Download Article - Archive ouverte UNIGE

Article
Smooth muscle alpha-actin is a marker for hair follicle dermis in vivo
and in vitro
JAHODA, C. A., et al.
Abstract
We have examined the expression of smooth muscle alpha-actin in hair follicles in situ, and in
hair follicle dermal cells in culture by means of immunohistochemistry. Smooth muscle
alpha-actin was present in the dermal sheath component of rat vibrissa, rat pelage and human
follicles. Dermal papilla cells within all types of follicles did not express the antigen. However,
in culture a large percentage of both hair dermal papilla and dermal sheath cells were stained
by this antibody. The same cells were negative when tested with an antibody to desmin.
Overall, explant-derived skin fibroblasts had relatively low numbers of positively marked cells,
but those from skin regions of high hair-follicle density displayed more smooth muscle
alpha-actin expression than fibroblasts from areas with fewer follicles. 2-D SDS-PAGE
confirmed that, unlike fibroblasts, cultured papilla cells contained significant quantities of the
alpha-actin isoform. The rapid switching on of smooth muscle alpha-actin expression by
dermal papilla cells in early culture, contrasts with the behaviour of smooth muscle cells in
vitro, and has implications for [...]
Reference
JAHODA, C. A., et al. Smooth muscle alpha-actin is a marker for hair follicle dermis in vivo and
in vitro. Journal of Cell Science, 1991, vol. 99, no. Pt 3, p. 627-636
PMID : 1939373
Available at:
http://archive-ouverte.unige.ch/unige:11044
Disclaimer: layout of this document may differ from the published version.
Smooth muscle a-actin is a marker for hair follicle dermis in vivo and
in vitro
COLIN A. B. JAHODA 1 *, AMANDA J. REYNOLDS2•*, CHRISTINE CHAPONNIER2,
JAMES C. FORESTER3 and GIULIO GABBIANI*
1
Department of Biological Sciences, University of Dundee, Dundee DD1 4HN, Scotland
^Department of Pathology, University of Geneva, 1211 Geneva 4, Switzerland
^Department of Surgery, Ninewells Hospital and Medical School, Dundee, Scotland
* Present address: Department of Biological Sciences, University of Durham, Durham DH1 3LE, England
Summary
We have examined the expression of smooth muscle
a-actin in hair follicles in situ, and in hair follicle
dermal cells in culture by means of immunohistochemistry. Smooth muscle a-actin was present in the
dermal sheath component of rat vibrissa, rat pelage
and human follicles. Dermal papilla cells within all
types of follicles did not express the antigen. However, in culture a large percentage of both hair
dermal papilla and dermal sheath cells were stained
by this antibody. The same cells were negative when
tested with an antibody to desmin. Overall, explantderived skin fibroblasts had relatively low numbers
of positively marked cells, but those from skin
regions of high hair-follicle density displayed more
smooth muscle a-actin expression than nbroblasts
from areas with fewer follicles. 2-D SDS-PAGE
confirmed that, unlike nbroblasts, cultured papilla
cells contained significant quantities of the a-actin
isoform.
The rapid switching on of smooth muscle a-actin
expression by dermal papilla cells in early culture,
contrasts with the behaviour of smooth muscle cells
in vitro, and has implications for control of expression of the antigen in normal adult systems. The
very high percentage of positively marked cultured
papilla and sheath cells also provides a novel marker
of cells from follicle dermis, and reinforces the idea
that they represent a specialized cell population,
contributing to the heterogeneity of fibroblast cell
types in the skin dermis, and possibly acting as a
source of myofibroblasts during wound healing.
Introduction
dermal papillae after papilla removal, or lower follicle
amputation (Oliver, 19666; Jahoda et al. 1991).
The finding that cultured rat vibrissa dermal papilla
cells (Jahoda and Oliver, 1981) retain their capacity to
stimulate hair growth (Jahoda et al. 1984; Home et al.
1986) distinguishes them as a culture system with
particular relevance to known biological function. Other
investigations have used cultured dermal components
from human hair follicles (Messenger, 1989; Arai et al.
1989 for reviews). The morphology, behaviour and biosynthetic activites of papilla cells reflect their distinct in situ
properties (Jahoda and Oliver, 1984; Messenger et al. 1986;
Couchman, 1986), and it has been suggested that papilla
cells represent a specialized skin fibroblast population
(Jahoda and Oliver, 1984). This view is supported by the
finding that fibroblasts derived from skin explant cultures
are incapable of hair growth stimulation (Jahoda et al.
1984; Home et al. 1986). However, while certain features
such as the synthesis of basement membrane extracellular
components (Couchman, 1986) help to distinguish papilla
cells from skin fibroblasts, none can be expressly described
as a specific papilla cell marker.
This question fringes on the broader problem of
categorization of the cells that are present in the skin
dermis, and which commonly appear fibroblast-like when
Skin appendage development is engineered through a
series of interactions involving dermal mesenchyme and
epithelial-derived epidermis (Sengel, 1986, for review). In
hair follicle formation, the mesenchymal cells that form
the dermal component of the structure first become visible
as a cell agglomeration just below a primary epidermal
thickening. At an early stage, cell division within the
mesenchymal aggregation apparently ceases (Wessels and
Roessner, 1965), and these cells form the dermal papilla
component during follicular downgrowth. Connected with
the base of the adult dermal papilla, and separated from
the outermost epidermal layer of the follicle by a thick
basement membrane, is the external element of the follicle
mesenchyme, termed the dermal sheath.
