1761
Journal of Cell Science 107, 1761-1772 (1994)
Printed in Great Britain © The Company of Biologists Limited 1994
Transfection of rat dermal papilla cells with a gene encoding a temperaturesensitive polyomavirus large T antigen generates cell lines retaining a
differentiated phenotype
Wendy Filsell*, Julie C. Little, Amanda J. Stones, Stewart P. Granger and Susan A. Bayley†
Unilever Research, Colworth House, Sharnbrook, Bedford MK44 1LQ, UK
*Author for correspondence
†Present address: MRC Cambridge Centre for Brain Repair, Department of Clinical Veterinary Medicine, University of Cambridge, Cambridge CB3 0ES, UK
SUMMARY
The dermal papilla is a discrete group of cells at the base
of the hair follicle and is implicated in controlling the hair
growth cycle. Early passage dermal papilla cells can induce
hair growth in vivo, but, upon further culturing, this
property is lost. In order to study the events occurring in
hair induction, a representative dermal papilla cell line was
required. We have transfected passage 1 rat vibrissa
dermal papilla cells with a polyomavirus large T gene
encoding a temperature-sensitive T antigen, and generated
permanent cell lines in which the immortalizing function
can be switched off by temperature shift. The cells established without crisis, resembled cells in the starting population, and retained the aggregative properties of early
passage dermal papilla cells. Growth studies were
performed on the immortalized cell lines, which showed
that transferring the cells to the restrictive temperature for
the large T gene product resulted in cell senescence or quiescence, and changes in morphology. Implantation of cell
pellets into the ears of immunologically compatible rats
showed that the immortal cells retained hair-inductive
ability.
Cytokines are believed to have an important role in the
control of hair growth. The pattern of cytokine gene
expression in the immortal cell lines was compared with
early passage dermal papilla cells and a non-hair-inducing
dermal papilla cell line, using reverse transcriptase-polymerase chain reaction. Epidermal growth factor, tumour
necrosis factor, and interleukin-1a were detected in the
immortalized and non-hair-inducing dermal papilla cell
lines, but were absent in passage 2 dermal papilla cells. All
other cytokines examined were detected in all the cell types
under study.
These results demonstrate that the polyomavirus large
Ttsa-immortalized dermal papilla cell lines are very similar
to passage 2 dermal papilla cells and thus provide a good
model for hair growth studies. Cytokine expression profiles
indicate that the expression of several cytokines may be
implicated in hair induction. Further studies are under way
to investigate the relationship between cytokine expression
and the hair growth cycle.
INTRODUCTION
epidermal interactions continue to occur in adult skin, by
experiments in which isolated rat vibrissa dermal papillae, or
early passage dermal papilla cells, combined with afollicular
or ear epidermis are implanted into rat ear pinnae and stimulate
new hair follicle formation (Oliver, 1970; Jahoda et al., 1984).
Dermal papilla cells must therefore be sending out specific hair
induction signals that can cause controlled epithelial cell differentiation events involved in de novo hair follicle formation.
Cytokines have been described to have several important
biological effects on both epithelial and mesenchymal cells
(Matsue et al., 1992; Peters et al., 1992), and as substantial proliferative activity is necessary for producing a mature hair
follicle, it is likely that they play an important regulatory role
in this process, and that one or more are among the signals
released by the dermal papilla to the proximal epithelium.
The ability of early passage dermal papilla cells to induce
Hair growth is a highly regulated process, involving rapid cell
division of the epidermal hair bulb matrix cells and their
resultant differentiation and keratinization. The mammalian
process is cyclical, consisting of three stages: anagen (growing
phase), catagen (regressive phase) and telogen (resting phase).
The cellular signals that cause the transition between one state
and the next are unclear, but it is thought that they resemble
signalling events that occur in embryonic development of the
hair follicle, where it has been suggested that epithelial cells
induce mesenchymal condensations, which form the dermal
papilla, and these in turn send messages to the epidermis,
causing it to invaginate into the dermis, associate with the
rudimentary dermal papilla, and form hair follicles (for review
see Hardy, 1992). It has been shown that these dermal-
Key words: immortalization, dermal papilla cell, hair induction
1762 W. Filsell and others
hair follicle formation upon implantation into rat ear pinnae is
lost with continued passaging (Jahoda et al., 1984), suggesting
dedifferentiation of the cells as they enter crisis. Stable dermal
papilla cell lines that retained a primary cell phenotype would
therefore be useful in determining the role of the dermal papilla
in hair growth.
We have constructed a series of immortal rat vibrissa dermal
papilla cell lines by transfecting them with a polyomavirus
large T gene encoding a temperature-sensitive protein. Here we
document the characteristics of these cell lines, comparing in
vitro and hair-inductive properties with early passage dermal
papilla cells and a post-crisis dermal papilla cell line.
MATERIALS AND METHODS
Cell culture
PVG rats were used as tissue donors throughout this study, and were
obtained from Harlan-Olac. Dermal papillae, dermal papilla cells and
immortal cell lines were maintained in Dulbecco’s modification of
Eagle’s medium (DMEM) supplemented with 15% foetal calf serum
(FCS), 2 mM L-glutamine and 1% non-essential amino acids (Flow
Laboratories).
Dermal papillae were isolated from adult male PVG rat vibrissa
follicles using microdissection as described by Oliver (1967), put into
6-well (3.5 cm) tissue culture plates (3 dermal papillae per well) and
allowed to explant for 12 to 15 days before passaging. A spontaneous
dermal papilla cell line, PVGDP2, was generated by passaging dermal
papilla cells through crisis (passage 8), as has also been described by
other workers (Reynolds et al., 1991).
Recombinant retrovirus generation and transfections
A gene encoding a temperature-sensitive polyomavirus large T
antigen, plttsa, was excised from pLTtsa (Rassoulzadegan et al., 1983)
by BglI, HincII restriction. The 2491 base pair (bp) fragment was
inserted into the retroviral vector pZIPNeoSV(X)1 (Cepko et al.,
1984) at the BamHI site to produce the retroviral vector pZITtsa.
