used as a specific marker for gastrointestinal brush cells
(Gerbe et al. 2009, 2011; Saqui-Salces et al. 2011). Since
DCLK1-immunoreactive cells corresponded with villin-
immunoreactive cells, immunohistochemistry for DCLK1
may also be used for the visualization of brush cells in the
laryngeal mucosa, similar to the small intestine.
Solitary chemosensory cells are the same cell popula-
tion as brush cells in the respiratory epithelium (Kaske et al.
2007; Krasteva et al. 2011; Saunders et al. 2013). In the
present study, immunoreactivities for GNAT3 and PLCβ2,
markers of solitary chemosensory cells, were not observed in
most villin-immunoreactive cells. Although some GNAT3-
immunoreactive solitary chemosensory cells had microvilli
that were weakly immunoreactive for villin, GNAT3-immu-
noreactive cells differed from villin-immunoreactive brush
Journal of Molecular Histology (2018) 49:63–73 71
Fig. 6  Double immunofluorescence for Espin with villin (a–c),
GNAT3 (d–f), and SNAP25 (g–i). The microvilli of epithelial cells
immunoreactive to villin (a–c) and GNAT3 (d–f) are immunoreactive
to espin. Arrowheads in a–c indicate espin-immunoreactive microvilli
cells in their cell shape and the arrangement of microvilli.
Therefore, GNAT3- or PLCβ2-immunoreactive solitary
chemosensory cells may be a different cell population from
brush cells, at least in the laryngeal epithelium. In the duo-
denum, it has been reported advillin is specifically expressed
in TRPM5 expressing cells, but villin is expressed in both
TRPM expressing and non-expressing cells by quantitative
RT-PCR technique (Bezençon et al. 2008). Further study is
needed to elucidate the expression of villin/gelsorin family
proteins to characterize the specific epithelial cell popula-
tions including brush cells.
Another non-ciliated cell, neuroendocrine cells were
reported in the laryngeal epithelium of the hamster (McDow-
ell et al. 1994), cat (Yu et al. 1996), and rat (Yamamoto
et al. 2000). This cell type was described as a Kulchitsky
cell in previous electron microscopic studies (Jeffery and
Reid 1975; Lewis and Prentice 1980), and is immunohis-
tochemically detectable using an antibody for SNAP25,
as well as serotonin and CGRP (McDowell et al. 1994; Yu
et al. 1996; Takahashi et al. 2016). In the present study,
in villin-negative cells. No espin-immunoreactive microvilli are pre-
sent in SNAP25-immunoreactive cells (g–i). Arrowheads in g–h indi-
cate espin-immunoreactive microvilli in SNAP25-negative cells
SNAP25-immunoreactive cells were not immunoreactive
for villin, and differed from villin-immunoreactive brush
cells and GNAT3-immunoreactive solitary chemosensory
cells in morphology. These results suggest that laryngeal
neuroendocrine cells are an independent cell population.
Triple immunofluorescence of a whole-mount preparation
of the laryngeal mucosa confirmed immunoreactivity for
villin, GNAT3, and SNAP25 in three different cell types:
brush cells, solitary chemosensory cells, and neuroendocrine
cells, respectively.
Espin immunoreactivity may also be useful for the clas-
sification of laryngeal non-ciliated epithelial cells. Espin
is an actin-binding molecule that is localized in the micro-
villi of certain cell types, e.g., cochlear hair cells, taste
receptor cells, and Merkel cells (Sekerková et al. 2006).
The arrangement of espin-immunoreactive microvilli may
also be characterized as three cell populations; brush cells,
solitary chemosensory cells, and neuroendocrine cells.
Although espin immunoreactivity was found in the micro-
villi of villin-immunoreactive brush cells and GNAT3- or
13

72
Fig. 7  Maximum intensity view of double immunofluorescence for
villin with P2X3 (a), SP (b), and CGRP (c). Insets indicate the lateral
projection view from arrows at the dotted line. A few nerve endings
with P2X3, SP and CGRP immunoreactivities are shown on the sur-
face of villin-immunoreactive cells
PLCβ2-immunoreactive solitary chemosensory cells, immu-
noreactivity for espin clearly revealed a different arrange-
ment of microvilli in the two cell types. Furthermore, the
microvilli of SNAP25-immunoreactive neuroendocrine cells
were not immunoreactive for espin. Therefore, the molecular
characteristics and arrangement of microvilli may also be
classified into the three types of laryngeal epithelial cells.