Microsurgical manipulation of the rat vibrissa follicle
model system has shown that the adult dermal papilla is a
prerequisite for hair growth initiation and maintenance
(Oliver, 1966a,6,1967). Moreover, the adult dermal papilla
retains powerful inductive capabilities in that it induces
new hair follicle formation when associated with afollicular or wound epidermis (Oliver, 1970; Reynolds, 1989). The
vibrissa follicle dermal sheath cells also possess distinctive properties, since they can be the source of regenerated
Journal of Cell Science 99, 627-636 (1991)
Printed in Great Britain © The Company of Biologists Limited 1991
Key words: smooth muscle or-actin, hair follicle, nbroblasts,
dermal heterogeneity.
627
put into culture. It has become increasingly apparent that
dermal subpopulations exist within adult skin, and that
explant culture in particular may produce heterogeneous
fibroblast outgrowth for investigative purposes. One group
has recently attempted to tackle this problem by adopting
a classification system that separates fibroblasts into a
series of differentiation states according to their morphological and behavioural properties (Bayreuther et al.
1988).
A distinctive fibroblastic cell type first described in
granulation tissue, and associated with retractive processes of soft tissues, is the myofibroblast. These cells
possess a microfilamentous system akin to that observed
in cultured fibroblasts or smooth muscle cells. More
recently, myofibroblasts have been described in normal
tissues, where it has been proposed that they exert
contractile activities (Sappino et al. 1990). The use of
cytoskeletal markers such as desmin, an intermediate
filament protein typical of most muscle cells; smooth
muscle ir-actin, an actin isoform found in smooth muscle
cells; and smooth muscle myosin, has enabled several
myofibroblastic subpopulations in normal and pathological tissues to be defined. Using a monocolonal antibody
that specifically recognizes smooth muscle o--actin Skalli et
al. (1986) noted that some hair follicle dermis stained
positively. This, and resemblences between cultured
dermal papilla cells and myofibroblasts prompted us to
investigate whether hair follicle mesenchyme had smooth-
muscle cell cytoskeletal characteristics, an idea previously
mentioned by Couchman (1986). In the current study we
used immunolabelling methods to examine human and rat
follicular elements in vivo and in vitro, and report that the
two principal dermal components, the dermal papilla and
the dermal sheath, both express vascular smooth muscle a
actin in vitro.
Materials and methods
Rat tissues were derived from an inbred strain of PVG rats
(Colony Dundee University), and from Wistar rats. Human
material came from groin, scalp and scrotal skin biopsies as a
result of surgical procedures.
Pieces of rat skin were taken from body regions with high hair
follicle density (mystacial pad), reduced hair follicle density (ear
skin), and without follicles (footpad skin). Isolated vibrissa
follicles were dissected from the mystacial pad region as
previously described (Jahoda and Oliver, 1981), and human
follicles were dissected from groin, scalp and scrotal skin with
watchmakers' forceps and curved iridectomy scissors.
For tissue immunohistochemistry, specimens were embedded
in Tissue tek HI (Miles Scientific) water-based mounting medium
and snap frozen over liquid nitrogen. Cryostat sections of 6 /on
were cut at -20°C and air dried prior to labelling.
Cell cultures
Dermal papilla cultures were established from intact papilla
explants (Fig. 1) dissected from vibrissa, and human follicles, as
Fig. 1. An isolated vibrissa dermal papilla prior to being put into culture. Note how papillae can be cleanly separated from the
follicle epidermis, and the dermal sheath. Phase-contrast; x360.
Fig. 2. A human dermal papilla still attached to dermal sheath (ds) that has been inverted during the dissection procedure. The
two are then separated at the narrow basal stalk region (arrowed). Phase-contrast; X170.
Fig. 3. Dermal papilla cell culture treated with pre-immune serum in the staining procedure as one of the immunohistochemical
controls.
628
C. A. B. Jahoda et al.
previously described (Jahoda and Oliver, 1981; Messenger, 1984).
Vibrissa dermal sheath cultures were initiated from the thin
circular envelope that remains attached to the dermal papilla
after removal of the epidermal component during dissection.
Human dermal sheath cells were cultured from all of the bulbar
dermis remaining after the papilla and epidermal matrix has
been excised (Jahoda et al. unpublished). Briefly, the bases of
anagen follicles were cleared of any adherent tissue, under a
binocular dissecting microscope (x20) in Minimal Essential
Medium containing penicillin/streptomycin at 50 units ml" 1
(both Gibco), and their bulb regions were excised and transferred
to dishes with fresh medium. The follicle bulbs were then inverted
using a combination of finely tipped watchmakers' forceps and
sharp tungsten needles. This invariably led to separation of the
epidermal component from the dermal papilla and attached
sheath of follicular dermis (Fig. 2). The sheath was then easily
separated from the papilla with tungsten needles.
Rat skin fibroblast cultures came from explants from the three
above mentioned regions of skin: mystacial pad, ear and footpad.
Where possible, fibroblast cultures were set up from the same
individuals as were used to provide papilla and dermal sheath
cells. Areas of skin were cleaned, and surface hair fibres removed,
before the tissue was chopped into small pieces. Fibroblast
cultures were set up from explants of human skin in a similar
manner.
For immunolabelling of cells during early outgrowth, explants
were established directly on top of glass coverslips in 35 mm
culture dishes. For later observation, passaged cells were seeded
at various densities on top of coverslips in 35 mm dishes. All
cultures were fixed and permeabilized by exposure to methanol
for 30min at — 20 "C before marking.