The vectors pZIPNeoSV(X)1 and pZITtsa were transfected into ψ2
cells using calcium phosphate coprecipitation, to generate recombinant retrovirus-producing cell lines as a source of rvNeo, and rvLTtsa
retroviruses, as described by Mann et al. (1983). Viral titres were
routinely between 104 and 105 colony forming units ml−1.
Primary cultures of dermal papilla cells were trypsinized, seeded in
24-well tissue culture plates and incubated at 37°C in an atmosphere
of 5% CO2 in air until subconfluent cultures were obtained. The cells
were infected with 1 ml of virus stock containing 8 µg ml−1 polybrene
per well and incubated for 2.5 hours, followed by a further overnight
Table 1. Characterization of RT-PCR primers and products
Location of
primers in
cDNA
Species
(Reference)
Sequence (5′→3′)
TGF-α
(Lee et al., 1985)
(+) ATGGTCCCCGCGGCCGGACAGCTC
(−) TCTCAGAGTGGCAGCAGGCAGT
TGF-β
(Wen Qian et al., 1990)
(+) GCCCTGGATACCAACTACTGCTTC
1246-1269
(−) TCAGCTGCACTTGCAGGAGCGCACGATCAT 1553-1585
145-158
592-613
Location in
exon
Predicted
fragment
size (bp)
Restriction
site
Predicted
restriction
sizes (bp)
1
469
BglII
217, 252
−
−
340
HaeIII
145, 195
IGF I
(+) ATCCCCAAAATCAGCAGTCTTCCA
(Shimatsu and Rotwein, 1987) (−) TTCTGAGTCTTGGGCATGTCACTG
73-96
450-473
1
3
401
AvaII
55, 145,
201
IGF II
(Ueno et al., 1988)
(+) CAGTGGGGAAGTCGATGTTGGT
(−) GGAAGTACGGCCTGAGAGGTAG
387-408
660-681
2
3
295
AluI
83, 98, 114
bFGF
(Kurokawa et al., 1988)
(+) CAGCATCACTTCGCTTCCCGCACT
(−) TCAGCTCTTAGCAGACATTGGAAG
265-288
695-718
−
−
454
HinfI
146, 308
NGF-β
(Whittemore et al., 1988)
(+) TCATCCACCCACCCAGTCTTCCAC
(−) CTGTTTGTCGTCTGTTGTCAACGC
653-676
917-940
−
−
288
HpaII
84, 204
EGF
(Dorrow and Simpson, 1988)
(+) ATGTATGTTGAATCCGTGGACCGC
(−) GTGTTCCTCTAGGACCACAAACCA
88-111
406-429
−
−
342
HaeIII
65, 116,
161
TNF
(Shirai et al., 1989)
(+) ATGAGCACGGAAAGCATGATCCGA
(−) AGTGTGAGGGTCTGGGCCATGGAA
807-830
1521-1544
1
2
239
AccI
93, 146
Il-1α
(Nishida et al., 1989)
(+) TGGCCAAAGTTCCTGACTTGTTTC
(−) GAGACTTTGAGAGTCACAGGAATA
63-86
607-630
−
−
568
HaeIII
5, 264, 299
Il-2
(McKnight et al., 1989)
(+) ATGTACAGCATGCAGCTCTCATCC
(−) TTACTGAGTCATTGTTGAGATCAT
15-38
459-482
−
−
468
−
−
Il-3
(Cohen et al. 1986)
(+) ATGGTTCTTGCCAGCTCTACCACC
(−) CCTCAGATTGGCTTTGTCGTCACT
1376-1399
1652-1675
1
2
300
AvaII
82, 218
Il-4
GenBank accession no.
X16058
(+) TCTCACGTCACTGACTGTAGAGAG
(−) CTTTCAGTGTTGTGAGCGTGGACT
1-24
383-406
−
−
406
−
−
Il-6
(Northemann et al., 1989)
(+) TCTCTCCGCAAGAGACTTCCAGCC
(−) GAGTCTTTTATCTCTTGTTTGAAG
73-96
520-543
1/2
4/5
471
BglII
174, 297
IFN-α1
(Dijkema et al., 1986)
(+) ATGGCTCGGCTCTGTGCTTTC
(−) TCAGGACTCATTTCTTTCTTC
130-150
−
−
579
−
−
IFN-γ
(Dijkema et al., 1986)
(+) ATGAGTGCTACACGCCG
(−) TCGAACTTGGCGATGATGCTCATGAAT
8-31
344-367
1
3
368
Immortalized dermal papilla cells 1763
incubation in virus stock diluted 1:1 with medium. Virus-containing
medium was then replaced with fresh medium and the cells grown
until nearly confluent before selection of transfectants with 0.2 mg
ml−1 G418. Cell lines were designated DPLTtsa with a number to distinguish them.
DPLTtsa1, DPLTtsa2 and DPLTtsa3 cell lines were transfected and
established at the semi-permissive temperature (37°C); DPLTtsa4 and 5
were transfected at 37°C, but selected and established at the permissive
temperature, 33°C; DPLTtsa6 and 7 were transfected and established at
33°C. DPLTtsa1, 3, 5 and 7 were prepared as clonal cell lines, DPLTtsa2,
4 and 6 as polyclonal cell lines. Dermal papilla cells were transfected,
selected and expanded clonally or polyclonally at the temperatures
described to determine the best regimen for generating conditionally
immortalized cell lines retaining a primary cell phenotype. Transfections were not performed at the restrictive temperature, 39°C, as at this
temperature the large T antigen should be absent, and therefore cell
immortalization would not be expected to occur.
Genomic integration of plttsa was examined using Southern blotting
(Southern, 1975). Genomic DNA was extracted from cells (passage
21) and restricted with AluI, BamHI, or with BamHI and EcoRI. The
DNA was electrophoresed through a 0.8% agarose gel and transferred
to Zeta Probe membrane (Bio-Rad). Membranes were probed with the
2491 plttsa gene fragment labelled with [α-32P]dCTP (3000 Ci mmol−1;
Amersham) using a Multiprime DNA labelling kit (Amersham): probe
specific activity was >108 dpm µg−1 DNA.
