Download Abnormalities of epidermal differentiation associated with

Journal of General Virology(1992), 73, 1251-1260. Printedin GreatBritain
1251
Abnormalities of epidermal differentiation associated with expression of
the human papillomavirus type 1 early region in transgenic mice
J. M. Tinsley, 1 C. Fisher 2 and P. F. Searle 1.
1Cancer Research Campaign Laboratories, Department of Cancer Studies, University of Birmingham Medical School,
Birmingham B15 2TJ, U.K. and 2The Upjohn Company, Kalamazoo, Michigan 49001, U.S.A.
The promoter region of a keratin 6 (K6) gene was used
to regulate expression of the early region of human
papillomavirus type 1 (HPV-1 e) in transgenic mice. In
one line of mice the K 6 - H P V l e transgene was
transcribed in several regions of the skin, the predominant transcript being a 1.1 kb RNA including the E4
open reading frame, and E l - E 4 protein was detected in
the upper suprabasal layers of the skin in paws and tail.
A 1-7 kb RNA corresponding to the E6/E7 transcript
was also prominent in tails of homozygous transgenic
animals. In young homozygous transgenic mice the
epidermis of the tail showed dysplasia and hyperplasia
of the suprabasal layers with both hyperkeratosis and
focal parakeratosis in the stratum corneum. A similar
though milder phenotype was also observed spor-
adically in hcmizygous transgenics. Analysis of the
pattern of mouse keratins present in the affected tail
skin showed strong up-regulation of the endogenous
keratins 6 and 16 throughout the basal and suprabasal
layers, suggesting a positive feedback mechanism for
the strong transgene activation. Expression of the
major differentiation-specific keratins 1 and 10 was
repressed. The pattern of E l - E 4 expression and the
perturbation of normal epithelial differentiation parallel many of the characteristics of HPV-1 warts or
verrucae, suggesting that H P V transgenic mice could
be useful for analysis of the interactions of H P V gene
products with cellular regulatory pathways within an
otherwise normal epithelium.
Introduction
produces the striking change in epithelial morphology is
unknown, although it seems likely that expression of the
viral early region genes within the lower levels of the wart
is involved. Viral D N A replication is detectable only in
the suprabasal layers; however it appears likely that the
viral episome is present within cells of the basal layer,
where early gene products might act to interfere with the
normal control of cell proliferation and differentiation
(Pfister, 1984; Howley et al., 1986).
Common warts on other regions of the cutaneous
epithelium are generally caused by HPV-2 or -4. Neither
the warts induced by these viruses nor those caused by
HPV-1 have been associated with malignant progression. Other types specific for the skin, such as HPV-5
and -8, are not commonly detected in the general
population but are found particularly in individuals with
the rare inherited disease epidermodysplasia verruciformis; in such individuals warts associated with these
viruses may progress to squamous cell carcinomas,
particularly at sites exposed to the sun (Syrj/inen et al.,
1987).
Other papillomaviruses are found specifically in
lesions of mucosal epithelia. Much interest has focused
on the frequent detection of HPV-16 or -18 D N A
The papillomaviruses are a widespread group of viruses
having dsDNA genomes of around 8 kb, including six to
eight 'early' and two 'late' open reading frames (ORFs)
all in the same orientation. At least 60 distinct types of
human papillomaviruses (HPVs) have been identified
(de Villiers, 1989). They infect the various cutaneous and
mucosal epithelia of the body where they cause benign
tumours with a broad spectrum of morphologies from
almost fiat dysplasias to large, hyperplastic papillomatous lesions (Broker & Botchan, 1986; Pfister, 1984;
Syrj/inen et al., 1987).
HPV-1 is associated with deep warts on the thickened
palmar and plantar skin. Such warts or verrucae are
typically very hyperplastic, with the proliferative basal
cell layer contorted into elongated papillae, thickening of
the suprabasal layers (acanthosis), marked hyperkeratosis and parakeratosis (Gross et al., 1982). HPV-1 is the
most productive of the human papillomaviruses, and the
virus particles form characteristic inclusion bodies in the
cell nuclei in the upper layers of the wart; cytoplasmic
granules containing the abundant E l-E4 protein are also
present in the upper layers. The mechanism that
0001-0584 © 1992SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
1252
J. M . Tinsley, C. Fisher and P. F. Searle
sequences in higher grade cervical lesions and invasive
carcinoma, whereas types 6 and 11, although frequently
present in condylomata acuminata and low grade lesions
of the cervix (CIN I), do not appear to be associated with
malignant progression (zur Hausen & Schneider, 1987;
Syrj/inen et al., 1987; Vousden, 1989). Cellular transformation functions have been assigned to the viral E6
and E7 gene products (Miinger et al., 1989; Halbert et al.,
1991), and malignant progression of the initially benign
lesions associated with types 16 and 18 appears to be
associated with chromosomal integration of part of the
HPV genome in a manner which preserves the function
of the E6 and E7 genes but removes some of the normal
transcriptional control over their expression (Schwarz et
al., 1985). The transforming activity of these proteins is
believed to be related to their ability to interact
respectively with the cellular p53 and retinoblastoma
gene proteins (Dyson et al., 1989; Werness et al., 1990;
Scheffner et at., 1990). Although the mechanisms are
currently unknown, both p53 and the retinoblastoma
protein are believed to function in the regulation of the
cell cycle. The E7 proteins of HPV-6 and -11 bind more
weakly to the retinoblastoma gene protein, but the E6
proteins of these viruses could not be shown t o interact
with p53. Comparable studies of the E6 and E7 proteins
of the cutaneous HPVs have not been reported.