Immunohistochemistry
Cultured cells and sections were incubated with a monoclonal
antibody to smooth muscle o^actin (anti-asm-1, Skalli et al. 1986)
diluted 1:10 in phosphate-buffered saline (PBS), pH7.4, at room
temperature for one hour. After thorough washing in PBS,
material was exposed to biotinylated anti-mouse antibody (BRL),
diluted 1:90, for one hour, washed again, and incubated with
fluorescein-linked streptavidin, diluted 1:100, for the same length
of time. Specimens were mounted in Citifluor (Agar aids)
medium, and examined and photographed on a Zeiss ICM 405
inverted microscope equipped with epifluorescence. Alternatively, for more permanent preparation the material was stained
as described above, but with horseradish peroxidase-linked
steptavidin. Estimations of positive smooth muscle actin marking
in cell cultures, were carried out by examinining one hundred
cells in randomly chosen fields. Rat tissues and cell cultures were
also labelled with a polyclonal antibody that recognizes all actin
isoforms (Skalli et al. 1986). The same material was tested for
desmin with an affinity-purified rabbit anti-desmin polyclonal
antibody (Kocher et al. 1984) and a monoclonal antibody to
vimentin (Transformation Research, Inc. Framingham, MA,
USA). As one of the control procedures, specimens were treated
with pre-immune sera. None of the control sections or cultures
treated in this fashion showed any significant fluorescence
(Fig. 3).
Gel electrophoresis
For 2-D PAGE, passage-two cultured dermal papilla cells and
fibroblasts were dissolved in sample buffer containing 1 % sodium
dodecyl sulphate, 1% dithiothreitol, limn Nff-p-tosyl-L-arginine
methyl ester and lmM phenylmethylsulphonyl fluoride. In the
first dimension, actin isoforms were separated by isoelectric
focusing according to OTarrell (1975). The pH gradient was
established with 2% preblended Ampholines. pH 4.0-6.5 (LKB
Co., Lucerne, Switzerland). The second dimension was run on
SDS-10% PAGE. For the different cell types, equal amounts of
sample protein were loaded, following protein determination by
the Bradford (1976) method.
Electron microscopy
Specimens were fixed in 2% glutaraldehyde in 0.1M cacodylate
buffer (pH7.2) and 0.05% Ruthenium Red (to enhance visualization of extracellular material) for two hours. Postfixation was in
1% osmium tetroxide, with 0.1M cacodylate buffer and 0.025%
Ruthenium Red, for three hours at 4°C. Material was dehydrated
in increasing concentrations of ethanol and propylene oxide
before being embedded in Epon resin. Ultrathin sections were
stained with uranyl acetate, followed by lead citrate, and
examined with a Zeiss EM 109 electron microscope.
Results
Staining of normal tissue sections
Rat vibrissa follicles. In anagen vibrissa follicles,
overall actin distribution, as revealed by the general
antibody, was widespread, with the outer root sheath of
the follicle epithelium strongly marked (Fig. 4).
Smooth muscle u-actin labelling of mid-anagen vibrissa
follicles revealed that the dermal papilla component was
virtually unmarked. However, the lower dermal sheath of
the follicle was clearly labelled, except for a region
surrounding part of the bulb at the base of the follicle
(Fig. 5). This marking clearly delineated a line of dermal
sheath cells just exterior to the glassy membrane (Fig. 5)
but from about half way up the follicle, staining became
patchy, and then petered out (Fig. 6). Follicles were not
labelled by the antibody to desmin but vibrissa sections
acted as an excellent positive control for this antibody,
because follicles could be left with some of their surrounding muscle tissue attached, and these muscle blocks
always displayed intense staining (Fig. 7).
Rat skin. Foot pad skin was virtually unmarked, with
only dermal vasculature and Meissner corpuscles positively labelled. Ear and whisker pad skin both showed
widespread follicular staining with the total actin antibody (Fig. 8). As in vibrissa follicles, smooth muscle aactin label was restricted to the dermal sheath, with no
expression at all inside the dermal papilla except for
occasional marking of blood vessels (Fig. 9). However, the
smaller pelage follicles differed from vibrissae in that
labelling extended right around the follicle base. There
was also an indication that staining of the dermal sheath
was absent or reduced in follicles that had shortened, and
were in the telogen phase of the hair cycle. Intriguingly,
longitudinal sections cut at the outside edge of follicles
revealed a highly organized pattern of smooth muscle aactin within the dermal sheath. The label showed up as a
stack of concentric horizontal rings around the follicle
circumference (Fig. 10). This tied in with parallel observations from median sections where the outer line of
dermal sheath fluorescence was punctuated rather than
continuous.
Human follicles. Anagen human hair follicles were also
marked around the dermal sheath and, like rat vibrissa
follicles, label was not always strong outside the lowest
part of the follicle bulb. In some specimens the basal stalk
below the papilla was strongly marked (Fig. 11), but
papillae were never stained. Unlike vibrissa follicles,
dermal sheath labelling extended into the upper half of
appendages. Detailed exmination showed that the antibody was only staining the internal cell layer of the
dermal sheath. When this was cut tangentially, it showed
the same circular horizontal labelling observed in the
dermal sheath of rat pelage follicles. The outer connective
tissue sheath remained unmarked, although blood vessels
peripheral to the follicle were often fluorescent (Fig. 12).
Smooth muscle a-actin in hair follicle dermis
629
Passaged cell culture labelling
Smooth muscle a-actin. In early-passage rat dermal
cells (passages 1 to 3), hair follicle-derived dermal papilla
and dermal sheath cells were both highly positive for
smooth muscle o--actin; inside prominent stress fibres
(Figs 13, 14). Although the amount of positive staining
was slightly variable between different lines, it was
estimated that over 75 % of papilla cells, and more than
95% of dermal sheath cells, were marked in all these
630
C. A. B. Jahoda et al.
cultures. Indeed, several dermal sheath cell cultures
appeared to have nearly every cell stained positively,
although within groups individual cells displayed different intensities of label. The distribution of marker within
cells reflected previous observations, in that marking of
stress fibres did not always extend to the periphery of the
cell projections. However, intricate patterns of cytoskeletal architecture were visible near the periphery of
many larger flattened cells, perhaps associated with
Figs 4-12. In situ labelling.