Morphology and immunohistochemistry
Cells were stained with Coomassie Blue using the method of Pena
(1980), and examined using a Leitz Labovert microscope.
Serial sections were processed for immunogold staining by deparaffinizing through xylene and descending grades of ethanol to water.
Sections were neutralized with 0.1 M ammonium chloride, blocked
with 10% goat serum in phosphate-buffered saline (PBS), and
incubated at room temperature overnight with a monoclonal antibody
to polyomavirus large T antigen, α-PyLT1 (Dilworth and Griffin,
1982), diluted 1:200 in PBS containing 1% bovine serum albumin. A
1 hour incubation with goat anti-rat immunogold conjugate (5 nm;
Biocell Research Laboratories) diluted 1:200 in PBS containing 15%
goat serum was followed by extensive washing in deionized, distilled
water. Silver enhancement was performed using a Biocell LM silver
enhancement kit. Control sections were prepared without either
antisera, or without α-PyLT1, plus or minus silver enhancement.
Sections were counterstained with haemotoxylin and eosin, and
observed using a Leitz Ortholux II microscope and a Bio-Rad MRC
500 confocal scanning laser microscope.
Temperature-shift studies
The effect of the large T protein on the growth rate and morphology of
DPLTtsa cells (passage 50) was examined by temperature switch timecourse experiments similar to those performed by Rassoulzadegan et al.
(1983). Passage 2 dermal papilla cells were grown under the same conditions for comparison. Cells were seeded into flasks and grown at the
permissive temperature, then transferred to the restrictive temperature
at days 1, 5, and 9 post-seeding. Cell counts were performed every 2
to 3 days, and cell morphology was noted. Cells were also stained with
Coomassie Blue for more detailed morphology studies.
Hair induction studies
Cells for implantation were seeded into 10 cm culture dishes and
allowed to grow to confluence. DPLTtsa-immortalized cells (passages
50 and 65) were grown at permissive, semi-permissive and restrictive
temperatures. Cells were drained of medium, and scraped up with a
rubber policeman to give clumps of cells with their associated extracellular matrix. The cell clumps were picked up with fine-tipped
watchmaker’s forceps and implanted into small pockets in rat ear
pinnae as described by Jahoda (1992). After 6 weeks the animals were
killed, and ears examined for abnormal hair growth using a WildHeerbrugge dissecting microscope.
For histological analysis, tissue was fixed for at least 24 hours in
phosphate-buffered formalin before dehydration through increasing
grades of ethanol and embedding in paraffin wax. Serial sections (68 µm) were cut, mounted on slides precoated with 0.01% poly-Llysine and stained with haematoxylin and eosin, or used for immunogold procedures.
Reverse transcriptase polymerase chain reaction (RTPCR)
Total RNA was isolated from cells (passage 50) by the single-step guanidinium thiocyanate method of Chomczynski and Saachi (1987), and
stored at −70°C as an ethanol precipitate. cDNA was synthesized from
1 µg total RNA and amplified using a GeneAmp RNA PCR kit (Perkin
Elmer). Samples of 15 pmoles each of the upstream and downstream
primers (Table 1) were used. Forty amplifications were performed, each
cycle comprising 1 minute at 95°C and 1 minute at 60°C. The final step
was 60°C for 7 minutes. The products were electrophoresed through a
1.5% agarose gel, and compared with 123 bp ladder markers (Gibco
BRL). Each product was digested with restriction endonucleases to
confirm its identity. Each experiment was performed 3 times.
If a PCR product of predicted size was not observed, 10 µl of the
original PCR mix was reamplified as before using 50 pmoles of each
primer; 35 cycles were performed, comprising 1 minute at 92°C, 1
minute at 50°C, 1.5 minutes at 72°C with the final step being 8.5
minutes at 72°C. The products were then examined as before.
The oligonucleotide primers and the size of their expected amplification product are listed in Table 1. In order to distinguish between
amplification of contaminating genomic DNA and cDNA, the primer
positions for each cytokine were selected so as to enclose at least one
intron on the genomic sequence as shown in Table 1. When the
Fig. 1. Southern hybridization analysis of transfected dermal papilla
cells. Lanes a-f were probed with 32P-radiolabelled LTtsa sequences.
Lanes a, c and e are BamHI digests of genomic DNA (10 µg) from
the cell lines DPLTtsa1, DPLTtsa2 and an untransfected cell line,
respectively, and show that the 2.5 kb LTtsa gene is present in the
DPLTtsa cell lines but not in the untransfected cell line. Lanes b, d
and f are BamHI, EcoRI digests of DNA from the same cell lines and
b and d confirm the detection of the LTtsa gene, showing the expected
1.1 kb and 1.4 kb fragments of the gene yielded by BamHI, EcoRI
restriction; lane f shows that no LTtsa sequences were detected in
DNA from the untransfected cells. Size markers (kb) are HindIII,
EcoRI restricted lambda DNA fragments.
1764 W. Filsell and others
genomic structure of a chosen cytokine was unknown, the RT-PCR
primers were designed from the cDNA sequence and selected to be
at least 300-400 bases apart because exons of a larger size than this
are relatively rare in vertebrates (Hawkins, 1988). This strategy
favoured selected primers spanning an intron.
RESULTS
Preparation of LTtsa-immortalized dermal papilla cells
Dermal papilla cells were passaged once and rapidly growing
subconfluent cell monolayers were infected with the recombinant retrovirus, rvLTtsa. G418 was used to select transfected
cells, and these were expanded into 7 clonal and polyclonal
cell lines. No crisis period was observed, and a fibroblastic
morphology was maintained. Infection and expansion of cells
took 90 to 100 days. In control experiments we confirmed that
uninfected dermal papilla cells and those transfected with
rvNeo could not be clonally expanded into cell lines. With the
exception of the Southern blot analysis, which was performed
on cells at passage 21, all experimental data were generated
from PVGDP2 or DPLTtsa cell lines that had been passaged at
least 50 times post-transfection. All 7 DPLTtsa cell lines have
been cultured at 37°C for over 120 passages and still retain a
primary dermal papilla cell phenotype.
