ORIGINAL STUDY
Clear Cells of Toker in the Developing Anogenital Region
of Male and Female Fetuses
Sebastian C. J. van der Putte, MD
Abstract: The clear cells of Toker are a mysterious population of
intra-epidermal glandular cells. They were originally described in
nipples, but were recently observed in the vulva as well. It was
hypothesized that intra-epidermal embryonic remnants or underlying
glands were a potential source. The embryological aspects were in-
vestigated by studying specimens of the anogenital region of 18 male
and 15 female fetuses between 12 and 39 weeks gestation. The search
for Toker cells was enhanced by cytokeratin (CK) 7 immunohisto-
chemistry. The investigation showed that Toker cell elements are
a normal, though highly variable constituent of the developing
anogenital region. The study revealed the following: (1) single intra-
epidermal glandular vesicles near follicular anlages in interlabial
sulcuses of female fetuses of 15 and 16.5 weeks gestation; (2) CK7+
solitary cells, clusters, and vesicles which were related to developing
intra-epidermal follicular canal tracks and tended to disperse inside
the epidermis in fetuses of approximately 18 weeks gestation; (3)
dispersed CK7+ cells in fetuses of 19–23 weeks gestation; (4)
characteristic CK7+ Toker cell proliferations in fetuses more than 23
weeks gestation. These observations indicate that in the anogenital
region, primordial follicular cells programmed to participate in the
formation of apocrine and mammary-like glands, become displaced
into the epidermis where they disperse, and proliferate into Toker cell
populations. However, the proximity of Toker cells to CK7+ cells in
excretory ducts of late fetal apocrine and mammary-like glands
suggested a possible additional source. Consequences for Toker cells
of the breast and primary Paget disease are discussed.
Key Words: development, mammary-like gland, perineum, scrotum,
Toker cell, vulva
(Am J Dermatopathol 2011;33:811–818)
INTRODUCTION
In 1970, Toker1 reported a novel population of cells
inside the epidermis of normal nipples of male and female
breasts. The cells differed from epidermal keratinocytes by
their more abundant ‘‘clear’’ cytoplasm and round nuclei with
scant karyoplasm. They manifested themselves as single cells,
clusters, and small tubular formations in the lower half of the
From the Department of Pathology, University Medical Center Utrecht,
Utrecht, The Netherlands.
The authors declare no conflicts of interest.
Reprints: Sebastian C. J. van der Putte, MD, Department of Pathology, H04.
312, University Medical Center, PO Box 85550, 3508 GA, Utrecht, The
Netherlands (e-mail: s.c.j.vandeputte@umcutrecht.nl).
Copyright Ó 2011 by Lippincott Williams & Wilkins
Am J Dermatopathol  Volume 33, Number 8, December 2011
epidermis and were interpreted as non-neoplastic mammary
gland elements, which could, however, convert into the
neoplastic cell proliferation of primary intra-epidermal Paget
disease. After a long interval, other studies confirmed his
observations.2–10 Additional electron microscopy showed that
the clear cytoplasm was due to a paucity of organelles and
filaments.5 Immunohistochemistry supported the glandular
character by revealing the presence of low molecular weight
cytokeratins.2–5,7–10
The sensitivity of the ‘‘clear cells’’ or ‘‘Toker cells’’ for
CK7 in particular, greatly facilitated their detection and revealed
a preference for the epidermis surrounding the ostia of the
lacteriferous ducts in the nipples.3–5,7 Similar ‘‘clear cell’’ or
‘‘Toker cell’’ elements were reported in the areola,11 in accessory
nipples in females11–13 and males,11,14 and in the axilla.15
Recently such Toker-cell populations were also described in the
vulva,16 where they were implicated in extramammary Paget
disease.
Origin and nature of the cells have remained obscure,
notwithstanding the increased knowledge of their morphology.
Some authors suggested a primarily intra-epidermal derivation
from abortive mammary differentiation within the basal layer
of the epidermis during embryonic or postnatal life1 or from
embryonic nests of cells related to the embryonal stratum
germinativum which gives rise to the eccrine and apocrine
structures.2 Others hypothesized that the Toker cells, because
of their resemblance to mammary gland epithelium and the
spatial relationship to the ostia of the glands, migrated from
inside the mammary excretory ducts into the epidermis of the
nipple.3–5,8
My interest in this aspect of the Toker cell was raised by
the recognition of ‘‘clear cells’’ very similar to Toker cells in
the epidermis of the anogenital region of a female fetus during
research into the development of the human perineum.17 This
incidental observation and the availability of sufficient
material stimulated a systematic investigation which might
offer an opportunity to disclose the origin and earliest stages of
at least the clear cells of Toker in the anogenital region.
MATERIAL AND METHODS
The investigation was performed on specimens from
the anogenital region of 18 male fetuses and 15 female fetuses
that were part of a previous study on the development of the
human perineum.17 The gestational ages were correlated to
crown-rump lengths (CRL) measurements.18 These values
were for the male fetuses as follows: 12.5 weeks (80-mm CRL),
13 (90), 17 (140), 18 (155), 19 (160), 20 (170), 22 (190),
www.amjdermatopathology.com | 811
van der Putte Am J Dermatopathol  Volume 33, Number 8, December 2011