Fig. 4. Longitudinal section through a rat vibrissa follicle
showing total actin expression. Within the follicle, the most
strongly marked elements are the epidermal outer root sheath
(arrowed) and the lower dermal sheath, just outside the
dermal-epidermal junction. x55.
Fig. 5. The same follicle labelled with smooth muscle oactin.
Apart from blood vessels outside of appendage, only the lower
dermal sheath is stained, by a discrete line of marking. This
label does not extend proximally around the base of the bulb,
becomes increasingly fragmented and then disappears, above
the lower third of the structure. The dermal papilla is totally
unmarked. X55.
Fig. 6. Detail of dermal sheath marking shows that
fluorescence is located just outside the glassy membrane. The
line of label is seen to be punctuated rather than continuous,
and becomes more diffuse and fades distally. x220.
Fig. 7. Vibrissa follicle stained with antibody to desmin.
Within the follicle neither the dermal papilla (p) nor the
dermal sheath is marked, but the muscle blocks outside the
follicle capsule are strongly labelled. xllO.
Fig. 8. Pelage follicles in longitudinal section stained with
antibody to total actin. Marking is clearest in the outer root
sheath of the epidermis, and the dermal sheath, x 120.
Fig. 9. The same region of skin stained with smooth muscle
a<-actin. As with vibrissa follicles, label is restricted to the
dermal sheath; however, in the case of anagen pelage follicles
staining extends right around the follicle bulb proximally. The
dermal papilla is not labelled. XllO.
Fig. 10. Closer view of the dermal sheath region of a pelage
follicle, stained with smooth muscle o^actin. Part of the section
shows that actin within the sheath cells is clearly organized in
a series of horizontally stacking concentric rings. X220.
Fig. 11. The base of an anagen hair follicle from human groin
skin labelled for smooth muscle a^actin. There a strong plug of
marking in the basal stalk of the papilla, no label in the
papilla (p) itself, and a thin line of marking on the inside of
the connective tissue (dermal) sheath. x90.
Fig. 12. Detail of the mid region of a human scalp follicle
labelled for smooth muscle <*-actin. As well as the inner dermal
sheath marking, which extends high up into the follicle
distally, associated blood vessels are frequently marked
(arrowed). X130.
adhesion plaques. Further passaging reduced the estimated number of papilla or sheath cells with positive
marking, and with subculture the intensity of label also
diminished within individual cells; however, passage six
or seven papilla and sheath cells still remained over 60 %
marked.
All of the fibroblast cell lines from the various regions of
skin had less than 20 % of their cells positively stained by
smooth muscle o"-actin. Foot pad and ear skin fibroblasts
contained 10 % or less of positive cells, and these were
often distributed in patches, presumably as a result of
colony formation around a single parent cell (Fig. 15).
Large areas of these cultures were often devoid of label, in
stark contrast to the universal staining of fibroblasts by
the total actin antibody (Fig. 16). Fibroblasts from the
mystacial region displayed a slightly higher number (up to
15%) of positively labelled cells. Once more, serial
passaging did not alter the basic staining pattern, so that
the fibroblasts retained small colonies of cells that
contained smooth muscle cr-actin after extended passaging, but again with considerable variation in staining
intensity between cells.
Desmin and vimentin. Vibrissa dermal papilla and
sheath cells, and rat skin fibroblasts, were all negative to
the anti-desmin antibody, but positively marked by the
antibody to vimentin (Fig. 17).
Explant outgrowth labelling
Vibrissa dermal papilla and dermal sheath. Cells
around fragments of dermal sheath tissue were universally marked by the smooth muscle o--actin antibody as
soon as they emerged from the explants. By contrast, cells
from dermal papilla explants that had attached and
flattened were initially negative. Then, between the first
and second week in culture as cells began to move and
spread out, their appearance altered. Cells at the
periphery of the explants became positive for the smooth
muscle a-actin antibody, while the central mass still
remained negative (Figs 18, 19). In papilla explant cultures that had been established for 14 days or more, when
cells had further dispersed, an increasingly high proportion of them were marked.
Mystacial pad skin. In the cell-covered substratum
surrounding pieces of mystacial pad skin, whole regions
were completely unmarked, and it was clear that positive
staining among the fibroblast cells was highly localized.
Further scrutiny of areas with labelled cells revealed that
label was usually found around pelage hair follicles.
Follicles were visible in the skin explants and groups of
marked cells were seen around the outline of these
follicular 'contaminants' (Fig. 20).
Human cells. Between passages 1 and 3, widespread
staining of the dermal papilla and dermal sheath cells
with smooth muscle <r-actin was evident. Around 90 % of
dermal sheath cells and 60 to 80 % of dermal papilla cells
were positively labelled (Fig. 21), but once again individual cells showed different staining intensity. Laterpassage human dermal papilla and dermal sheath cultures from groin skin follicles had reduced numbers of
strongly marked cells - in one passage 7 papilla culture as
low as 40%. This phenomenon was less apparent with
papilla cells from scrotal skin follicles. Human skin
fibroblast cultures from all of the studied regions contained very small numbers of marked cells (Fig. 22): in
some cases less than 2 %.
Explant cultures showed precisely the same trends as
previously described for rat follicles. At an early stage in
culture, dermal sheath explants produced arrays of cells in
parallel lines, all positively marked and emanating from
the central tissue (Fig. 23). At the same time, dermal
papilla cultures were again unmarked in the central
papilla core, and only sparse labelling was seen where
cells had initially settled and spread outwards (Fig. 24).
Fibroblast cells that emerged from pieces of skin in
explant culture were again almost totally negative. As
with rat material, positively stained cells were localized
most clearly in the vicinity of hair follicles present in the
skin.