Regarding innervation, a few nerve endings immuno-
reactive to P2X3, SP, and CGRP appeared to attach to
villin-immunoreactive cells. Previous studies reported that
these nerve endings are densely distributed around soli-
tary chemosensory cells and neuroendocrine cells in the
laryngeal mucosa (Yamamoto et al. 2000; Finger 2005;
Tizzano et al. 2011; Takahashi et al. 2016). The results
of the present study suggested that villin-immunoreactive
cells were scarcely innervated by sensory nerves. Villin-
immunoreactive cells appear to differ from chemosensory
13
Journal of Molecular Histology (2018) 49:63–73
and neuroendocrine cells from the viewpoint of sensory
innervation.
In conclusion, the present results indicate that brush
cells are independent of other non-ciliated epithelial cells
in the rat laryngeal mucosa, such as solitary chemosensory
cells and neuroendocrine cells. Although brush cells are
stated as the same cell type to the chemosensory cells in
other organs, e.g., the gastrointestinal tract and trachea,
their characteristics need to be re-considered for the clas-
sification of the epithelial cell lineage. New nanostruc-
tures of brush cells were recently reported by serial block
face electron microscopy and a tape-collecting ultrami-
crotome with scanning electron microscopy techniques:
lateral ‘cytospinules’ projecting to the nuclei of neighbor-
ing epithelial cells and the interconnected tubular network
in the supranuclear cytoplasm (Hoover et al. 2017). The
exact functions of brush cells in the respiratory tract need
to be elucidated in further studies.
Acknowledgements  The authors are grateful to Ms. Eri Ishiyama,
Technical Support Center for Life Science Research, Iwate Medical
University for her technical support with the EM study. This work was
supported by the JSPS KAKENHI: Grant Number 15K07759 (YY)
from the Japan Society for the Promotion of Science (JSPS), Japan.
Compliance with ethical standards 
Conflict of interest  The authors declare that they have no conflict of
interest with regard to the content of this article.
References
Bezençon C, Fürholz A, Raymond F et al (2008) Murine intestinal cells
expressing Trpm5 are mostly brush cells and express markers of
neuronal and inflammatory cells. J Comp Neurol. 509:514–525
Bjerknes M, Khandanpour C, Möröy T et al (2012) Origin of the brush
cell lineage in the mouse intestinal epithelium. Dev Biol 362:194–
218. https://doi.org/10.1016/j.ydbio.2011.12.009
Deckmann K, Kummer W (2016) Chemosensory epithelial cells in
the urethra: sentinels of the urinary tract. Histochem Cell Biol
146:673–683. https://doi.org/10.1007/s00418-016-1504-x
Finger TE (2005) ATP Signaling is crucial for communication from
taste buds to gustatory nerves. Science 310:1495–1499. https://
doi.org/10.1126/science.1118435
Gerbe F, Brulin B, Makrini L et al (2009) DCAMKL-1 expression
identifies Tuft cells rather than stem cells in the adult mouse
intestinal epithelium. Gastroenterology 137:2179–2180. https://
doi.org/10.1053/j.gastro.2009.06.072
Gerbe F, van Es JH, Makrini L et al (2011) Distinct ATOH1 and Neu-
rog3 requirements define tuft cells as a new secretory cell type
in the intestinal epithelium. J Cell Biol 192:767–780. https://doi.
org/10.1083/jcb.201010127
Hass N, Schwarzenbacher K, Breer H (2007) A cluster of gustducin-
expressing cells in the mouse stomach associated with two dis-
tinct populations of enteroendocrine cells. Histochem Cell Biol
128:457–471. https://doi.org/10.1007/s00418-007-0325-3
Journal of Molecular Histology (2018) 49:63–73 73
Hijiya K, Okada Y, Tankawa H (1977) Ultrastructural study of the
alveolar brush cell. J Electron Microsc 26:321–329. https://doi.
org/10.1093/oxfordjournals.jmicro.a050078
Höfer D, Drenckhahn D (1992) Identification of brush cells in the ali-
mentary and respiratory system by antibodies to villin and fim-
brin. Histochemistry 98:237–242
Höfer D, Drenckhahn D (1996) Cytoskeletal markers allowing dis-
crimination between brush cells and other epithelial cells of
the gut including enteroendocrine cells. Histochem Cell Biol
105:405–412
Höfer D, Drenckhahn D (1998) Identification of the taste cell G-pro-
tein, α-gustducin, in brush cells of the rat pancreatic duct system.