23 (200), 24 (210), 26 (230), 26.5 (240), 28 (260), 28 (265),
35 (320), 35 (320), 36 (330), and 38 (350), including 2 fetuses
of 26.5 and 28 weeks gestation with anorectal malformations.
For the female fetuses, the values were as follows: 12 weeks
(70-mm CRL), 13 (87), 14 (105), 15 (110), 16 (126),
16.5 (130), 18 (150), 20 (175), 21 (182), 22 (190), 24 (210),
26 (230), 28 (260), 35 (320), and 39 (360).
The specimens were fixed in 4% formalin and embedded
in paraplast. The specimens smaller than 150-mm CRL were
embedded in toto; the specimens larger than 150-mm CRL
were cut into transverse slices or cut sagitally into a right and
left half, and sectioned at 5–7 microns. The plane of sectioning
was transverse, frontal, or sagittal to the surface of the
perineum. Sections were mounted in complete series, series of
step sections (1:5 to 1:10) or as single sections from slices in
male fetuses more than 20 weeks gestation or as combinations.
Most series could be complemented by further sectioning of
the embedded rest tissue, which also allowed immunohisto-
chemical testing. The sections were stained with hematoxylin
and eosin.
An anti-CK7 antibody was selected for the detection of
Toker cells because of its high sensitivity and specificity in
relation to the epidermal keratinocytes. CK20 was tested to
discriminate Toker cells from CK7+, CK20+, Merkel cells if
present.3 The tests were performed on rehydrated 4-mm
paraffin-embedded sections using an automatic Bond-Max
immunohistochemistry system (Leica, Newcastle, United
Kingdom). Slides were heated for 10 minutes at 56°C. CK7
was brought in a pH 6.0 citrate buffer for 20 minutes at 99°C,
CK20 was similarly treated by ethylene diaminetetraacetic
acid pH 9. The anti-CK7 clone OVTL 12/30 (Biogenex, San
Ramon, CA) was used at a dilution of 1:800, and the anti-
CK20 clone Ks 20.8 (DAKO, Glostrup, Denmark) at a dilution
of 1:200. The primary antibody incubation was 15 minutes.
It was followed by postprimary incubation for 8 minutes,
polymer—peroxidase for 8 minutes, and diaminobenzidine for
10 minutes (Bond polymer refine detection system) at room
temperature. Sections were counterstained with hematoxylin.
Tissues that were previously shown to demonstrate reactivity
with the 2 antibodies were used as positive controls.
RESULTS
Female Anogenital Region
The earliest intra-epidermal glandular structures reminis-
cent of Toker cell elements were observed in fetuses of 15 and
16.5 weeks gestation at the bottom of the interlabial sulcuses of
the vulva where the first few follicular germs of the region had
formed after 12 weeks gestation (Fig. 1A). Some germs had
elongated into club-shaped primordial follicles with a cap of
stromal cells characteristic of hair follicle differentiation and
showed a small lateral bud in some indicating early gland
formation (Fig. 1B). Intra-epidermal extensions of inner
follicular cells formed tracks representing the anlages for the
future follicular canals (Fig. 1B).
The first intra-epidermal glandular structures occurred in
the vicinity of these primordial structures in the form of a single
vesicle in the basal layer at the bottom of a single sulcus in
fetuses of 15 and 16.5 weeks gestation (Fig. 1C and D). Their 2-
layered epithelium was highly suggestive of a glandular nature,