Southern blot analysis was used to show the presence of the
LTtsa gene within the transfected cells. Large T sequences were
found in genomic DNA from rvLTtsa-infected cells, but not in
that from control cells. Fig. 1 shows that restriction with the
endonuclease BamHI released the entire inserted sequence for
detection (lanes a and c). Confirmation of the identity of the
fragment was provided by additional restriction with EcoRI,
which cuts asymmetrically within the LT gene to yield 1.1 kb
and 1.4 kb fragments (lanes b and d).
Expression of the immortalizing gene was shown using
immunofluorescence staining. Using the monoclonal antibody,
α-PyLT1, the large T antigen was detected in the nuclei of
rvLTtsa-immortalized cells, but not the nucleoli or cytoplasm.
The thermolability of the large T protein could be demonstrated by immuno-staining of cells grown at the permissive,
semi-permissive and restrictive temperatures: at 33°C and
37°C cell nuclei were positive for large T protein, whereas cells
cultured at 39°C for 48 hours stained weakly, and upon further
culturing (5 days) all cells became negative for the large T
protein. No staining was seen with passage 2 dermal papilla
cells (Fig. 2).
Morphology of DPLTtsa cells
Primary and early passage dermal papilla cells have a characteristic pattern of aggregation in culture, quite unlike other
Fig. 2. Immunofluorescence detection of large Ttsa antigen. Using an antibody against polyomavirus large T protein, expression was
demonstrated to be localized to the nuclei, but not the nucleoli of DPLTtsa6 cells grown at the permissive temperature, 33°C, (A) and the semipermisive temperature, 37°C (B); cells grown at the restrictive temperature, 39°C, for 5 days were negative for the antigen (C), as were passage
2 dermal papilla cells (D). Bar, 10 µm.
Immortalized dermal papilla cells 1765
Fig. 3. Aggregation in culture. (A) Early passage dermal papilla cells
aggregate (arrowhead) in a characteristic manner as confluence is
approached; (B) DPLTtsa6 cells aggregate (arrowhead), form ridges
and clear the substratum in a similar manner; (C) the spontaneously
transformed dermal papilla cell line, PVGDP2 show minimal
aggregates. Coomassie Blue. Bar, 100 µm.
fibroblasts, which is lost upon continued passaging. We have
compared the aggregative behaviour of the immortalized
dermal papilla cell lines with passage 2 dermal papilla cells
and the post-crisis dermal papilla cell line, PVGDP2.
In culture, passage 2 dermal papilla cells behaved as
described by Jahoda et al. (1984): nearly confluent cells grew
in multilayers, formed aggregates and withdrew from the
substrate so that generally confluence was not reached; in
contrast, PVGDP2 cells reached confluence without any significant aggregation, did not clear the substratum and ultimately detached from the substratum as a confluent sheet (Fig.
Fig. 4. Aggregation (arrowheads) of passage 2 dermal papilla cells at
(A) 33°C, (B) 37°C, (C) 39°C. Bar, 100 µm.
3). The small cell conglomerates sometimes observed probably
arose from loss of contact inhibition by these spontaneously
transformed cells.
Subconfluent cultures of DPLTtsa cells demonstrated multilayering and formed aggregates that were stable for many days,
with ridging and clearing of the substratum occurring when the
cells reached confluence (Fig. 3). The cleared substratum was
then filled in by growing cells and by one week post-confluence the cells started to round up and detach.
Effect of temperature shift on DPLTtsa cells
To determine the effect of the immortalizing antigen on the
DPLTtsa cell lines, temperature-shift growth studies were
1766 W. Filsell and others
performed, and the growth rates and morphologies of the cells
observed. In parallel, passage 2 dermal papilla cultures were
transferred to the permissive and restrictive temperatures to
assess their behaviour.
When compared with passage 2 dermal papilla cells grown at
37°C, no alteration in growth rate or cell morphology (Fig. 4)
was noted with cells cultured at 39°C. At 33°C, a slight decrease
in growth rate was noted, but cell morphology was unaltered.
Fig. 5. Effect of temperature shift on growth and morphology of DPLTtsa cell lines. Growth of DPLTtsa2 (A), DPLTtsa6 (B), and DPLTtsa4 (C)
at 33°C (•) and after shifts to 39°C at day 1 (u), day 5 (n), or day 9 (s). Morphology of DPLTtsa2 (D,G), DPLTtsa6 (E,H), and DPLTtsa4 (F,I)
cells after incubation at 39°C for 1 day (D,E,F), and 10 days (G,H,I). Bar, 100 µm.
Immortalized dermal papilla cells 1767
Table 2. Observation of hair-induction events
Cell type implanted
Total number of implants
Evidence of hair induction
Dermal papilla cells (passage 2)
24
18
PVGDP2
24
0
DPLTtsa1
12 (33°C), 12 (37°C), 12 (39°C)
4 (33°C), 4 (37°C), 3 (39°C)
DPLTtsa2
12 (33°C), 12 (37°C), 12 (39°C)
0 (33°C), 0 (37°C), 0 (39°C)
DPLTtsa3
12 (33°C), 12 (37°C), 6 (39°C)
0 (33°C), 0 (37°C), 0 (39°C)
DPLTtsa4
12 (33°C), 12 (37°C), 6 (39°C)
2 (33°C), 4 (37°C), 0 (39°C)
DPLTtsa5
12 (33°C), 12 (37°C), 12 (39°C)
5 (33°C), 2 (37°C), 2 (39°C)
DPLTtsa6
12 (33°C), 20 (37°C), 12 (39°C)
6 (33°C), 14 (37°C), 3 (39°C)
DPLTtsa7
12 (33°C), 12 (37°C), 12 (39°C)
2 (33°C), 4 (37°C), 1 (39°C)
The growth patterns of the immortal dermal papilla cell lines
DPLTtsa1, 2, 5, 6 and 7 were similar in the temperature-switch
time course experiments: upon transferring the cells to the
restrictive temperature, growth rate slowed markedly, and the
cells seemed to enter a period of quiescence that lasted for upto
15 days before they were able to recommence growth at 39°C
(Fig. 5A,B). Morphologically, cell lines DPLTtsa1, 5, 6 and 7
retained a fibroblastic appearance throughout the growth study,
with a few cells within the population being large, flattened
and granular. DPLTtsa2 cells showed a mixture of shapes after
growth at the restrictive temperature: some cells remained
fibroblastic, some had a more flattened appearance and were
larger with predominant cytoskeletons, and other cells were
smaller and cuboidal.