A full understanding of the molecular function of
papillomaviruses in both benign and malignant lesions
must include the effects of their gene products within the
context of the normally highly ordered pathway of
epithelial cell differentiation, and for this reason several
groups have produced transgenic mice carrying chromosomally integrated papiUomavirus sequences. In the first
such study, the entire genome of a bovine papillomavirus
(BPV-1, which induces fibropapillomas in cattle) was
introduced into the mouse genome (Lacey et al., 1986).
Transgenic mice in this line initially appeared normal,
but later developed patches of abnormal skin and
protuberant tumours at sites of epithelial wounding.
These appeared to involve primarily the fibroblasts of
the dermis, and extrachromosomal copies of the viral
D N A were found in the lesions. Characteristic karyotypic abnormalities were found to be associated with the
locally invasive fibrosarcomas (Lindgren et al., 1989). In
another study, the E6 and E7 genes of HPV-16 were
introduced into transgenic mice under the control of the
mouse mammary tumour virus promoter region. Male
mice in several lines carrying this construct were found to
develop seminomas (Kondoh et al., 1991).
We were interested in comparing the function of
'oncogenic' and 'non-oncogenic' HPVs in epithelial
tissues of transgenic mice. Initially we produced three
lines of transgenic mice carrying the entire genome of
H P V - I ; however none of these produced detectable
ET
° 1 ~ E7
Sail
Keratin 6 promoter
E2
a
RV Smal
Xh
{79 bp)
I
Full-length primary transcript
3
4
kb
AAARA 4"0 kb
E6, E7, Eli^E4, E5a
Eli^E4, E5a
2
~
llflflRfl
1'7 kb
aRRAn
I'1 kb
Fig. 1. The bK6-HPVle transgene. The stippled box represents 2.2
kbp of 5' flankingDNA from the bK6 gene. The thick black linejoined
to this represents HPV-1 DNA, from 75 bp on the viral genome; the
scale marks (in kb) also relate to the viral genome. Selectedrestriction
enzyme sites are indicated. Open boxes in the upper part of the figure
indicate the positionsof the viral ORFs. The extent of the expectedfulllength primary transcript from the keratin promoterto the early region
poly(A)site of the virus is shownin the lowerpart of the figure,together
with spliced RNAs referred to in the text.
HPV transcripts, perhaps due to the observed methylation of the viral D N A (unpublished results). We
therefore adapted our approach, and have produced
transgenic mice carrying fusion constructs in which the
early region transcription units of HPV-1 and -16 were
placed under the transcriptional control of epidermisspecific promoters from keratin genes. In this paper we
describe a line of transgenic mice in which high level
expression of the HPV-1 early region in the tail is
associated with epidermal hyperplasia, hyperkeratosis
and parakeratosis, strong up-regulation of the hyperproliferation-associated keratins and down-regulation of
keratins associated with normal epithelial differentiation. Similar changes are also characteristic of HPV-1induced warts. Hence H P V transgenic mice may be
useful for studying the disruption of the normal control of
epidermal cell proliferation and differentiation induced
by human papillomaviruses.
Methods
Plasmid construction. Plasmid pbK6-CAT (kindly provided by Dr J.
L. Jorcano and Dr M. Blessing) contains approximately2.2 kb of 5'
sequence from the bovine keratin 6 (bK6) gene (formerlydesignated
keratin IV), joined to the chloramphenicol acetyltransferase(CAT)
gene, and inserted between the BamHI and Sinai sites of pUC18
(Blessing et al., 1989).The pUC-bK6 vector fragment was purified by
electrophoresis after excising the CAT gene with XhoI and Sinai.
Plasmid pPXL-HPV1 (kindlyprovided by Dr T. Broker)contains an
almost complete HPV-1 genome, except that the early promoter had
been replacedwith an XhoIlinker at bp 75 of the viral genome;this had
been clonedinto a simian virus40 late promoter construct(Chowet al.,
1987). The promoterless HPV-t early region from this XhoI site to
EcoRV [at 4527 bp, downstreamof the early region poly(A)signal]was
cloned between the XhoI and EcoRV sites of pBluescript (Stratagene),
then recloned using the XhoI and Sinai (polylinker) sites into the
pUC-bK6 vector fragment alreadydescribed, to producethe construct
pbK6-HPVle.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
Epidermal abnormalities in HPV-1 transgenics
(a)
(b)
(c)
I
Ta
M
E
S
B
kb
: ....
..:.....
• .....
•
..
.......
•
•
.....
~/
(d)
+/- I+/+
To
1253
P
S
E
Ta
kb
,~_
, : :i',
:
- 5-0
: - 4-0
- 5.0
5-0
5.0
- 4.0
4.0
4.0
- 3.0
3.0
- Z-03'0
- 2.0
- 1.8
1-8
~,::
kb
•
,
-
,/:-1"3
2.0
1.8
1.3
3-0
2.0
1.8
1.3
1.3
- 0.8
0'8
,0.8
0.8
Fig. 2. Northern blots of RNA from bK6-HPVle transgenic mice. (a) RNA samples from tails from animals representative of the five
independent lines are as indicated. (b) RNA samples from tissues of a mouse in line 27-7; P, paws; I, intestine; Ta, tail; M, skeletal
muscle; E, ears; S, dorsal skin; B, brain; To, tongue. (c) Comparison of RNAs from tails of hemizygous (+ / - ) and homozygous ( + / +)
transgenic mice in line 27-7. (d) RNAs from tissues of bomozygous 27-7 transgenic mouse; P, paws; S, dorsal skin; E, ears; Ta, tail. The
main panels show the results using an HPV-1 specific probe; filters were reprobed with a p-actin probe to control for variation in RNA
loading, as shown in the lower panels of (a), (c) and (d).