FIGURE 1. Follicular germ (A), primordial follicle (B), and vesicles (C and D) at the bottom of interlabial sulcuses in vulvas of fetuses
of 13,17.5, 15, and 16.5 weeks gestation, respectively. A, Follicular germ showing a pronounced basal layer and a small number of
inner germ cells (large arrow) (hematoxylin and eosin). B, Elongated primordial follicle consisting of a hair follicular part (large
arrow) covered by dense stroma of future hair papilla, a glandular primordium (small arrow) and an early track for the future
follicular canal (small arrowhead) inside the epidermis (hematoxylin and eosin). C, Single vesicle (large arrowhead) consisting of
a 2-layered glandular epithelium and occurring near early germinative epithelium (large arrow) at the bottom of an interlabial
sulcus. D, Similar vesicle (large arrowhead) situated near a follicular primordium (large arrow) (surrounded by a shrinkage artefact)
with a glandular protrusion (small arrow) (hematoxylin and eosin).

812 | www.amjdermatopathology.com q 2011 Lippincott Williams & Wilkins

Am J Dermatopathol  Volume 33, Number 8, December 2011 Toker Cells in the Fetal Anogenital Region

which was confirmed by CK7 reactivity in the 1 specimen
tested.
More marked CK7+ elements were encountered inside
the epidermis at the bottom of both sulcuses of a fetus of 18
weeks gestation. These elements consisted of isolated cells,
small clusters, and vesicular structures lined by a thin layer
of epithelial cells (Fig. 2). They were linked to the intra-
epidermal follicular canal tracks of small epithelial cells that
originated from follicles with large primordial glands,
resembling mushrooms by their prominent end-buds (Figs.
2A–D). Immunohistochemical tests demonstrated that most
cells of the tracks were weakly reactive for CK 7 similar to
the cells in the center of the superficial part of the follicle.
The vesicles and a few isolated cells at the periphery of the
tracks were, however, strongly positive. Some vesicles
occurred in the center of the tracks, others had a more
peripheral position and could extend into the surrounding
epidermis (Fig. 2D).

FIGURE 2. Toker cell elements and primordial mammary-like glands in female fetuses of 18 (A–E) and 20 (F and G) weeks gestation.
A, Primordial hair follicle (large arrow) and mammary-like gland with a prominent end-bud (small arrow) (connection is out of the
plane of sectioning) embedded in the dense connective tissue beneath the bottom of the interlabial sulcus. The follicular canal track
(small arrowhead) contained a small vesicle (large arrowhead) (inset: detail) (hematoxylin and eosin). B, Detail of follicular canal
track (small arrowhead) revealing a small CK7+ vesicle and cluster (large arrowheads). C, Follicular canal track (small arrowhead)
with large CK7+ vesicle (large arrowhead). Large arrow: primordial hair follicle (mammary-like gland out of plane of sectioning). D,
Two CK7+ vesicles (large arrowheads) extending into the surrounding epidermis. Small arrowhead: track of follicular canal. E,
follicular canal tracks (small arrowheads) in the perineum revealing a CK7+ vesicle and cluster. F, Small isolated cluster of CK7+ cells.
G, Almost completely regressed primordial follicle (large arrow) and a mammary-like gland (small arrow), which opens directly at
the surface of the bottom of the interlabial sulcus (hematoxylin and eosin).