The growth patterns of the dermal papilla cell lines
DPLTtsa3 and 4 resembled each other, in that after the switch
to the restrictive temperature cell numbers progressively
decreased as the cells appeared to enter senescence. But
although growth trends were similar between the two cell lines,
cell morphology changes were not. DPLTtsa3 cells were fibroblast-like at 39°C, but rapidly rounded and detached from the
culture vessel; DPLTtsa4 cells initially appeared as enlarged,
flattened fibroblasts with prominent cytoskeletons (characteristic of a senescent phenotype), but with time changed to small
cuboidal cells with long processes, which soon started to round
and detach (Fig. 5). When kept at 39°C for extended periods,
eventually these two cell lines died.
Hair induction by immortal dermal papilla cells
DPLTtsa cells were implanted into rat ear pinnae to determine
their hair-inductive abilities; control implantations of passage
2 dermal papilla cells and PVGDP2 were performed, as were
incision-only tests to show the wound response engendered by
the operational procedure alone (Table 2 gives details of operations performed). The use of inbred PVG rats as both a source
of primary dermal papilla cells and as recipients of implanted
cells derived from them prevented any rejection-mediated
immune response.
Implantation of passage 2 dermal papilla cells resulted in
induction of whisker fibre production in the ear, which could
be clearly seen at the macroscopic level: histological analysis
confirmed the presence of large induced hair follicles (Fig.
6A). As expected, PVGDP2 cells did not demonstrate any hairinductive ability (Fig. 6B), and implanted cells were not distinguishable from indigenous dermal fibroblasts, possibly
because of their inability to form large or stable aggregates.
Incision-only controls resembled PVGDP2 implant sites (data
not shown).
In tissue that had been implanted with the DPLTtsa cell lines,
although no large whisker hairs were visible, histological examination of the implantation sites revealed evidence of hair
induction (Table 2). The conditionally immortal dermal papilla
cell lines DPLTtsa1 and DPLTtsa6 were able to induce large hair
follicles (Fig. 6), similar to those that arose after passage 2
dermal papilla cell implants. Tumours were not observed in any
of the test sites implanted with DPLTtsa cells. This was as
expected because polyomavirus large T protein alone is not sufficient to cause cellular transformation, but requires a second
oncoprotein, such as the product of the polyomavirus middle T
gene, for this event to occur (Rassoulzadegan et al., 1982).
The DPLTtsa cell lines were implanted after culturing at the
permissive, semi-permissive and restrictive temperatures for
the large T protein. Cells grown at the restrictive temperature
did not reach confluence, looked unhealthy (as described
earlier), and had poor extracellular matrices, which reduced the
number of cells it was possible to implant. However, this did
not appear to affect their hair-inductive abilities in the in vivo
model: as can be seen in Fig. 6D,E and F, DPLTtsa1 cells grown
at all three temperatures are capable of inducing large hair
follicles.
Immunogold staining of the large T protein was performed
to establish the location of the DPLTtsa cells in relation to
induced hair follicles. The use of confocal microscopy allowed
visualization of the different planes of the sections. Comparison of interference reflection and phase-contrast micrographs
showed that the implanted DPLTtsa cells had formed dermal
papillae, from which new hair follicles had arisen. Fig. 7 shows
this for a DPLTtsa6 cell implant: the abnormally large follicle
identified by haematoxylin- and eosin-stained sections of the
implant site (7A,B,C), when stained with an antibody to the
large T antigen, is shown by immunogold labelling (7D,F;
immunostaining shown in red, phase-contrast shown in green)
to have a dermal papilla made up of the implanted cells.
Control sections of PVGDP2 cell implants did not show any
staining for the large T antigen, and standard immunogold
controls were also negative.
Cytokine profiles of DPLTtsa cell lines
Total RNA isolated from the immortal dermal papilla cell
lines, passage 2 dermal papilla cells and PVGDP2 appeared
1768 W. Filsell and others
to be intact as determined by the banding pattern after agarose
gel electrophoresis (data not shown). The amount of total
RNA isolated was within the range of 25-90 µg per 160 cm2
flask.
Gene expression was analysed using the technique of RTPCR. As shown in Table 3 and Fig. 8, passage 2 dermal papilla
cells, DPLTtsa cell lines and PVGDP2 cells had similar
cytokine expression profiles, except for epidermal growth
Fig. 6. Histological examination of implanted tissue for hair induction. Passage 2 dermal papilla cells (A); the non-hair-inducing dermal papilla
cell line, PVGDP2 (B); DPLTtsa6 cultured at 37°C prior to implantation (C); DPLTtsa1 cultured at 33°C (D), 37°C (E), and 39°C (F) prior to
implantation. Arrowheads indicate induced hair follicles. Bar, 200 µm.
Immortalized dermal papilla cells 1769
factor (EGF), tumour necrosis factor (TNF), and interleukin 1a
(IL1a), whose expression could not be detected in passage 2
dermal papilla cells even after further amplification.
DISCUSSION
In this report we have described the establishment of rat
vibrissa dermal papilla cell lines that retain specialized
functions, using a polyomavirus large T gene that encodes a
thermolabile protein. The use of a retroviral shuttle vector
system to transfer the gene allowed efficient transfection of
small starting populations of cells. Transfected cells were
expanded into cell lines, and their retention of specific dermal
papilla cell characteristics was examined. By employing a thermolabile immortalizing antigen we were able to study the
effects of this protein on the properties of the dermal papilla
cell lines.