2-D PAGE. When the pattern of actin isoforms was
examined by 2-D PAGE, cultured dermal papilla and
dermal sheath cells were found to contain three spots
corresponding to a--, ft- and y-actins, as previously
described (Kocher et al. 1984; Skalli et al. 1987) (Fig. 25A).
Only two spots, identified as £- and y-actins, were seen in
gels prepared from equivalent skin fibroblast samples
(Fig. 25B).
Transmission electron microscopy. Under phase-contrast microscopy, stress fibres are prominent in rat dermal
papilla and dermal sheath cell monolayers. A behavioural
feature of these cell types is their propensity to form
aggregates, and ultrastructural examination of cells on
plastic susbtrata, or on top of other cells within clumps,
revealed thick microfilament bundles (Fig. 26). More
intriguing is the finding that in situ dermal papilla cells
Smooth muscle a-actin in hair follicle dermis
631
displayed extensive cytoplasmic processes, many of which
were packed with microfilaments. Sometimes, elongated
cells were observed with bulbous regions of cytoplasm
bridged by narrower sections rich in microfilaments.
These cells were undoubtedly in the process of moving
(Figs 27, 28), and appeared most often during periods of
changing activity, at the beginning or end of the hair
cycle.
Figs 13-17. Staining of passaged rat cells.
Fig. 13. Passage 3 rat dermal papilla cells, many of which are overlapping, show widespread labelling with smooth muscle oactin.
x90.
Fig. 14. Rat dermal sheath cells after three passages display strong marking with the smooth muscle a-actin antibody. Note the
broad flattened morphology of this cell type. x90.
Fig. 15. One a-actin-labelled cell against a field of unstained cells in a passage 2 ear skin fibroblast culture. x56.
Fig. 16. Antibody to total actin marks stress fibres in every cell in this passage 2 skin fibroblast culture. X220.
Fig. 17. Vimentin filaments labelled in a passage 2 dermal papilla cell. x380.
Fig. 18. A dermal papilla explant after 11 days. The centre of
the explant (e) has collapsed, and papilla cells have attached to
the substratum but are still relatively small and tightly
bunched. There is almost no smooth muscle o^actin staining in
this central area but some cells are marked towards the edge
of the field. X90.
Fig. 19. Lower magnification view of the same explant culture.
Cells at the core of the explant are unmarked, but towards the
periphery, where they have become more spaced out and
flattened, there is widespread marking. x56.
Fig. 20. Low magnification view, showing two hair follicles at
the edge of a skin explant, and two clumps of smooth muscle
cr-actin-labelled cells in close proximity (arrowed). In fibroblast
explant culture, the only large groups of positively stained
cells were invariably associated with hair follicles, and were
usually in aggregated form. X56.
Fig. 21. Human groin follicle dermal papilla cells after their
632
C. A. B. Jahoda et al.
second passage, marked with smooth muscle a^actin. Note the
relatively broad morphology of some cells. x90.
Fig. 22. A second-passage groin skin fibroblast culture,
showing small numbers of labelled cells, against an unstained
background of confluent cells. Many parts of human skin
fibroblast cultures were completely unmarked. x90.
Fig. 23. Groin dermal sheath cells positively stained for aactin streaming outwards from a central tissue explant, which
is itself well marked. Ten day culture. x90.
Fig. 24. A human groin dermal papilla explant (p) equivalent
to that in Fig. 23, is not positively labelled by the tr-actin
antibody at ten days. Of the cells that have emerged and settled
in the vicinity, a relatively small proportion are marked. x225.
Fig. 25. (A),(B) Two-dimensional polyacrylamide gel
electrophoresis (2-D PAGE) of second-passage cultured cells.
Dermal papilla cells contain a<-, /?- and y-actins (A), wheras
skin fibroblasts only show /S- and ^actin isoforms (B).
Discussion
It appears increasingly clear that the cells that are
collectively called fibroblasts or stromal cells display a
phenotypic diversity that was not previously suspected on
morphological grounds (Sappino et al. 1990). As the
predominant cell type in connective tissue, the fibroblast
fulfils a variety of essential functions, and plays a pivotal
role in tissue repair. It is also implicated in mesenchymal-
epithelial interactions that are known to govern organogenesis. The recognition of heterogeneity in the cytoskeletal differentiation programme of fibroblastic cells
may therefore help us to understand further the specialized activities of the different subpopulations. Cytoskeletal markers are useful for this purpose, and are
particularly suited to identifying potential contractile
capacities of fibroblastic cells.
As briefly observed previously (Skalli et al. 1986), the
Smooth muscle a-actin in hair follicle dermis
633
Fig. 26. Transmission electron microscopy of a cultured dermal papilla cell sitting on top of another within a typical papilla cell
aggregate. A prominent microfilament band (arrowed) lies at the base of this cell, x 16 000.
Fig. 27. Transmission electron microscopy of a vibrissa follicle dermal papilla at catagen. One papilla cell (arrowed) is highly
elongated, and is apparently in the process of moving position. X4000.