q 2011 Lippincott Williams & Wilkins www.amjdermatopathology.com | 813

van der Putte
The association of strongly CK7+ cells and the follicular
canal tracks was not restricted to the deep part of the interlabial
sulcuses but to a lesser degree also occurred in the rest of the
anogenital region, that is, in the superficial part of the sulcuses,
in the short midperineal region (also called ‘‘female
perineum’’) (Fig. 2E), and around the anus, where primordial
follicles developed small glandular tubules without prominent
end-buds.
The tracks had differentiated into follicular canals in the
vulvas of the somewhat older fetuses of 20, 21, and 22 weeks
gestation. Intra-epidermal CK7+ cells were present but in
small numbers and scattered as individual cells and a few small
clusters (Fig. 2F). The concentration was highest at the bottom
of the sulcuses, but a few elements were also observed in the
rest of the anogenital region. The fetus of 20 weeks showed
also follicular structures in the final stage of regression and
large glands opening directly at the surface at the bottom of its
sulcuses (Fig. 2G).
In sharp contrast, the well-differentiated epidermis of the
vulvas in the fetuses between 24 and 39 weeks gestation
displayed marked proliferations of CK7+ cells in most cases.
The cells showed the characteristic morphology of Toker cells
revealing a moderate amount of lightly stained cytoplasm and
round nuclei with scant karyoplasm (Fig. 3). They were easily
identified in hematoxylin and eosin staining when forming
clusters or vesicular structures but were difficult to recognize
as individual cells in most cases. Immunohistochemistry
showed strong reactivity for CK7 and unreactivity for CK20,
which precluded confusion with Merkel cells.
The variation in number and distribution of the Toker cell
elements was great. Predominantly, the elements occurred
irregularly dispersed in the basal half of the epidermis, with
the highest concentration in the deep parts of the interlabial
sulcuses (Figs. 4A, B). Some fetuses revealed an additional
FIGURE 3. Characteristic Toker cell proliferation at the bottom of an interlabial sulcus in a female fetus of 35 weeks gestation.
A, Typical clear cells (small arrowheads) (hematoxylin and eosin). B, Vesicle (large arrowhead) (hematoxylin and eosin). C, CK7
reactivity revealing a much more extensive Toker cell population than was expected in the hematoxylin and eosin staining.
814 | www.amjdermatopathology.com
Am J Dermatopathol  Volume 33, Number 8, December 2011
conspicuous presence in the walls of the infundibulum of
folliculosebaceous units associated with a small or large type of
gland (Figs. 4C–E). These Toker cell elements were separated
from the CK7+ secretory epithelium by a long segment of
multilayered squamous epithelium of the secretory duct.
However, one large gland, in the oldest fetus studied, revealed
CK7+ clusters in the terminal part of the duct near a concentration
of Toker cells (Fig. 4F).
The small glands associated with the folliculosebaceous
units apparently represented still immature apocrine glands (Fig.
4E). They developed from buds on the primordial follicles, which
formed the anlages for tubular glands but mostly regressed almost
immediately. The remaining buds grew into straight tubules,
which formed secretory coils directly beneath the developing
sebaceous glands. The narrow excretory ducts opened always in
the infundibulum of folliculosebaceous units. These apocrine
glands developed predominantly in the superficial part of
the sulcus interlabialis, in the midperineal region, and around
the anus.
The large associated glands can be interpreted as
immature mammary-like glands. They developed from the
large mushroom-shape primordial glands with prominent end-
buds at the bottom of interlabial sulcuses of the female fetuses.
The wide excretory ducts might open in a reduced follicu-
losebaceous structure or directly at the surface without this
structure and often showed their secretory coils at a consider-
able distance from the orifice.
Number and size of the Toker cell elements were much
smaller in the rest of the anogenital region. CK7+ cells were
only sporadically observed in the surrounding skin, with
the exception of a narrow adjacent zone in the pubic region of
the oldest fetus.
No Toker cells were found in relation to eccrine glands
developing from the epidermis between folliculosebaceous units.
q 2011 Lippincott Williams & Wilkins
Am J Dermatopathol  Volume 33, Number 8, December 2011 Toker Cells in the Fetal Anogenital Region

FIGURE 4. Toker cell proliferations in interlabial sulcuses of female fetuses of 24 weeks (A) and 39 weeks (B–F) gestation. A, The
earliest observation of an extensive infiltration of CK7+ Toker cell elements showing some dispersed solitary cells and many large
vesicles in the deep part of a sulcus. B, Dispersed CK7+ Toker cell elements in the deep part of an interlabial sulcus. C, Concentration
of CK7+ Toker cell elements around the wide opening of a mammary-like gland (small arrow) into the infundibulum of
a folliculosebaceous unit (large arrow). D, Similar configuration of Toker cell elements (large arrowheads) in a hematoxylin and
eosin stained section. E, Toker cell elements (large arrowheads) around the orifice of an apocrine gland (small arrow) (hematoxylin
and eosin). F, mammary-like gland (small arrow) showing proximity of Toker cells (large arrowhead) and CK7+ glandular cells
(small arrowheads) within the stratified squamous epithelium of the wide excretory duct (inset: detail).