Production and screening of transgenic mice. Transgenic mice were
produced by microinjection of the bK6-HPVle DNA fragment shown
in Fig. 1 (isolated after SalI and SmaI digestion) into fertilized F2 eggs
from CBA/C57B1 mice, essentially as described (Hogan et al., 1986).
Transgenic mice were identified by dot and Southern blotting of DNA
samples obtained from tail biopsy. All procedures involving animals
were carried out under the authority of appropriate Home Office
Project and Personal Licences.
RNA isolation and analysis. Standard procedures were followed
(Maniatis et al., 1982); briefly, RNA was prepared from freshly excised
tissue by homogenization in guanidine thiocyanate and ultracentrifugation through a cushion of CsC1. For Northern blotting, an estimated
15 pg RNA (unless otherwise stated) was denatured using glyoxal,
electrophoresed on 1.5 % agarose gels in phosphate buffer, transferred
to nitrocellulose and hybridized using probes (usually the entire HPV-1
early region), labelled using [c~-3zP]dCTP and a random primer
labelling kit (Pharmacia).
For mapping the site of transcription initiation directed by the
keratin promoter, a 297 bp probe was prepared by end-labelling at the
BclI site (nucleotide 226 of the viral sequence) within the HPV-1 E6
ORF using [~-3zp]ATP and T4 polynucleotide kinase, after treatment
of the Bcllqinearized plasmid with calf alkaline phosphatase.
Following digestion with XbaI which cuts within the bK6 promoter
region, the 297 bp BclI-XbaI DNA fragment spanning the junction
between the bK6 promoter region and the HPV-1 early region was
isolated by gel electrophoresis. The hybridization reactions containing
a molar excess of the denatured radioactive probe and 30 Ixg of total
RNA were incubated for 16 h at 53 °C in 80% formamide, 0.4 M-NaC1,
40 mM-PIPES pH 6.4. After treatment with S1 nuclease (250 units,
30 min at 37 °C), the products were separated by electrophoresis on a
denaturing 5 % polyacrylamide gel. A portion of the probe was treated
with dimethylsuphate followed by piperidine cleavage (Maniatis et al.,
1982) to generate a G ladder.
Histology and immunofluorescence. Tissues for histology were fixed in
Carnoy's fixative, embedded in paraffin wax and 8 Ixm sections were
stained with haematoxylin and eosin.
For immunofluorescence, tissues were frozen in liquid nitrogen.
Keratins were detected in 4 p.m frozen sections using monospecific
rabbit sera (kindly provided by Dr S. Yuspa; Roop et al., 1984); the
HPV-1 El-E4 protein was detected using a mouse monoclonal
antibody (Doorbar et al., 1988), after post-fixation of the sections with
4% paraformaldehyde. Antigens were visualized using appropriate
fluorescein isothiocyanate-conjugated secondary antibodies (Harlow &
Lane, 1988).
Results
Production o f transgenic mice and transgene expression
T h e D N A c o n s t r u c t i n t r o d u c e d i n t o t r a n s g e n i c m i c e in
t h i s s t u d y c o n t a i n e d 2.2 k b o f u p s t r e a m r e g u l a t o r y
s e q u e n c e s f r o m t h e b K 6 g e n e j o i n e d to 4.5 k b o f H P V - 1
DNA, encompassing the entire early region transcript i o n u n i t (Fig. 1). F i v e f o u n d e r t r a n s g e n i c m i c e w e r e
obtained carrying the bK6-HPVle
transgene, and
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
1254
J. M. Tinsley, C. Fisher and P. F. Searle
Southern blots confirmed the presence of the intact gene
in each of the mice (data not shown), at one or two copies
per cell (mice 27-5 and 48-16), at approximately 10 copies
(mice 48-10 and 48-11) and at 15 to 20 copies (mouse
27-7). Each of these founder animals was mated to
provide transgenic progeny for further analysis.
To look for papillomavirus gene expression in the skin
of transgenic mice descended from each of the founder
animals, RNA was extracted from the tails and analysed
by Northern blotting for the presence of HPV early
region transcripts. As shown in Fig. 2(a), low levels of
transcripts were detected in the sample from line 27-7;
the most prominent of these was approximately 1.1 kb in
length; faint bands at about 1.7 and 4 kb were just visible
on the original autoradiograph. No transgene expression
could be detected in any of the other lines, and we
therefore concentrated on line 27-7 in our subsequent
studies.
Fig. 2(b) shows a Northern blot of RNA from tissues
of a 27-7 hemizygous mouse. The tail contained the
highest levels of the 1.1 kb transcript, but it was also
detectable in paws, ears, dorsal skin and tongue. No
HPV transcripts were detected in the intestine (the high
Mr smear in this and other lanes was attributed to DNA
contamination), skeletal muscle or brain. The transgenic
mouse used as the source of these RNA samples was
normal in appearance. We have occasionally observed
the development of a very flaky appearance of the tail
skin of young hemizygous transgenic mice, apparently
associated with epidermal hyperproliferation (see
below). A similar though more pronounced phenotype
was reproducibly observed in homozygous transgenic
mice in this line.
A comparison of HPV early region transcripts in
paws, ears, dorsal skin and tail of a young homozygous
transgenic mouse from line 27-7 is shown in Fig. 2(d).
The intensity of the bands suggested that the level of
RNA expression was at least an order of magnitude
higher than that seen in samples from hemizygotes.