Fig. 28. Higher magnification view of the microfilament structure (arrowed) in the same cell. X22000.
dermal sheath region of the hair follicle labels positively
with smooth muscle a'-actin, and present work shows that
in vibrissa follicles dermal sheath marking is often
restricted to the lower half of the follicle. This observation
may relate to experimental work where, on removal of the
lower third of vibrissa follicles, a new papilla is re-formed
from lower dermal sheath cells. The latter are observed by
light and electron microscopy migrating from the sheath
region through the thick specialized basement membrane
of the follicle to the site of papilla formation (Oliver,
19666: Jahoda et al. 1991). When the follicle is ablated
higher up, papilla regeneration does not take place
(Oliver, 1966a). While sheath cells can definitely alter
their phenotype to become papilla cells, it is less clear
whether the reverse is true, although characteristic
Alcian Blue marking of papilla cells suggests that they at
least partially assume the role of dermal sheath in new
follicles induced by papilla interaction with skin epidermis (Oliver, 1970). Human follicle papilla regeneration
can also be elicited (Jahoda et al. 1989), which suggests
that a similar papilla/sheath cell relationship exists in the
lower human follicle. The finding that only the internal,
horizontally aligned layer of the human dermal sheath
has smooth muscle a'-actin marking, emphasizes that this
structure essentially consists of two layers (Kligman,
1988), each of which may have a separate function. This
may also hold true for the rodent pelage follicle sheath,
and indeed for other follicle types.
Here we report for the first time that cultured cells from
both the dermal papilla and the dermal sheath are marked
positively with smooth mucle o--actin-specific antibodies,
but react negatively with desmin antibody. Like skin
fibroblasts, both sheath and papilla cells contained
vimentin. While counts of numbers of smooth muscle a'actin positive cells acted as a useful indicator, they could
not be regarded as accurate assessments of overall content
because of individual variation in labelling intensity.
However, 2-D PAGE analysis confirmed that the smooth
muscle a'-actin isoform is a major component in the
cytoskeletal make up of follicle dermal cells, but not an
appreciable element in skin fibroblasts.
634
C. A. B. Jahoda et al.
The change of papilla cells from negatively stained in
vivo to positively marked in vitro has a parallel in cultured
eye lens cells (Schmitt-Graff et al. 1990), and it distinguishes these special types of fibroblastic cells from
adult smooth muscle cells whose smooth muscle a'-actin
content decreases when placed in culture. During atherosclerosis, smooth muscle cells also lose smooth muscle a'actin and revert to having a /3-isoform predominance,
similar to that seen in embryonic tissue (Gabbiani et al.
1984; Kocher and Gabbiani, 1987). While dermal sheath
cells did not alter their smooth muscle a--actin expression
in culture, a feature of the current work was the very rapid
switch on of smooth muscle o--actin expression by papilla
cells around the second week of explant culture. The
finding that cells that moved into spaces away from the
original papilla mass immmediately started to express
smooth muscle o--actin, and that the whole culture could
transform in a matter of days, depending on the speed of
dispersal, was intriguing. It suggests that perhaps the
papilla cells are held in a state of non-expression by
modulatory or inhibitory factors within the follicle, which
are lost in the culture environment; this is unusual in
relation to control of smooth muscle differentiation. On
this point it is worth noting that the dermal papilla has a
glycosaminoglycan (GAG)-rich extracellular matrix, and
that heparin is one extracellular component that is known
to influence smooth muscle a^actin expression in vivo
(Clowes et al. 1988), and in vitro (Desmouliere et al. 1990).
The sudden change in marking also provided a possible
link between smooth muscle a<-actin expression and cell
motility, since papilla cells took up this label at the time
when they began to move. In this context, ultrastructural
observations of migratory papilla cells in situ stress the
need to consider the papilla cell population as dynamic and
not static, particularly during periods of morphogenetic
activity.
Another striking finding was the relative homogeneity
of the dermal sheath cell cultures, where an exceptionally
high proportion of cells stained positively with smooth
muscle a'-actin antibody. This means that dermal sheath
could become a valuable source of culture material for
studies of smooth muscle a--actin as a cell cytoskeletal
component.
In relation to papilla cell cultures, the reduction in
numbers of positively stained cells over time could be a
reflection of the ageing of the cell lines. It has previously
been observed that human papilla cells have a relatively
short lifespan before they become 'senescent' (Messenger,
1989). The low numbers of marked cells seen in a passage 7
human papilla cell culture coincided with a slowing down
in cell division and other signs of senescence at the time of
testing.
Although rat skin fibroblast cultures generally had low
numbers of cells containing smooth muscle a--actin, those
from hairy skin (whisker pad) appeared to have more
labelled cells than cultures from less hairy regions
(footpad or ear). This tied in with the observation that
early on in skin explant culture positively stained cells
had a patchy distribution, and were seen grouped at the
periphery of hair follicles. In rat fibroblast cell lines we
have observed heterogeneity of morphology, and from time
to time have noted aggregations of cells, similar to those
observed in dermal papilla (Jahoda and Oliver, 1984) or
dermal sheath (Jahoda et al. unpublished) cultures. From
these observations, one idea is that fibroblast cell lines
from skin explants that incorporate hair follicles are
'contaminated' by follicular mesenchyme cells. This could
have significance with regard to fibroblast lineage relationships. It also might need to be taken into consideration in experimental work on skin fibroblasts, where the
original tissue source is hair-bearing skin. However, the
finding of the same phenomenon in human explants raises
a potentially more important idea. Myofibroblasts that are
involved in skin wounding have specific properties, some
of which resemble smooth muscle cells (Skalli and
Gabbiani, 1988). Their origin in wounds appears to be local
(Ross et al. 1970), and they are thought to be derived from
fibroblasts (Skalli and Gabbiani, 1988, for review). A
recent study of open cutaneous wounds in the rat has
suggested that fibroblasts in granulation tissue transform
gradually into myofibroblasts, and during a restricted
period express a smooth muscle marker (Darby et al. 1990).
In superficial cutaneous wounds, a principal source of
replacement skin epidermis comes from outer root sheath
epidermal cells of the hair follicle (Eisen et al. 1955;
Bullough, 1970; Krawczyk, 1971; Pang et al. 1978). Given
the similarity between myofibroblast and dermal sheath
cell characteristics, it is reasonable to suggest that the
dermal sheath cell layer (previously shown to be a
reservoir of papilla cells for the follicle, and situated
adjacent to the outer root sheath epidermal cells) might be
a specific source of myofibroblasts in cutaneous wounds.