Fig. 7. Localization of implanted DPLTtsa6 cells. (A), (B), (C) haematoxylin- and eosin-stained serial sections showing large induced hair
(arrowhead). Bar, 200 µm. (D) Interference reflection, (E) phase-contrast, and (F) superimposed (D) on (E) confocal micrographs of induced
dermal papilla seen in (A), after immunogold staining with αPyLT. Bars, 25 µm.
1770 W. Filsell and others
Table 3. Cytokine expression profiles of passage 2 dermal papilla (DP) cells, DPLTtsa and PVGDP2 cell lines as
determined by RT-PCR
Cell type:
Culture temp.
TGF-α
TGF-β
IGF I
IGF II
bFGF
NGF-β
EGF
TNF
Il-1α
Il-2
Il-3
Il-4
Il-6
IFN-α1
IFN-γ
DPLTtsa1
(oC)
DPLTtsa2
DPLTtsa3
DPLTtsa4
DPLTtsa5
DPLTtsa6
DPLTtsa7
p.2DP cells
PVGDP2
33
37
33
37
37
37
37
33
37
39
37
37
33/37/39
+
+
+
+
+
+
+
+R
+
−R
+
−R
+
−R
−R
+
+
+
+
+
+
+
+
+
−R
+
−R
+
−R
−R
+
+
+
+
+
+
+R
+R
+
−R
+
−R
+R
−R
−R
+
+
+
+
+
+
+R
+
+
−R
+
−R
+
−R
−R
+
+
+
+
+
+
+R
+
+
−R
+
−R
+R
−R
−R
+
+
+
+
+
+
+R
+R
−R
−R
+
−R
+R
−R
−R
+
+
+
+
+
+
+
+
+
−R
+
−R
+
+/−
−R
+
+
+
+
+
+
+R
+R
+R
−R
+
−R
+R
−R
−R
+
+
+
+
+
+
+
+
+
−R
+
−R
+
+/−
−R
+
+
+
+
+
+
+R
+
+
−R
+
−R
+R
−R
−R
+
+
+
+
+
+
+R
+R
+
−R
+
−R
+R
−R
−R
+
+
+
+
+
+
−R
−R
−R
−R
+
−R
+R
−R
−R
+
+
+
+
+
+
+
+R
+
−R
+
−R
+
−R
−R
+, PCR product identified; −, no PCR product; R, additional 35 amplification cycles; °C, temperature at which the DPLTtsa cell lines were cultured.
At the permissive and semi-permissive temperatures for the
large T antigen, DPLTtsa cells behaved as early passage dermal
papilla cells in culture, being fibroblastic in appearance, having
a comparable growth rate, and forming aggregates that were
stable for several days. The characteristic aggregation of early
passage cells in culture has also been described for sheep
(Withers et al., 1986) and human dermal papilla cultures
(Messenger et al., 1986), and is reminiscent of the condensation of mesenchymal cells during hair follicle development.
That they are able to do so in the absence of epithelial cells
suggests that the signal for foetal follicle formation may arise
from these cells.
Temperature-induced removal of the large T protein had
profound effects on both growth rate and morphology of the
conditionally immortal DPLTtsa cell lines. In contrast, there
appeared to be no temperature effect on passage 2 dermal
papilla cells, with retention of a fibroblastic appearance and an
unaltered growth rate. After the immortalizing element was
lost from DPLTtsa cells they either quiesced for several days
before recommencing growth, or appeared to senesce before
eventually dying. In every cell line changes in cell shape, size
and spreading were observed. Similar findings for a rat fibroblast cell line immortalized with this polyomavirus LT gene
have been reported previously (Rassoulzadegan et al., 1983).
This suggests that the changes observed in the immortal cells
were due to the influence of the large T antigen, and that its
presence is necessary for maintenance of the immortal
phenotype, as has also been demonstrated in other cell types
using mutant viral genes encoding thermolabile proteins (Jat
and Sharp, 1989; Wynford-Thomas et al., 1990). These
workers observed a cessation of growth upon switching conditionally immortalized cells to the restrictive temperature,
although the cells remained metabolically active for several
days; cell enlargement and increased spreading was also
reported, as we have found with LTtsa-immortalized dermal
papilla cell lines.
However, DPLTtsa1, 2, 5, 6 and 7 cell lines were able to
recover growth potential after several days at the restrictive
temperature. Thus upon reverting to the non-immortal
phenotype they were able to recommence normal growth in
culture at 39°C, just as untransfected dermal papilla cells are
able to (Fig. 4). DPLTtsa cell lines 3 and 4 were not able to
survive the loss of the immortalizing antigen: the original populations of dermal papilla cells from which these 2 cell lines
Fig. 8. RT-PCR amplification. PCR products obtained by RT-PCR
and agarose gel electrophoresis of 1 µg DPLTtsa1 (37°C) for TGF-α
(lane a), TGF-β (lane c), IGF I (lane e), IGF II (lane g), bFGF (lane
i), NGF-β (lane k), EGF (lane m), TNF (lane o), IL-1α (lane q), IL-2
(lane s), IL-3 (lane t), IL-4 (lane v), Il-6 (lane w), IFNα1 (lane y),
IFN-γ (lane z), indicated by arrows. Restricted products are indicated
in lanes b, d, f, h, j, l, n, p, r, u, and x. Size markers (M) are a 100 bp
ladder.
Immortalized dermal papilla cells 1771
were derived may not have contained those cell subpopulations
capable of recovering from reversion of immortalization.