Highest expression was again observed in the tail
sample, followed by paws, skin, then ears (taking the
intensity of the fl-actin signal as an internal control for
RNA loading). In addition to the major 1.1 kb mRNA,
the 1-7 kb RNA was clearly visible in the paw sample,
and was particularly prominent in the tail. The 4 kb
RNA visible in ears, paws and particularly tail is the
expected length for unspliced, full-length early region
transcripts. A direct comparison between the levels of
HPV RNA in a hemizygous animal (of normal appearance) and a homozygous 27-7 mouse, both 14 days old, is
shown in Fig. 2(c). The 1.1 kb RNA was at least 10-fold
higher in the homozygous tail compared with the
hemizygote, and the 1-7 kb RNA was increased by a
greater factor in the homozygote, being barely detectable
in the hemizygote. Such an increase in the level of
transgene expression in the homozygous mice was
considerably greater than expected from the twofold
difference in transgene copy number. As discussed
further below it appears likely that the keratin promoter
of the transgene was up-regulated as a specific response
to epithelial hyperproliferation.
The 1.1 kb HPV mRNA detected in the transgenic
mice (Fig. 2) corresponded in size to the spliced
EWE4,E5a mRNA detected as a major species in
HPV-1 warts as indicated in the lower part of Fig. 1
(Chow et al., 1987). This is believed to initiate at a
promoter towards the end of the E7 ORF. We found that
our 1-1 kb RNA can be detected with an E4 sequence
probe; furthermore use of the polymerase chain reaction
and primers designed to detect possible novel splices
between the E6/E7 region and E4 provided evidence for
use only of the known splice sites indicated in the lower
part of Fig. 1 (results not shown). It therefore seemed
most likely that the 1.1 kb RNA in the transgenics
resulted from initiation not at the bovine keratin
promoter, but at a natural viral site of mRNA initiation
within the HPV E7 sequence, with appropriate splicing
shortly downstream of the beginning of E1 to the E4
ORF.
The 1.7 kb RNA corresponds in size to the
E6,E7,EIi^E4,E5a detected in warts (Chow et al., 1987).
Although the normal viral initiation site for this m R N A
is included in the construct, the viral promoter has been
replaced by that of the bovine keratin gene. S1 nuclease
mapping using a probe end-labelled at nucleotide 226 of
the viral sequence (within the E 6 0 R F ) produced a
cluster of three strong and five faint bands 21 to 28 bp
downstream of the sequence TATATAA within the
keratin promoter region (Fig. 3a, lane 2; bold arrow). As
indicated in Fig. 3(b), the most abundant RNA
transcript therefore contained a total of 18 nucleotides
corresponding to the keratin gene and linker sequences
before entering the HPV-1 sequence at nucleotide 75 of
the viral genome. A very faint band (indicated by the
small arrow in Fig. 3a, lane 2) was also observed,
corresponding to a low level of initiation at viral
nucleotide 76; this is about 20 bp upstream of the
probable start site within the viral DNA. The major
RNA species initiating from the keratin promoter in the
transgenic mice are therefore approximately 38 nucleotides longer at their 5' end relative to the corresponding
viral transcripts; this extra sequence does not contain an
AUG initiation codon.
Epithelial abnormalities
Homozygous transgenic mice in line 27-7 appeared
normal at birth; however by day 7 the skin of the tail
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
Epidermal abnormalities in HPV-1 transgenics
(a)
1
2
3
4
5
1255
6
(al
(b)
Fig. 4. Tail morphology. Tails of 9 day old mice are shown: (a) normal;
(b) homozygousbK6-HPVle transgenic mouse from line 27-7.
5 '-TTCCATATATAAGCTGCTGCCGGGAAGC
3 '---GG
.........
G--G--GG
AT CTCGAGGCAGGTAAGACTCT
- -G-G
.... G ........
TCC T CTCATAGAT
......
G-GG-G-G
GCACAGGAC
G-G--G-G
......
CAGATGG
.... GG ......
CTGG
G---
CGACAC
C- 3 '
G- -G-GG-
5 '
Fig. 3. Mapping the transcription initiation site. (a) Autoradiograph of
S1 nuclease mapping gel. Lanes 1 and 4 show cleavage of the Y-endlabelled DNA probe at G residues. Lane 2 shows protection of the
probe by tail RNA from a homozygous27-7 bK6-HPVle transgenic
mouse; the bold arrow indicates the length of probe protected by the
most abundant transcript; the smaller arrow indicates the positionof a
weak band corresponding to initiation within viral sequence. Lane 3,
control S1mappingreaction using RNA from a non-transgenicmouse.
Lanes 5 and 6, size markers producedby digestingthe probe with BgllI
and XhoI respectively. (b) DNA sequence around the junction of the
bK6 promoterand HPV-1 sequence. Viral sequence is indicated by the
hatched bar above the sequence. The bK6 TATA box and the ATG
initiation codon of the HPV-1 E6 gene are shown in bold type. The
BgllI and XhoI sites used to generate size markers are underlined. On
the lower strand only G residues are indicated, to correspond with the
G ladder shown on the gel. The positions and relative sizes of the
arrows indicate the sites of transcription initiation in the transgenic
mice, and their relative usage as determined by S1 mapping.
became abnormally hard; a few days later the surface
broke up into large flakes, often forming a series of rings
around the tail (Fig. 4b; compare with normal in Fig.
4a). At later times the distal region of the tail generally
failed to grow normally and atrophied, leaving a tail
stump usually between 30 to 70 % of the normal length by
the time the mice were 3 weeks old. At the same time, the
skin on the remaining region of the tail reverted to a
normal appearance. The appearance of hemizygous
transgenic animals was identical to that of control
animals at this age.