This hypothesis is currently the subject of study.
Concerning hair follicles, it is worth considering that
minoxidil, the only human hair growth-promoting compound to have had reported success (Katz, 1988), was
developed as an antihypertensive drug, with its principal
mode of action on the vascular system (Zinns, 1988), and a
known capacity in vivo to relax vascular smooth muscle
yia an active metabolite, minoxidil sulphate (Meisheri et
al. 1988). More intriguingly, smooth muscle cell features
have been described in the fibroblasts of normal organs
where some tissue reorganization occurs either continuously or periodically (Sappino et al. 1990, for a review), and
where contractile function might be involved. One of us
(Reynolds et al. unpublished data) has observed changes in
smooth muscle a^actin expression in the dermal sheath
during the hair growth cycle, and has suggested that the
smooth muscle characteristics of dermal sheath cells could
relate to contractile processes that might control hair
follicle shortening and upward fibre movement during the
hair growth cycle. In this context, the observed alignment
of smooth muscle a^actin in the inner follicle dermal
sheath as a series of concentric circular rings clearly
implies functional importance, particularly as this organization appears to be common to many follicle types.
This study provides a new marker for hair follicle
dermis, and possibly an opportunity of examining the hair
follicle from a slightly different evolutionary and developmental point of view. It will certainly assist in those
investigations where hair follicle cells need to be distinguished.
C.A.B.J. held a Royal Society, 1983 University Research
Fellowship. This work was supported in part by the Swiss
National Science Foundation, grant nos 3.108- 0.88 and 3126614.89, and the Wellcome Trust.
References
ARAI, A , KATSUOKA, K., KIESWETTER, F., SCHELL, H. AND HORNSTEIN, O.
P. (1989). Effects of testosterone, dihydrotestosterone and estradiol on
the growth behavior of cultured hair bulb papilla cells and root
aheath flbroblaste. In Trends m Human Hair Growth and Alopecia
Research (ed. D. Van Neste, J. M. Lachapelle and J. L. Antoine), pp.
99-104. London: Kluwer.
BAYREUTHER, K., RODKMANN, H. P., FRANCZ, P. I. AND MAIER, K. (1988).
Differentiation of fibroblast stem cells. In Stem Cells (ed. B. I. Lord
and T. M. Dexter), J. Cell Sci. Suppl. vol 10, pp. 115-130. Cambridge:
The Company of Biologists Ltd.
BRADFORD, M. M. (1976). A rapid and sensitive method for the
quantification of microgram quantities of protein utilizing the
principle of protein-dye binding. Analyt. Biochcm. 72, 248—254.
BULLOUQH, W. S. (1970). Epithelial Repair and Regeneration (eds. J. E.
Dunphy and W. Van Winkle), pp. 35-46. New York- McGraw Hill.
CLOWES, A. W., CLOWES, M. M., KOCHER, O., ROPRAZ, P., CHAPONNIER,
C. AND GABBIANI, G. (1988). Arterial smooth muscle cells in vivo.
relationship between actin isoform expression and mitosis and their
modulation by heparin. J. Cell Biol. 107, 1937-1945.
COUCHMAN, J. R. (1986). Rat hair follicle dermal papillae have an
extracellular matrix containing basement membrane components.
J. invest. Derm. 87, 762-767.
DARBY, I., SKALJJ, O. AND GABBIANI, G. (1990). ^Smooth muscle actin is
transiently expressed by myofibroblasts during experimental wound
healing. Lab. Invest 63(1), 21-29.
DESMOULIERE, A., RUBBIA-BRANDT, L. AND GABBIANI, G. (1991).
Modulation of actin isoform expression in cultured arterial smooth
muscle cells by heparin and culture conditions. Arteriosclerosis (in
press).
EISEN, A. Z., HOLYOKE, J. B. AND LOBITZ, W. C. (1955). Responses of the
superficial portion of the human pilosebaceous apparatus to controlled
injury. J. invest. Derm 25, 145-156.
GABBIANI, G., KOCHER, O., BLOOM, W. S., VANDERKERKHOVE, J. AND
WEBER, K. (1984). Actin expression in smooth muscle cells of rat
aortic intimal thickening, human atheromatous plaque and cultured
rat aortic media. J. elm. Invest. 73, 148-152
HORNE, K. A., JAHODA, C. A. B. AND OLIVER, R. F. (1986). Whisker
growth induced by implantation of cultured vibrissa dermal papilla
cells in the adult r a t J. Embryol. exp. Morph. 97, 111-124.
JAHODA, C. A. B., HORNE, K. A., MAUCEE, A., BARD, S. AND SENQEL, P.
(1991). Cellular and extracellular involvement in lower rat vibrissa
follicle regeneration. Development (in press)
JAHODA, C. A. B., HORNE, K. A. AND OLIVER, R. F. (1984). Induction of
hair growth by implantation of cultured dermal papilla cells. Nature
311, 560-562.
JAHODA, C. A. B. AND OLIVER, R. F. (1981). The growth of vibrissa
dermal papilla cells in vitro. Br J. Derm. 105, 623-627.
JAHODA, C. A. B. AND OLIVER, R. F. (1984). Vibrissa dermal papilla cell
aggregative behaviour in vivo and m vitro. J. Embryol. exp Morph.
79, 211-224.
JAHODA, C. A. B., OLIVER, R. F., FORRESTER, J. C. AND HORNE, K. A.
(1989). Regeneration in human hair follicles grafted onto athymic
mice. J. invest Derm. 93<3), 452.
KATZ, H. I. (1988). Topical Minoxidil: Review of Efficacy. In Clinics in
Dermatology. Androgenic Alopecia From Empiricism to Knowledge (ed.