Implantation of pellets of DPLTtsa cells into rat ear pinnae
have shown that 5 of the 7 cell lines cells retain at least partial
hair-induction ability. DPLTtsa2 and DPLTtsa3 cell lines have
not produced any obvious features of hair induction. Large
follicles reminiscent of vibrissa follicles were observed after
implantation of DPLTtsa1 and DPLTtsa6 cells. This is as
expected, as it has recently been reported that fibre type is
dependent on the site of origin of its dermal papilla (Jahoda,
1992), and the cell lines we have produced originate from
vibrissa dermal papillae. That macroscopic evidence of hair
formation was not seen after implantation with DPLTtsa cells
might be due to the orientation of induced follicles causing the
fibres to remain within the dermis, as is suggested by Fig. 6C
and D; or it may be that it takes longer for the large T-immortalized cells to condense into a functional dermal papilla than
passage 2 cells, resulting in less time available during the
experimental period to develop a fibre. The hair induction
capacity of the DPLTtsa cell lines 1 and 6 was not affected by
growing them at the restrictive temperature prior to implantation, suggesting that the presence of the immortalizing protein
is not necessary, in the short term at least, for the cells to retain
their hair-inductive properties. However, it is important to note
that although the cells were grown at the three different temperatures, for the 6 week implantation study they were at 37°C
or slightly lower (rat ear temperature), so, although it is
unlikely that the large T antigen is involved in the hair
induction process, this cannot be proved by these experiments.
The location of the DPLTtsa cells in sections showing
induced hair follicles was revealed using immunogold
detection of the large T antigen as a cell marker. It would
appear from our histological observations that the immortal
cells condense to form a new dermal papilla (Fig. 7) that is
capable of interacting with cells in the in vivo environment,
including epithelium brought into close contact with the
immortal cells by the implantation operation, to cause de novo
hair follicle synthesis. Jahoda and colleagues (1993) have
recently reported similar observations with early passage
dermal papilla cells. This interactive process suggests that the
immortal cells are behaving as in follicle neogenesis, by both
producing and responding to specific cellular signals.
In order to try to identify candidate signal molecules
involved in the hair induction process, cytokine expression
profiles of the DPLTtsa cell lines were compared with those of
early passage dermal papilla cells, and the non-hair-inducing
dermal papilla cell line, PVGDP2. EGF, TNF and IL1a were
expressed in the hair-inducing immortal and non-hair-inducing
cell lines, but not in passage 2 dermal papilla cells. It may be
that these cytokines are expressed in the DPLTtsa and PVGDP2
cell lines as a consequence of their establishment in culture,
where regulatory mechanisms predetermining cell apoptosis
must be overridden. It is also possible that the DNA synthesisstimulating and gene-transactivating ability of large T antigen
may be involved in the DPLTtsa cell lines.
The expression of all of the other cytokines examined was
similar between the cells, although in some cases it was
necessary to reamplify in order to detect the cytokine. From
this it was concluded that these molecules are expressed at
lower levels in some cell lines than others. This explanation is
more likely than the alternative of inconsistencies in the RT-
PCR reactions because the sole variant in each reaction was
the RNA sample, and not the primers or the reaction conditions. We are currently investigating the use of semi-quantitative RT-PCR to confirm these variations in level of cytokine
transcripts.
The observation that there is no obvious correlation between
cytokine expression and hair-inductive ability of the cell lines
investigated in this study could be interpreted as indicating that
cytokines are not important in the ability of the dermal papilla
to stimulate new hair follicle formation. However, this is
unlikely as there are many reports implicating cytokines in the
control of hair growth (du Cros, 1993; Guo et al., 1993;
Akhurst et al., 1988; Green and Couchman, 1984) and many
cytokines have been described to have important biological
effects on both epithelial and mesenchymal cells (Lyons et al.,
1990; Guo et al., 1993). Alternatively, cytokines may be
involved, but it is the level of expression of one or more factors
that is crucial in determining hair-inductive capacity of the
dermal papilla rather than simply the presence or absence of a
particular gene transcript.
The immortal dermal papilla cell lines described in this
report have retained the specialized phenotype of their primary
counterparts, including hair-inductive capacity, for over 120
passages to date. Further study of this readily available source
of cells should yield important information about the role of
the dermal papilla in the process of hair growth.
We thank Dr R. Mulligan for the retroviral vector pZIPNeoSV(X)1
and the ψ packaging line; Dr F. Cuzin for pLTtsa; Prof. B. Griffin for
the monoclonal antibody α-PyLT1. We are grateful to Miss A. Scarborough for synthesizing the oligonucleotide primers, and Mr D. Ferdinando for help with confocal microscopy. We are particularly
indebted to Dr A. Reynolds for her invaluable assistance with the ear
implantation experiments.
REFERENCES
Akhurst, R., Fee, F. and Balmain, A. (1988). Localized production of TGF-β
mRNA in tumour promoter-stimulated mouse epidermis. Nature 331, 363365.
Cepko, C. L., Roberts, B. E. and Mulligan, R. C. (1984). Construction and
applications of a highly transmissible murine retrovirus shuttle vector. Cell
37, 1053-1062.
Cohen, D., Hapel, A. and Young, I. (1986). Cloning and expression of the rat
interleukin-3 gene. Nucl. Acids Res. 14, 3641-3658.
Chomczynski, M. and Saachi, N. (1987). Method of RNA isolation by acid
guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem.
162, 156-159.
Dijkema, R. van der Meide, P., Dubbeld, M., Caspers, M., Wubben, J. and
Schellekens, H. (1986). Cloning, expression and purification of rat IFNgamma. Meth. Enzymol. 119, 453-463.
Dilworth, S. M. and Griffin, B. E. (1982). Monoclonal antibodies against
polyoma virus tumour antigen. Proc. Nat. Acad. Sci. USA 79, 1059-1063.
Dorrow, D. and Simpson, R. (1988). Cloning and sequence analysis of a
cDNA for rat epidermal growth factor. Nucl. Acids Res. 16, 9338.
du Cros, D. L. (1993). Fibroblast growth factor influences the development and
cycling of murine hair follicles. Dev. Biol. 156, 444-453.
Green, M. R. and Couchman, J. R. (1984). Distribution of EGF receptors in
rat tissues during embryonic skin development, hair formation and the adult
hair growth cycle. J. Invest. Dermatol. 83, 118-123.
Guo, L., Yu, Q.-C. and Fuchs, E. (1993). Targeting expression of keratinocyte
growth factor to keratinocytes elicits striking changes in epithelial
differentiation in transgenic mice. EMBO J. 12, 973-986.