Male Anogenital Region
The development of Toker-cell populations in the
anogenital region of male fetuses was quite similar to the
histogenesis of these cells in female fetuses.
The first Toker-cell elements were revealed by CK7
reactivity in the perianal zone and scrotum of a fetus of 19
weeks gestation in the form of solitary cells and small clusters
and vesicles. They were often related to the developing intra-
epidermal follicular canals (Figs. 5A, B). The number and size
of such CK7+ elements was still very small in the fetuses of 20
and 22 weeks gestation. Appearances greatly changed at
approximately 23 weeks gestation when vesicles were easily
identified in hematoxylin and eosin sections and the popula-
tion including solitary cells and clusters became a conspicuous
feature in tests for CK7 (Fig. 5C).
The Toker-cell elements had developed throughout the
anogenital region, including the perianal zone, midperineal
region, scrotum, and penis, but not in the peripheral skin, with
the exception of an adjacent zone in the pubic area of the fetus
of 36 weeks gestation. Number and density greatly varied
within the same anogenital region and between the anogenital
regions of different fetuses. Most elements did not show

q 2011 Lippincott Williams & Wilkins www.amjdermatopathology.com | 815

van der Putte Am J Dermatopathol  Volume 33, Number 8, December 2011

FIGURE 5. CK7+ Toker cell proliferations in the perineum of male fetuses of 19 weeks gestation (A and B), and 36 weeks gestation
(C and D). A and B, Small CK7+Toker cell elements (large arrowhead) linked to early follicular canals (small arrowheads). C,
Concentration of dispersed CK7+ Toker cell vesicles lateral to the perineal raphe. D, CK7+ Toker cell elements (large arrowheads) at
a short distance from a cluster of CK7+ glandular cells (small arrowhead) inside the stratified squamous epithelium of the wide
excretory duct of a perianal mammary-like gland (large arrow), which was not associated with a folliculosebaceous structure. Small
arrow: orifice of mammary-like gland.

a special relationship to epidermal appendages in fetuses less
than 35 weeks gestation with only an occasional Toker-cell
element observed at the openings of folliculosebaceous units.
Such an association was regularly observed in older fetuses,
revealing concentrations of Toker cells around the openings of
apocrine glands into large sebaceous glands with often
underdeveloped hair follicles situated around the anus and
in the grooves bounding the perineal raphe. This feature was
most striking in the anogenital region of the fetus of 36 weeks
gestation (Figs. 5C, D) and compared well with the female
fetus of 39 weeks gestation (Figs. 4C–F). In this fetus, several
glands showed Toker cells very close to CK7+ cells inside the
stratified squamous epithelium of excretory ducts of mam-
mary-like glands (Fig. 5D) and of apocrine glands.
DISCUSSION
This investigation demonstrated in the first place that the
clear cells of Toker were a highly variable but normal
constituent of the developing anogenital region in male and
female fetuses. Toker-cell populations in fetuses more than 24
weeks gestation in particular were similar to those described in
normal nipples1–10 and areola11 of the breast, accessory
nipples,11–14 and vulvas.16 To the best of our knowledge, Toker
cells have not yet been described in the anogenital region of
the male. The strong reactivity with the anti-CK7 antibody in
an otherwise unreactive epidermis proved to be of great
practical value for detecting Toker cells, including solitary
cells and small clusters which were difficult to identify in
routine hematoxylin and eosin–stained sections of these
fetuses. This antibody demonstrated that the highest concen-
tration of Toker-cell elements occurred in the deep part of the
interlabial sulcuses of the developing vulva. This involvement
was in accordance with the observations of Willman et al16 in
the adult vulva where Toker cells were found to be associated
with the ostia of mammary-like glands. These ostia are
positioned in the borderland of labia minora and majora, which
corresponds to the deep part of the fetal interlabial sulcuses.17
The strong preference for this small and confined space in
combination with the strong CK7 reactivity was a great asset
for detecting the inconspicuous intra-epidermal glandular
elements in female fetuses less than 23 weeks gestation and to
trace them back to their origin.
The pattern of Toker-cell elements in the sulcuses of the
female fetus of 18 weeks gestation in particular highlighted
the major role of the follicle-born intra-epidermal tracks of the
future follicular canals in the histogenesis of Toker cells in
dissiminating the glandular elements inside the epidermis of
both sexes. Such tracks extended from an intrafollicular cell
mass, which provided at that stage not only the elements for
the future folliculosebaceous unit and intra-epidermal follic-
ular canal but also for a future gland. The observations suggest
that in a ‘‘mainstream’’ of primordial follicular cells
differentiating into the early follicular canal, cells programmed
to participate in gland formation became displaced into the
epidermis, dispersed, and differentiated into glandular ele-
ments. It seems plausible that the single vesicles at the bottom
of the interlabial sulcuses in the younger fetuses of 15 and 16.5
weeks gestation derived from the younger follicular germ
cells. The small number of widely dispersed mostly solitary
CK7+ cells observed during this investigation in the epidermis
of the older male and female fetuses of approximately 20–22
weeks gestation can be understood as representing a more
modest proliferation which may form Toker-cell clusters and