Histologically the abnormalities of 27-7 homozygote
tails were most conspicuous between about days 7 and
14; longitudinal sections through the tails of 7 day old
mice are shown in Fig. 5. In a normal tail (Fig. 5 a), the
epithelium had a well ordered appearance; the replicative basal layer of the epidermis followed a regular
'saw-tooth' profile, which would lead later to the
characteristic scale pattern of the adult mouse tail. The
tail epithelium of a homozygous transgenic mouse (Fig.
5c) appeared disorganized by comparison; the basal
layer followed a more highly indented, less regular course
and the thickness of the suprabasal layer was increased
around twofold, due to an increase in the number of cell
layers and also to the fact that many of the cells retained a
cuboidal shape rather than flattening out shortly after
leaving the basal layer. The nucleus to cytoplasm ratio
also appears higher than in the normal tail. Whereas the
normal tail at this age showed a fairly uniform granular
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
1256
J. M. Tinsley, C. Fisher and P. F. Searle
Fig. 5. Histology of skin from tail and lower back. (a), (b) and (c) show longitudinal sections through tail skin; (d), (e) and ( f ) show
sections through dorsal skin. All tissues were collected from 7 day old animals, and 8 pm sections were stained with haematoxylin and
eosin. (a) and (d) are from normal (non-transgenic) mice; (b) and (e), hemizygous 27-7 transgenic; (c) and (f), homozygous 27-7
transgenic (see text). The opposed open and filled arrowheads indicate respectively the transition between the granular (suprabasal)
layer of the epidermis and the dead stratum corneum (~7), and the position of the basal membrane separating the basal layer of the
epidermis (above) from the dermis (A). All are shown at the same magnification; the bar marker represents 100 pm.
layer about two cells thick between the living suprabasal
cells and the dead stratum corneum, the granular layer in
the affected transgenics in some regions could be up to
four cells thick, whereas elsewhere it was not apparent.
Finally, the cornified layer in the affected tails was
hyperkeratotic, generally two to three times the thickness
of that in normal tails and also rather more compact or
dense in appearance. Regions of parakeratosis in which
the terminal differentiation of the keratinocytes was
abnormal leading to the persistence of densely staining
cell nuclei in the outer layers were also commonly
observed in the affected transgenic animals' tails (not
shown).
Most hemizygous 27-7 transgenics were indistinguishable from normal animals; however, occasionally the tail
skin of an entire litter of hemizygous 27-7 transgenics
developed a flaky appearance at around 2 weeks of age
(not shown). This lasted only a few days, and was always
milder than the homozygous phenotype. This sporadic
phenotype has been seen in litters from several crosses
between a homozygous 27-7 transgenic mouse of either
sex, mated to a normal animal; both previous and
subsequent litters born to the same parents were
generally normal in appearance. The appearance of the
hemizygous phenotype appeared to correlate with
increased levels of the HPV RNAs presumably induced
by an unidentified environmental cofactor (data not
shown). Fig. 5 (b) shows a stained section through the tail
skin of a day 7 hemizygous mouse in which hyperplasia
of the suprabasal layers was apparent, although the
organization is clearly less abnormal than in the
homozygote (Fig. 5 c). This section probably represents
the development of the sporadic 'flaky' phenotype, since
the tail skin of the surviving hemizygous littermates of
this animal became very flaky for a few days around days
12 to 14 after birth.
In homozygous 27-7 transgenics a transient flakiness
of the skin on other regions of the body, particularly
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
Epidermal abnormalities in HPV-1 transgenics
1257
Fig. 6. Detection of El-E4, K10 and K6 proteins on longitudinalsectionsof tail skin. The antigens visualizedby immunofluorescence
are HPV-1 El-E4 (a, b and c; the fluorescenceof the dermisin these sectionsis non-specific,see text); K 10 (d, e and f ) and K6 (g, h and
i). Sections(a), (d) and (g) show normal (non-transgenic)tissue; (b), (e) and (h) are from a hemizygous27-7 transgenicmouse, and (c),
(f) and (i) are from a homozygous27-7 transgenic mouse. All tissues are from 7 day old animals, and all panels show comparable
sections at the same magnification; the bar marker represents 100 gm.
on the lower back and around the limb joints was
also generally observed around day 7 after birth, subsequently reverting to apparent normality after a few
days. As shown in Fig. 5 ( f ) , this was also associated
with epidermal hyperplasia. Both at a gross level and
histologically, these regions of the skin of hemizygotes
(Fig. 5 e) never differed appreciably from non-transgenic
animals (Fig. 5d).
E l - E 4 protein expression
The E l - E 4 protein is a major protein in HPV-1 warts,
where it may contribute up to 30% of the total protein
content (Doorbar et al., 1986). As discussed above, the
1.1 kb transcript detected in the 27-7 transgenic mice
appears to correspond to the viral 1.1 kb transcript which
encodes the E l - E 4 protein. Using a monoclonal antibody against E4 (Doorbar et al., 1988), the viral protein
was detected in the upper suprabasal layers of the skin in
both hemizygous and homozygous mouse tails (Fig. 6b,
c), whereas no specific reactivity was seen in controls
(Fig. 6a; the relatively high background fluorescence of
the underlying dermal tissue was also seen when t h e E 4
primary antibody was omitted, and is attributed to the
use of anti-mouse secondary antibody). The layer of cells
expressing E l - E 4 is appreciably thicker in the homozygous t h a n the hemizygous tails; assuming this to be a
refection of the distribution of the 1.1 kb m R N A , this
greater thickness may contribute to the greater abundance of the 1.1 kb m R N A in homozygous mice, since
the R N A was prepared from the whole tails. Weaker E4
immunofluorescence was also observed in the footpads
of both hemizygous and homozygous 27-7 transgenic
mice (results not shown).