Hardy, M. H. (1992). The secret life of the hair follicle. Trends Genet. 8, 55-61.
Hawkins, J. (1988). A survey on intron and exon lengths. Nucl. Acids Res. 16,
9893-9908.
1772 W. Filsell and others
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, 560562.
Jahoda, C. A. B. (1992). Induction of follicle formation and hair growth by
vibrissa dermal papillae implanted into rat ear wounds: vibrissa-type fibres
are specified. Development 115, 1103-1109.
Jahoda, C. A. B., Reynolds, A. J. and Oliver, R. F. (1993). Induction of hair
growth in ear wounds by cultured dermal papilla cells. J. Invest. Dermatol.
101, 584-590.
Jat, P. S. and Sharp, P. (1989). Cell lines established by a temperaturesensitive simian virus 40 large-T-antigen are growth restricted at the
nonpermissive temperature. Mol. Cell. Biol. 9, 1672-1681.
Kurokawa, T., Seno, M. and Igarashi, K. (1988). Nucleotide sequence of rat
basic fibroblast growth factor cDNA. Nucl. Acids Res. 16, 5201.
Lee, D., Rose, T., Webb, N. and Todaro, G. (1985). Cloning and sequence
analysis of a cDNA for rat transforming growth factor-alpha. Nature 313,
489-491.
Lyons, K. M. and Hogan, B. L. M. (1990). TGF-β-like genes in mammalian
development. In Genetics of Pattern Formation and Growth Control (ed. A.
P. Mahowald), pp. 137-156.
Mann, R., Mulligan, R. C. and Baltimore, D. (1983). Construction of a
retrovirus packaging mutant and its use to produce helper-free defective
retrovirus. Cell 33, 153-159.
Matsue, H., Cruz, P. D., Bergstresser, P. R. and Takashima, A. (1992).
Cytokine expression by epidermal cell subpopulations. J. Invest. Dermatol.
99, 42s-45s.
McKnight, A., Mason, D. and Barclay, A. (1989). Sequence of rat interleukin
2 and anomalous binding of a mouse interleukin 2 cDNA probe to rat MHC
classII-associated invariant chain mRNA. Immunogenetics 30, 145-147.
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. Dermatol. 114, 425-430.
Nishida, T., Nishino, N., Takano, M., Sekiguchi, Y., Kawai, K., Mizuno, K.,
Nakai, S., Masui, Y. and Hirai, Y. (1989). Molecular cloning and
expression of rat interleukin-1a cDNA. J. Biochem. 105, 351-357.
Northemann, W., Braciak, T., Hatton, M., Lee, F. and Fey, G. (1989).
Structure of the rat interleukin 6 gene and its expression in macrophagederived cells. J. Biol. Chem. 264, 16072-16082.
Oliver, R. F. (1967). The experimental induction of whisker hair 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.
Pena, S. D. J. (1980). A new technique for the visualization of the cytoskeleton
in cultured fibroblasts with Coomassie blue R250. Cell Biol. Int. Rep. 4, 149153.
Peters, K. G., Werner, S., Chen, G. and Williams, L.T. (1992). Two FGF
receptor genes are differentially expressed in epithelial and mesenchymal
tissues during limb formation and organogenesis in the mouse. Development
114, 233-243.
Rassoulzadegan, M., Cowie, A., Carr, A., Glaichenhaus, N. and Kamen, R.
(1982). The role of individual polyoma virus early proteins in oncogenic
transformation. Nature 300, 713-718.
Rassoulzadegan, M., Naghasfar, Z., Cowie, A., Carr, A., Grisoni, M.,
Kamen, R. and Cuzin, F. (1983). Expression of the large T protein of
polyoma virus promotes the establishment in culture of ‘normal’ rodent
fibroblast cell lines. Proc. Nat. Acad. Sci. USA 80, 4354-4358.
Reynolds, A., Oliver, R. F. and Jahoda, C. A. B. (1991). Dermal cell
populations show variable competence in epidermal cell support: stimulatory
effects of hair papilla cells. J. Cell Sci. 98, 75-83.
Shimatsu, A. and Rotwein, P. (1987). Sequence of two rat insulin-like growth
factor I mRNAs differing within the 5′ untranslated region. Nucl. Acids Res.
15, 7196.
Shirai, T., Shimizu, N., Horiguchi, S. and Ito, H. (1989). Cloning and
expression in Escherichia coli of the gene for rat tumour necrosis factor.
Agric. Biol. Chem. 53, 1733-1736.
Southern, E. M. (1975). Detection of specific sequences among DNA
fragments separated by gel electrophoresis. J. Mol. Biol. 98, 503-517.
Ueno, T., Takahashi, K., Matsurguchi, T., Endo, H. and Yamamoto, M.
(1988). Transcriptional deviation of the rat insulin-like growth factor II gene
initiated at three alternative leader exons between neonatal tissues and ascites
hepatomas. Biochim. Biophys. Acta 950, 410-411.
Wen Qian, S., Kondaiah, P., Roberts, A. and Sporn, M. (1990). cDNA
cloning by PCR of rat transforming growth factor β-1. Nucl. Acids Res. 18,
3059.
Whittemore, S., Friedman, P., Larhammar, D., Persson, H., GonzalezCarvajal and Holets, V. (1988). Rat β-nerve growth factor sequence and
site of synthesis in the adult hippocampus. J. Neurosci. Res. 20, 403-410.
Withers, A. P., Jahoda, C. A. B., Ryder, M. L. and Oliver, R. F. (1986).
Culture of wool follicle dermal papilla cells from two breeds of sheep. Arch.
Dermatol. Res. 279, 140-142.
Wynford-Thomas, D., Bond, J. A., Wyllie, F. S., Burns, J. S., Williams, E.
D., Jones, T., Sheer, D. and Lemoine, N.R. (1990). Conditional
immortalization of human thyroid epithelial cells: a tool for analysis of
oncogene action. Mol. Cell. Biol. 10, 5365-5377.
(Received 29 October 1993 - Accepted, in revised form,
17 March 1994)