816 | www.amjdermatopathology.com q 2011 Lippincott Williams & Wilkins

Am J Dermatopathol  Volume 33, Number 8, December 2011
vesicles later or remain as single cell formations or may
represent abortive elements leading to anogenital regions
without an appreciable Toker-cell presence.
These observations strongly linked Toker cells to the
tubular glandular element of primordial follicles. The study
revealed that these elements differentiated into apocrine glands
and mammary-like glands, which differed in development,
size, relations, and topography. There was no indication that
Toker-cell elements were related to developing eccrine glands.
The histogenesis of mammary-like glands was not
reported before to the best of our knowledge. Of special
interest were the conspicuous end-buds, which were not
observed in early apocrine glands but did figure in a description
of developing accessory mammary glands in the areola of
the female breast.19 The convincing derivation of mammary-
like glands from follicular structures may seem to contradict
the close relationship between mammary-like glands and
eccrine glands in adults, which was suggested by the
occasional presence of mammary-like gland segments in
otherwise typical eccrine glands in the vicinity of characteristic
mammary-like glands20,21 and by some expression of receptors
for oestrogen, progesterone, and androgen. Such expression
was less than in the neighbouring mammary-like glands but
formed a contrast with the lack of reactivity in the usual type
of eccrine gland outside the zone with mammary-like glands
(20, S.C.J.P. unpublished observations). The incontestably
mixed character of those glands, however, can be considered as
mammary-like gland metaplasia in originally eccrine glands.
Such a phenomenon was presumed by Pfeifer et al22 to explain
the development of mammary-like gland–related lesions
outside the vulva and does reminds of the existence of
apocrine metaplasia in eccrine glands manifesting as ‘‘apoec-
crine glands’’.23 It could not be established that such eccrine
glands with mammary-like gland metaplasia transformed into
the characteristic mammary-like glands. It is, therefore, quite
well possible that characteristic mammary-like glands which
are positioned nearest to the the labia minora represent the
original fetal mammary-like glands.
The association of Toker cells with the ostia of mammary-
like glands in adult women16 prompted the question if this
configuration could be the result of migration of CK7+ cells
from deeper parts of the glands as was suggested by the relation
between Toker cells and mammary glands.3–5,7,9,10 It is as
a matter of fact still unknown if an original follicle-related fetal
Toker-cell population such as observed in the present study
persists throughout life or is supplemented and/or replaced by
CK7+ cells originating from more or less differentiated apocrine
glands and/or mammary-like glands. But the present investiga-
tion did reveal proximity of CK7+ Toker cells concentrating
around the ostia of apocrine glands and mammary-like glands
and solitary cells and clusters of CK7+ cells inside the
multilayered squamous epithelium of the excretory ducts of
some of these glands in the oldest male and female fetuses,
which could indicate a contribution of Toker cells by the glands.
Toker cell development in its different ways occurred
throughout the anogenital region of male and female fetuses
and not beyond as was also reported in adults by Willman
et al16 This suggests a special local stimulation and/or high
sensitivity of the epithelium of primordial or more mature
q 2011 Lippincott Williams & Wilkins
Toker Cells in the Fetal Anogenital Region