Changes in expression of mouse keratins
The keratins are characteristic markers of the state of
epithelial cell differentiation, and we therefore examined
the 27-7 b K 6 - H P V l e transgenic animals' tails for
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
1258
J. M. Tinsley, C. Fisher and P. F. Searle
abnormalities of keratin expression using a panel of
monospecific antibodies (Roop et al., 1984). Several
changes in keratin expression were observed in the tails
of homozygous transgenic mice as shown in Fig. 6 (d) to
(i); these changes were most clearly defined at around
day 7 after birth.
Keratin K10 is generally associated with normal
differentiation of squamous epithelium (Moll et al.,
1982; Stoler et al., 1988), and as expected normal mouse
tail showed strong expression of K10 throughout the
suprabasal layers (Fig. 6d). The epitope appears to be
less accessible in the outer cornified layers. This same
pattern of expression was seen in the 27-7 hemizygote
(Fig. 6e). In homozygous animals K10 was strongly
down-regulated; this appeared to be somewhat asynchronous in different regions of the tail, as in some
regions cells still containing K10 could be seen in the
upper layers of the skin. However as shown in Fig. 6 (f),
by day 7 K10 was virtually undetectable throughout
much of the hyperplastic epithelium of the homozygous
animal. Keratin K 1, which is normally coexpressed with
K10, showed a similar distribution (results not shown).
Expression of keratins K6 and K16 has been linked
with epithelial hyperproliferation (Weiss et al., 1984;
Stoler et al., 1988), although as shown in Fig. 6(g), K6 is
expressed in the outer root sheath of the hair follicles in
normal epidermis and low levels of K 16 were detected in
the basal layer (not shown). Strikingly, both K6 (Fig. 6i)
and K16 (result not shown) were strongly expressed
throughout the basal and suprabasal layers of the
hyperplastic tail epithelium of the 27-7 homozygote. In
the hemizygote, K16 was not detectable above control
levels (results not shown) and only occasional clusters of
cells expressing K6 were observed in the epidermis,
generally around the join between the outer root sheath
of the hair follicles and the interfollicular epidermis (Fig.
6h). The hemizygous sections shown in this figure were
from a litter of normal appearance, thus despite the
El-E4 expression epidermal hyperplasia was not observed. The dramatic induction of the endogenous mouse
K6 expression in the homozygous 27-7 transgenics correlated with the strong increase in expression of the
transgene observed in these animals from the bovine K6
promoter used in this construct.
Discussion
Young mice homozygous for the bK6-HPVle transgene
locus in line 27-7 have been shown to develop a striking
transient disturbance of normal epithelial differentiation. The effect was greatest on the tail, where the
epithelium became hyperproliferative in appearance,
with several layers of irregular or cuboidal cells above the
basal layer, an increase in the total number of cell layers,
and abnormal cornification. A similar transient flaky
appearance also associated with epidermal hyperplasia
was observed on other regions of the skin around 7 days
of age. The hyperproliferation-associated keratins K6
and K16 were shown to be highly induced in the tail,
while the high Mr keratins K1 and K10 were downregulated to virtually undetectable levels.
The effects described were observed in only one line of
transgenic mice (the only line in which transcripts of the
HPV transgene were detectable), and were consistently
observed only in mice homozygous for the transgene. It
could therefore be argued that the phenotype may be due
to the mutation of a host gene affecting the control of
epithelial growth and differentiation, associated with the
essentially random process of transgene integration.
However for several reasons discussed below, it appears
considerably more likely that the phenotype is induced
by HPV gene products. First, it is clear that the capacity
for normal epithelial differentiation has not been
irretrievably lost, because although in most cases the
epithelial abnormality led to the loss of a significant
length of the tail, the remainder of the tail reverted to a
normal appearance, as did the skin on the rest of the
mouse. Second, there was strong evidence linking
expression of HPV early region transcripts with the
abnormality. For example the highest level of HPV
transcripts was found in the tail coincident with the
greatest disturbance of epidermal differentiation, and
the strong up-regulation of expression in the homozygous
transgenics both correlated with the abnormality and
provided a rationale for the difference in phenotype
between hemizygotes and homozygotes. Furthermore,
although the appearance of the hemizygotes was generally normal, we have occasionally observed a flakiness of
the tail of young hemizygous animals also, which was
associated with a similar epidermal hyperplasia. This is
believed to be due to up-regulation of the transgene by an
unidentified environmental factor, since this flakiness of
hemizygous animals' tails has been seen only occasionally but tended to affect an entire hemizygous litter when
it was observed. Previous and subsequent litters born to
the same parents were unaffected. Although this
sporadic hemizygous phenotype has not been observed to
develop to the same extent as is usual in the homozygores, it is consistent with the interpretation of the
difference between hemizygotes and homozygotes being
related to the level of transgene expression, rather than to
insertional mutagenesis of a host gene.
The great increase in transgene expression in homozygous 27-7 transgenic mice was initially unexpected;
normally (and in the absence of negative feedback
mechanisms) a transgenic mouse with two copies of a
transgene locus on the two homologous chromosomes
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
Epidermal abnormalities in H P V - 1 transgenics
would be expected to express twice the level of
transcripts as a mouse hemizygous for that transgene
locus. However in this particular case we suggest that t h e
transgene expression can induce a hyperproliferative
state within the cell, and that the transgene expression is
further induced as a response to this state. For reasons
discussed above, we believe it was HPV-1 proteins
encoded by the transgene which induced the epidermal
keratinocytes of the tail to a state of hyperproliferation,
as evidenced by the histological appearance and upregulation of keratins 6 and 16. It is most likely that this
state was induced by the E6 and/or E7 proteins, encoded
by the 1.7 kb m R N A . The up-regulation of the mouse K6
(and K16) may have occurred either as a direct response
to viral proteins, or perhaps more probably as a
secondary consequence of the hyperproliferation.