R. L. de Villez), pp 195-199. Philadelphia: Lippincott.
Smooth muscle a-actin in hair follicle dermis
635
KLIOMAN, A L. (1988). The comparative histopathology of male-pattern
baldness and senescent baldness. In Clinics m Dermatology.
Androgenic Alopecia From Empiricism to Knowledge (ed. R. L. de
Villez), pp. 108-118. Philadelphia: Lippincott.
KOCHKR, 0. AND GABBIANI, G. (1987). Analysis of (^smooth muscle actin
mRNA expression in rat aortic smooth-muscle cells using a specific
cDNA probe. Differentiation 34, 201-209.
KOCHKE, 0., SKALLI, O., BLOOM, W S. AND GABBIANI, G. (1984).
scanning and transmision electron microscopic study. Am. J. Anat.
153, 177-192.
REYNOLDS, A. J. (1989). In vivo and in vitro studies of isolated and
interacting dermal and epidermal components of the integument.
Ph.D. thesis. University of Dundee.
Ross, R., EVERETT, N. B. AND TYLER, R. (1970). Wound healing and
collagen formation. VI. The origin of the wound fibroblast studied in
parabiosis. J. Cell Biol. 44, 645-654.
Cytoskeleton of rat aortic smooth muscle cells. Normal conditions and
experimental intimal thickening. Lab. Invest. SO, 645-652.
KRAWCZYK, W. S (1971). A pattern of epidermal migration during
wound healing. J. Cell Bwl. 49, 247-263.
SAPPINO, A. P., SCHURCH, W. AND GABBIANI, G. (1990). The
MSISHERI, K. D , CIPKTJS, L. A AND TAYLOR, C. J. (1988). Mechanisms of
SCHMITT-GRAFT, A., PAU, H., SPAHR, R., PIPER, H , SKALLI, O. AND
action of minoxidil sulfate-induced vasodilation: a role for increased
K+ permeability. J. Pharm. exp. Ther. 245, 751-760.
MESSENGER, A. G (1984). The culture of dermal papilla cells from
human hair follicles. Br. J. Derm. 110, 685-689.
MESSENGER, A. G. (1989). Isolation, culture and in vitro behaviour of
cells isolated from papillae of human hair follicles. In Trends in
Human Hair Growth and Alopecia Research (ed. D. Van Neste, J. M.
Lachapelle and J. L. Antoine), pp. 57-66. London- Kluwer.
GABBIANI, G. (1990). Appearance of alpha-smooth muscle actin in
human eye lens cells of anterior capsular cataract and in cultured
bovine lens forming cells. Differentiation 43, 115-122.
SKNGKL, P. (1986). Epidermal—dermal interaction. In Biology of the
Integument, vol. 2, Vertebrates (ed. J Bereiter-Hahn, A. G. Matoltsy,
and K. S. Richards), pp. 374-408. Berlin, Heidelberg: Springer-Verlag.
SKALLI, O. AND GABBIANI, G. (1988). The biology of the myofibroblast
relationship to wound contraction and fibrocontractive disease. In The
Molecular and Cellular Biology of Wound Repair (ed. R A. F. Clark
and P. M. Henson), pp.373-402. New York, London: Plenum press.
MESSENGER, A. G., SENIOR, H. J. AND BLEEHEN, S. S. (1986). The in vitro
properties of dermal papilla cell lines established from human hair
follicles. Br. J. Derm. 114, 426-430.
OTARRELL, P. H. (1975). High resolution two-dimensional
electrophoresis of proteins. J. biol. Chem. 250, 4007-4021.
OLIVER, R. F. (1966a). Whisker growth after removal of dermal papilla
and lengths of the follicle in the hooded rat. J. Embryol. exp. Morph.
15, 331-347.
OLIVER, R. F. (19666). Histological studies of whisker regeneration in
the hooded r a t J. Embryol. exp. Morph. 18, 43-51.
OLIVER, R. F. (1967). The experimental induction of whisker growth in
the hooded rat by implantation of dermal papillae. J. Embryol. exp.
Morph. 18, 43-51
OLIVER, R F. (1970). The induction of follicle formation in the adult
hooded rat by vibrissa dermal papillae. J. Embryol. exp. Morph. 23,
219-236.
differentiation repertoire of flbroblaatic cells: expression of
cytoskeletal proteins as marker of phenotypic modulations. Lab.
Invest. 63, 144-161.
SKALLI, 0., ROPRAZ, P., TRZECIAK, A., BKNZONANA, G., GILLESSEN, D
AND GABBIANI, G. (1986). A monoclonal antibody against cr-Bmooth
muscle actin: A new probe for smooth muscle differentiation. J. Cell
Biol. 103 (no. 6, P t 2), 2787-2796.
SKALLI, 0., VANDEKERCKOVE, J. AND GABBIANI, G. (1987). Actin-isoform
pattern as a marker of normal or pathological smooth muscle and
fibroblastic tissues. Differentiation 33, 232-238.
WESSELB, N. K. AND ROESSNER, K. D. (1965). Non-proliferation in dermal
condensations of mouse vibrissae and pelage hairs. Devi Biol. 12,
419-433.
ZINNS, G. R. (1988). The history of the development of minoxidil. In
Clinics in Dermatology. Androgenic Alopecia From Empiricism to
Knowledge (ed. R. L. de Villez), pp. 132-143. Philadelphia: Lippincott.
PANG, S. C , DANIELS, W. H. AND BUCK, R. C. (1978). Epidermal
migration during the healing of suction blisters in rat skin: a
636
C. A. B. Jahoda et al.
(.Received 13 December 1991 - Accepted, in revised form, 5 April 1991)