apocrine glands and mammary-like glands. It may relate to the
special developmental histological and physiological proper-
ties of the area. As a derivative of the cloaca and supportive
mesenchymal structures, the anogenital region demonstrates
from the beginning of its development special histological
features in the structure of the epidermis and its appendages
and in the presence of fascial tissue instead of the usual dermis
and subcutis.17 In addition, influence by sex hormones is
indicated by variable expression of receptors for estrogen,
progesterone, and androgen in the stromal cells, epidermis,
and mammary-like glands in the adult (S.C.J.P. unpublished
observations). Such expression was not found in the fetal
tissues during the present investigation with the exception of
weak reactivity for androgen receptors in sebaceous, apocrine,
and mammary-like glands. This lack of expression may have
been the result of the immature character of the tissues and/or
tissue preservation, which might have been insufficient for the
sensitive immunoreactivity.
No differences were noticed between Toker-cell pop-
ulations in the anogenital region and those in the nipples and
areola and accessory nipples in cytology, histology, and
distribution pattern.1,3,5,7,9,11,14 Immunocytochemical data
seem to be in line although comparison was restricted by
the still limited number of tests in anogenital Toker cells and
the variable results in Toker cells of the nipple areola.2–5,7,12–14
Various theories have been proposed with respect to the
source of these mamma-related Toker cells. These theories
name intra-epidermal abortive embryonic or postnatal elements
related to mammary glands,1 embryonic germinative cells
which give rise to apocrine and eccrine glands,2 and mammary
duct cells extending through the ostia into the epidermis.3–5,7,9,10
It is interesting to note that these separate theories came together
in the present study. That study shows that to find answers, it
may be worthwhile to perform an (immuno) histological study
of successive developmental stages of mammary and accessory
mammary glands complemented with a detailed investigation of
continuity, if any, between CK7+ intraglandular cells and Toker
cells in the interphase between the terminal lacteriferous ducts
and the epidermis in fetuses and adults.
Recently, Toker cells of the vulva were proposed as the
most likely precursor of primary anogenital Paget disease, that
is, unrelated to intra-epidermal extension of urogenital or
anorectal carcinomas because of the striking similarities in
topography, growth and distribution pattern, cytomorphology
and enzyme, and immunohistochemistry.6,16,24
Considering the presence of Toker-cell infiltrations in
male fetuses and the consequential likelyhood that Toker cells
will be found in adult males, that proposition may apply to the
disease in males as well.
A pathological role for Toker cells is precluded in the
large majority of the patients with Paget disease of the breast
because these cases represent secondary involvement of the
epidermis by underlying carcinomas of mammary glands.
Primary intra-epidermal Paget disease, which is by far the
most prevalent in the anogenital region was rarely reported.11
Histopathological data were scarce, but a striking presence of
Toker cells indistinguishable from anogenital Toker cells
which was noted in the nipple-areola area of the 1 case studied
in detail11 supported the idea of a relationship.
www.amjdermatopathology.com | 817
van der Putte
ACKNOWLEDGMENTS
We are grateful to J.A.S van Ginkel of the Laboratory of
Histpathology, and Mr K. van der Ven and Mr J. H. de Pinth,
of the Laboratory of Molecular and Immunopathology, and
their colleagues for their technical assistance.