Because of the use of a (bovine) keratin 6 promoter to
control the transgene, it is understandable that transgene
expression should have been increased in parallel to the
observed expression of mouse K6, completing a positive
feedback loop to increase the transgene expression. The
question remains why this positive feedback loop was
usually only activated in homozygous animals. We
suggest that the establishment of the hyperproliferative
state requires a certain threshold level of the H P V
proteins, and that hemizygous transgenics in this line
generally fail to reach this threshold (but may do so if
there is an initial environmental stimulus, to account for
the sporadic appearance of the flaky phenotype in
hemizygotes). In homozygous animals on the other hand
we propose that an initial level of transgene expression
twice that in hemizygotes exceeds the threshold that
leads to hyperproliferation, thus activating the positive
feedback loop. Analysis of the histology and keratin
expression in younger mice suggests that the cells
become hyperproliferative with up-regulation of K6 at
around 2 to 3 days after birth (unpublished data).
It is striking that the epithelial abnormalities described are transient, with the skin reverting to a normal
appearance by about 3 weeks after birth. This could
indicate the existence of a negative feedback pathway
which operates to limit the hyperproliferative response.
Alternatively, the tails and skin of normal mice undergo
rapid growth and differentiation during the period when
the abnormalities were observed in the transgenic mice;
thus the scale pattern is established in the tail with upregulation of further differentiation-specific keratins
(Schweizer et al., 1987), and the fur develops over the
body. It could be that the HPV-1 proteins can induce
the effects described only in skin undergoing these
changes, or in conjunction with internal factors present
for a limited time while such changes are occurring. A
similar effect might underlie the specific association of
HPV-1 with the thickened epithelium of the hands and
1259
feet in man. Investigation of these questions m a y provide
interesting insight into the interactions of papillomaviruses with the cellular and tissue environment. It is also
not clear at present what causes the distal region of the
tail to degenerate in the homozygous transgenics; it is
possible that this is a side effect not directly related to the
cellular effects of the viral proteins; for example the
physical restraint imposed by the hard hyperkeratinized
rings around the tail at a time when it is growin~ ;apidly
may restrict the proper development of the blood supply
to the distal region.
The histological appearance of the affected mouse tails
strongly suggests that the epidermal cells are hyperproliferative, and expression of keratins 6 and 16 as seen
in the tails is generally associated with epidermal hyperproliferation (Weiss et al., 1984; Stoler et al., 1988).
Concomitantly with the expression of keratins 6 and 16,
the differentiation-specific keratins 1 and 10 were downregulated to undetectable levels, as observed in
the productively infected cells within HPV-1 warts
(Breitburd et al., 1987). The E l - E 4 protein was
expressed in the upper layers of the transgenic skin,
mimicking its distribution within warts. The ability to
reproduce these characteristic disturbances in the control of epithelial cell proliferation and differentiation in
H P V transgenic mice indicates that this m a y be a useful.
model system to study the mechanism of action of H P V
proteins within an otherwise normal epithelium.
We thank Dr J. L. Jorcano and Dr M. Blessingfor the kind gift of the
plasmid pbK6-CAT, and Dr T. Broker for the plasmid pPXL-HPV-1.
Monospecific anti-mouse keratin antisera were the generous gift of Dr
S. H. Yuspa, and the E4 monoclonal antibody was kindly provided by
Dr J. Doorbar. We thank Professor P. H. Gallimore, Dr J. L. Jorcano,
Dr J. Schweizerand Dr S. H. Yuspa for helpful discussions. This work
was supported by the Cancer Research Campaign.
References
BROKER, T. R. & BOTCHAN,M. (1986). Papillomaviruses: retrospectives and prospectives. Cancer Cells 4, 17-36.
BLESSING, M., JORCANO, J. L. 8~ FRANKE, W. (1989). Enhancer
elements directing cell-type-specificexpression of cytokeratin genes
and changes of the epithelial cytoskeletonby transfections of hybrid
cytokeratin genes. EMBO Journal 8, 117-126.
BREITBURD,F., CROISSANT,O. & ORTH, G. (1987). Expression of
human papillomavirus typed E4 gene products in warts. Cancer
Cells 5, 115-122.
CHow, L. T., REILLY,S. S., BROKER,T. R. & TAICHMAN,L. B. (1987).
Identification and mapping of human papillomavirus type 1 RNA
transcripts recovered from plantar warts and infected epithelial cell
cultures. Journal of Virology 63, 1913-1918.
DEVILLIERS,E. M. (1989). Heterogeneityof the human papillomavirus
group. Journal of Virology 63, 4898-4903.
DOORBAR,J., CAMPBELL,D., GRAND,R. J. A. & GALL1MORE,P. H.
(1986). Identification of the human papillomavirus type-la E4
products. EMBO Journal 5, 355-362.
DOORBAR,J., EVANS,H. S., CONERON,I. & CRAWFORD,L. V. (1988).
Analysis of HPV-1 E4 gene expression using epitope defined
antibodies. EMBO Journal 7, 825-833.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22
1260
J. M . Tinsley, C. Fisher and P. F. Searle
DYSON, N., HOWLEY,P~ M,, MUN6ER, K. & HARLOW,E. (1989). The
human papillomavirus-16 E7 oncoprotein is able to bind to the
retinoblastoma gene product. Science 243, 934-936.