Histochem Cell Biol 110:303–309
Höfer D, Püschel B, Drenckhahn D (1996) Taste receptor-like cells in
the rat gut identified by expression of α-gustducin. Proc Natl Acad
Sci USA 93:6631–6634
Hoover B, Baena V, Kaelberer MM et al (2017) The intestinal tuft
cell nanostructure in 3D. Sci Rep 7:1652. https://doi.org/10.1038/
s41598-017-01520-x
Jeffery PK, Reid L (1975) New observations of rat airway epithelium:
a quantitative and electron microscopic study. J Anat 120:295–320
Kaske S, Krasteva G, König P et al (2007) TRPM5, a taste-signaling
transient receptor potential ion-channel, is a ubiquitous signaling
component in chemosensory cells. BMC Neurosci 8:49. https://
doi.org/10.1186/1471-2202-8-49
Kasper M, Höfer D, Woodcock-Mitchell J et al (1994) Colocalization
of cytokeratin 18 and villin in type III alveolar cells (brush cells)
of the rat lung. Histochemistry 101:57–62
Krasteva G, Canning BJ, Hartmann P et  al (2011) Cholinergic
chemosensory cells in the trachea regulate breathing. Proc
Natl Acad Sci USA 108:9478–9483. https://doi.org/10.1073/
pnas.1019418108
Lewis DJ, Prentice DE (1980) The ultrastructure of rat laryngeal epi-
thelia. J Anat 130:617–632
Luciano L, Reale E, Ruska H (1968) Über eine “chemorezeptive”
Sinneszelle in der Trachea der Ratte. Z Zellforsch Mikrosk Anat
85:350–375. https://doi.org/10.1007/BF00328847
Luciano L, Reale E, Ruska H (1969) Bürstenzellen im Alveolarepithel
der Rattenlunge. Z Zellforsch Mikrosk Anat 95:198–201. https://
doi.org/10.1007/BF00968452
Marin ML, Lane BP, Gordon RE, Drummond E (1979) Ultrastructure
of rat tracheal epithelium. Lung 156:223–236
McDowell EM, Sorokin SP, Hoyt RF (1994) Ontogeny of endocrine
cells in the respiratory system of Syrian golden hamsters. I. Lar-
ynx and trachea. Cell Tissue Res 275:143–156
Meyrick B, Reid L (1968) The alveolar brush cell in rat lung - a third
pneumonocyte. J Ultrastruct Res 23:71–80
Reid L, Meyrick B, Antony VB et al (2005) The mysterious pulmonary
brush cell: a cell in search of a function. Am J Respir Crit Care
Med 172:136–139
Saqui-Salces M, Keeley TM, Grosse AS et al (2011) Gastric tuft cells
express DCLK1 and are expanded in hyperplasia. Histochem Cell
Biol 136:191–204. https://doi.org/10.1007/s00418-011-0831-1
Sato A (2007) Tuft cells. Anat Sci Int 82:187–199. https://doi.
org/10.1111/j.1447-073X.2007.00188.x
Sato A, Miyoshi S (1997) Fine structure of tuft cells of the main excre-
tory duct epithelium in the rat submandibular gland. Anat Rec
248:325–331
Saunders CJ, Reynolds SD, Finger TE (2013) Chemosensory brush
cells of the trachea. A stable population in a dynamic epithelium.
Am J Respir Cell Mol Biol 49:190–196. https://doi.org/10.1165/
rcmb.2012-0485OC
Sbarbati A, Osculati F (2005a) A new fate for old cells: brush
cells and related elements. J Anat 206:349–358. https://doi.
org/10.1111/j.1469-7580.2005.00403.x
Sbarbati A, Osculati F (2005b) The taste cell-related diffuse che-
mosensory system. Prog Neurobiol 75:295–307. https://doi.
org/10.1016/j.pneurobio.2005.03.001
Sekerková G, Zheng L, Loomis PA et al (2006) Espins and the actin
cytoskeleton of hair cell stereocilia and sensory cell micro-
villi. Cell Mol Life Sci 63:2329–2341. https://doi.org/10.1007/
s00018-006-6148-x
Souma T (1987) The distribution and surface ultrastructure of airway
epithelial cells in the rat lung: a scanning electron microscopic
study. Arch Histol Jpn 50:419–436
Taira K, Shibasaki S (1978) A fine structure study of the non-ciliated
cells in the mouse tracheal epithelium with special reference to the
relation of “brush cells” and “endocrine cells”. Arch Histol Jpn
41:351–366. https://doi.org/10.1679/aohc1950.41.351
Takahashi N, Nakamuta N, Yamamoto Y (2016) Morphology of P2
× 3-immunoreactive nerve endings in the rat laryngeal mucosa.
Histochem Cell Biol 145:131–146. https://doi.org/10.1007/
s00418-015-1371-x
Tizzano M, Cristofoletti M, Sbarbati A, Finger TE (2011) Expression
of taste receptors in solitary chemosensory cells of rodent airways.
BMC Pulm Med 11:3. https://doi.org/10.1186/1471-2466-11-3
Yamamoto Y, Kusakabe T, Hayashida Y et al (2000) Laryngeal endo-
crine cells: topographic distribution and adaptation to chronic
hypercapnic hypoxia. Histochem Cell Biol 114:277–282
Yu YC, Miyazaki J, Shin T (1996) Neuroendocrine cells in the cat
laryngeal epithelium. Eur Arch Otorhinolaryngol 253:287–293
13