REFERENCES
1. Toker C. Clear cells of the nipple epidermis. Cancer. 1970;25:601–610.
2. Nagle RB, Lucas DO, McDaniel KM, et al. Paget cells. New evidence
linking mammary and extramammary Paget cells to a common cell
phenotype. Am J Clin Pathol. 1985;83:431–438.
3 Lundquist K, Kohler S, Rouse RV. Intraepidermal cytokeratin expression is
not restricted to Paget cells but is also seen in Toker cells and Merkel cells.
Am J Surg Pathol. 1999;23:212–219.
4. Yao DX, Hoda SA, Chiu A, et al. Intraepidermal cytokeratin 7 immuno-
reactive cells in the non-neoplastic nipple may represent interepithelial
extension of lactiferous duct cells. Histopathol. 2002;40:230–236.
5. Marucci G, Betts CM, Golouh R, et al. Toker cells are probably precursors
of Paget cell carcinoma: a morphological and ultrastructural description.
Virchows Arch. 2002;441:117–123.
6. Fernandez-Flores A. Toker cell pathology as a unifying concept.
Histopathol. 2008;52:889–904.
7. Di Tommaso L, Franchi G, Destro A, et al. Toker cells of the breast.
Morphological and immunohistchemical characterization of 40 cases.
Human Pathol. 2008;39:1295–1300.
8. Garijo MF, Val D, Val-Bernal JF. An overview of the pale and clear cells of
the nipple epidermis. Histol Histopathol. 2009;24:367–376.
9. Park S, Suh Y-L. Useful immunohistochemical markers for distinquishing
Paget cells from Toker cells. Pathology. 2009;41:640–644.
10. Nofech-Mozes S, Hanna W. Toker cells revisited. Breast J. 2009;15:394–398.
11. Van der Putte SCJ, Toonstra J, Hennipman A. Mammary Paget’s disease
confined to the areola and associated with multifocal Toker cell
hyperplasia. Am J Dermatopathol. 1995;17:487–493.
818 | www.amjdermatopathology.com
Am J Dermatopathol  Volume 33, Number 8, December 2011
12. Martin VG, Pellettiere EV, Gress D, et al. Paget’s disease in an adolescent
arising in a supernumerary nipple. J Cut Pathol. 1994;21:283–286.
13. Decaussin M, Laville M, Mathevet P, et al. Paget’s disease versus Toker
cell hyperplasia in a supernumerary nipple. Virchows Arch. 1998;432:
289–291.
14. Willman JH, Golitz LE, Fitzpatrick JE. Clear cells of Toker in accessory
nipples. J Cut Pathol. 2003;30:256–260.
15. Makino T, Nakamura S, Nakayama H, et al. Genital Paget’s disease
with clear cells in the epidermis of the axilla. J Cut Pathol. 1998;25:
568–571.
16. Willman JH, Golitz LE, Fitzpatrick JE. Vulvar clear cells of Toker.
Precursors of extramammary Paget’s disease. Am J Dermatopathol. 2005;
27:185–188.
17. Van der Putte SCJ. The development of the perineum in the human. Adv
Anat Embryol Cell Biol. 2005;177:1–131.
18. Maroun LL, Graem N. Autopsy standards of body parameters and fresh
organ weights in unmacerated and macerated human fetuses. Pediatr
Devel Pathol. 2005;8:204–217.
19. Pinkus F. Die Entwicklungsgeschichte de Haut. In: Keibel F, Mall FP, eds.
Entwicklungsgeschichte des Mensschen. Leipzig, Germany: S. Hirzel;
1910:279–286.
20. Van der Putte SCJ. Anogenital ‘‘sweat’’ glands. Histology and pathology
of a gland that may mimic mammary glands. Am J Dermatopathol. 1991;
13:557–567.
21. Van der Putte SCJ. Mammary-like glands of the vulva and their disorders.
Int J Gynecol Pathol. 1994;13:150–160.
22. Pfeifer JD, Barr RJ, Wick MR. Ectopic breast tissue and breast-like sweat
gland metaplasias: an overlapping spectrum of lesions. J Cut Pathol.
1999;26:190–196.
23. Van der Putte SCJ. Apoeccrine glands in nevus sebaceous. Am J
Dermatopathol. 1994;16:23–30.
24. Belousova IF, Kazakov DV, Michal M, et al. Vulvar Toker cells: the long-
awaited missing link. A proposal for an origin-based histogenetic
classification of extramammary Paget disease. Am J Dermatopathol.
2006;28:84–86.
q 2011 Lippincott Williams & Wilkins