GROSS, G., PFISTER, H., HAGEDORM, M. & GISS~ANN, L. (1982).
Correlation between human papillomavirus (HPV) type and
histology of warts. Journal of Investigative Dermatology 78, 160--164.
HALBERT,C. L., DEMERS,G. W. & GALLOWAY,D. A. (1991). The E7
gene of human papinomavirus type 16 is sufficient for immortalization of human epithelial cells. Journal of Virology 65, 473-478.
HARLOW,E. & LANE,D. (1988). Antibodies: A Laboratory Manual. New
York: Cold Spring Harbor Laboratory.
HOGAN, B., CONSTANTINI,F. & LACY, E. (1986). Manipulating the
Mouse Embryo: A Laboratory Manual. New York: Cold Spring
Harbor Laboratory.
HOWLEY, P. M., YANG, Y.-C., SPALHOLZ, B. A. & R.ARSON, M. S.
(1986). Papillomavirus transforming proteins. In Papillomaviruses:
Ciba Foundation Symposium, vol. 120, pp. 39-52. Edited by
D. Evered and S. Clark. Chichester: John Wiley & Sons.
KONDOH, G., MURATA,Y., AOZASA,K., YUTSUDO,M. & HAKURA,A.
(1991). Very high incidence of germ cell tumorigenesis (seminomagenesis) in human papillomavirus type 16 transgenic mice. Journal of
Virology 65, 3335-3339.
LACEY,M., ALPERT,S. & HANAHAN,D. (1986). Bovine papillomavirus
genome elicits skin tumours in transgenic mice. Nature, London 322,
609-612.
LtNDGREN, V., SIPPOLA-THIELE, M., SKOWRONSKI, J., WETZEL, E.,
HOWLEr, P. M. & HANAnAN, D. (1989). Specific chromosomal
abnormalities characterize fibrosarcomas of bovine papillomavirus
type 1 transgenic mice. Proceedings of the National Academy of
Sciences, U.S.A. 86, 5025-5029.
MANIATIS, T., FRITSCH, E. F. & SAMBROOK, J. (1982). Molecular
Cloning: A Laboratory Manual. New York: Cold Spring Harbor
Laboratory.
MOLL, R., FRANKE,W. W., SCHILLER,D. L., GEIGER, B. & KREPLER,
R. (1982). The catalog of human cytokeratins: patterns of expression
in normal epithelia, tumors and cultured cells. Cell 31, 11-24.
MUNGER, K., PHELPS,W. C., BUBB,V., HOWLEY,P. M. & SCI-ILEGEL,
R. (1989). The E6 and E7 genes of the human papillomavirus type 16
together are necessary and sufficient for transformation of primary
human keratinocytes. Journal of Virology 63, 4417-4421.
PFISTER, H. (1984). Biology and biochemistry of papiUomaviruses.
Reviews of Physiology, Biochemistry and Pharmacology 99, 111-181.
RooP, D. R., CHENG, C. K., TITrERTON,L., MAYERS,C. A., STANLEY,
I. R., STEINERT, P. M. & YUSPA, S. H. (1984). Synthetic peptides
corresponding to keratin subunits elicit highly specific antibodies.
Journal of Biological Chemistry 259, 8037-8040.
SCHEFFNER,M., WERNESS,B. A., HUIGBREGTSE,J. M., LEVINE,A. J. &
HOWLEY, P. M. (1990). The E6 oneoprotein encoded by human
papillomavirus types 16 and 18 promotes the degradation of p53. Cell
63, 1129-1136.
SCnWARZ, E., FREESE, U. K., GlssMANN, L., MAYER, W.,
ROGGENEUCK, B., STREMLAU, A. & ZUR HAUSEN, H. (1985).
Structure and transcription of human papillomavirus sequences in
cervical carcinoma cells. Nature, London 314, 111-114.
SCHWmZER, J., FVRSTENBERGER,G. & WINTER, H. (1987). Selective
suppression of two postnatally acquired 70 kD and 65 kDa keratin
proteins during continuous treatment of adult tail epidermis with
vitamin A. Journal oflnvestigative Dermatology 89, 125-131.
STOLER, A., KOPAN, R., DUVIC, M. & Fucrts, E. (1988). Use of
monospecific antisera and cRNA probes to localize the major
changes in keratin expression during normal and abnormal
epidermal differentiation. Journal of Cell Biology 107, 427-446.
SYRJ,~NEN,K., GISSMANN,L. & KOSS.L.G. (editors) (1987). Papillomaviruses and Human Disease. Berlin & Heidelberg: Springer-Verlag.
VOUSDEN,K. (1989). Human papillomaviruses and cervical carcinoma.
Cancer Cells 1, 43-50.
WERNESS,B. A., LEVlNE,A. J. & HOWLEY,P. M. (1990). Association of
human papillomavirus types 16 and 18 E6 proteins with p53. Science
248, 76-79.
WEISS, R. A., EICnNER, R. & SUN, T. T. (1984). Monoclonal antibody
analysis of keratin expression in epidermal diseases: a 48- and
56-kdalton keratin as molecular markers for hyperproliferative
keratinocytes. Journal of Cell Biology 98, 1397-1406.
ZUR HAUSEN,H. & SCHNEIDER,A. (1987). The role of papiUomaviruses
in human anogenital cancer. In The Papovaviridae, vol. 2, Papillomaviruses, pp. 245-263. Edited by P. M. Howley & N. P. Salzman. New
York: Plenum Press.
(Received 29 August 1991 ; Accepted 7 January 1992)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 03 May 2017 20:53:22