| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,2
,4
* Institut für Molekulare Medizin und Zellforschung, Albert Ludwigs Universität Freiburg, 79106 Freiburg, Germany;
Hautklinik, Charité, Humboldt Universität, 10117 Berlin, Germany;
Abteilung Biochemie II, Zentrum Biochemie und Molekulare Zellbiologie, Georg August Universität Göttingen, 37073 Göttingen, Germany;
§ Allgemeine Pathologie und Neuropathologie, Tierärztliche Fakultät der Ludwig Maximilians Universität München, 80539 München, Germany; and
¶ Institut für Ultrastrukturforschung der Haut, Hautklinik der Ruprecht Karls Universität Heidelberg, 69115 Heidelberg, Germany
5Correspondence: Institut für Molekulare Medizin und Zellforschung, Albert Ludwigs Universität Freiburg, Hugstetter Strasse 55, 79106 Freiburg, Germany. E-mail: petersc{at}mm11.ukl.uni-freiburg.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: epidermal hyperplasia • hair follicle development • lysosomal cysteine proteinase
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. This was suggested by their high abundance in lysosomes,
ubiquitous expression in mammalian tissues, and by the observation that
a substantial portion of intracellular protein degradation can be
suppressed by protease inhibitors with broad specificity for cysteine
proteases (3)
. Besides the highly abundant, ubiquitously
expressed proteinases cathepsin L (CTSL), B, and H (4
5
6
7)
four additional members of this family with ubiquitous expression,
cathepsins C, O, Z, and F, have been characterized at the molecular
level during recent years (8
9
10
11)
.
Beyond these proteinases with broad tissue distribution, at least four
papain-like mammalian lysosomal cysteine proteinases with a restricted
expression pattern have been described. Cathepsin L2 is expressed in
thymus and testis (12)
and cathepsin W is predominantly
expressed in CD8+ T lymphocytes
(13)
, which may suggest as yet undefined specific
functions in cellular physiology. Cathepsin K is highly expressed in
osteoclasts (14)
and appears to be up-regulated at sites
of inflammation (15)
. It has potent collagenolytic and
elastinolytic activities (14)
and the identification of
mutations in its gene in patients with pycnodysostosis, an autosomal
recessive osteochondrodysplasia with osteosclerosis and short stature,
underscores its essential function in extracellular matrix remodeling
(15
, 16)
. Cathepsin S is expressed in spleen, lymphocytes,
monocytes, and other cells positive for major histocompatibility
complex (MHC) class II. Using a cathepsin S-specific inhibitor, it has
been demonstrated that this proteinase is essential for degradation of
the invariant chain of MHCII (Ii) and subsequent loading of antigenic
peptides into the antigen binding groove in peripheral
antigen-presenting cells (APCs) (17)
. Recently it was
shown that mice lacking the lysosomal cysteine proteinase cathepsin S
display a profound inhibition of Ii degradation in professional APCs
resulting in an impaired MHC class II peptide loading (18
, 19)
.
In addition to nonspecific bulk proteolysis, more specific functions in
physiological and pathophysiological processes—such as prohormone- and
antigen-processing and -presentation, atherosclerosis, pulmonary
emphysema, and cancer invasion and metastasis—have been postulated for
ubiquitously expressed papain-like lysosomal cysteine proteinases
(15
, 20)
.
Recently it has been shown that CTSL is essential for Ii processing in
cortical thymic epithelial cells but not in bone marrow-derived
antigen-presenting cells. Consequently, positive selection of
CD4+ T cells is reduced in CTSL-deficient mice
(21)
. Here we show that mice lacking CTSL develop periodic
hair loss with alteration of hair follicle morphogenesis and cycling as
well as hyperplasia and hyperkeratosis of the epidermis. Both
observations are attributed to hyperproliferation of hair follicle
epithelial cells and basal keratinocytes. Furthermore, the molecular
defect of the spontaneous mouse mutant furless is identified
as a mutation in the ctsl gene.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-FIXTMII-teratocarcinoma
library (Stratagene, La Jolla, Calif.) using a ctsl cDNA
(22)
. A 6 kb SacI-SacI DNA fragment
containing exons 1–5 was subcloned into pBluescript SKII+ and
characterized by restriction mapping and sequencing of the intron/exon
boundaries. A neomycin resistance cassette (23)
was
inserted at the NcoI site in exon 3 using a SalI
linker containing stop codons in all reading frames in both
orientations, giving rise to the targeting vector pMCL4neo (orientation
of the neo gene opposite to the ctsl gene). The
targeting vector was linearized prior to electroporation with
EcoRV (Fig. 1
) The targeting vector was introduced into E-14–1 embryonic stem (ES)
cells (24)
by electroporation. ES cells were cultured as
described (25)
. G418-resistant ES cell colonies were
screened for homologous recombination by Southern blotting using 5' and
3' external probes (Fig. 1B
). Targeted ES cell clones were
microinjected into blastocysts of C57BL/6J mice. Resulting male
chimeras were mated to C57BL/6J females, and heterozygous offspring
were intercrossed for generation of
ctsl-/ctsl-
mice. Mice were genotyped for the ctsl mutation by Southern
blot analyses of BglII-digested genomic tail DNA, using the
3'probe (Fig. 1C
). The animals used in all studies
([129Sv/J x C57BL/6J]F2 and FSB/GnEi, Jackson Laboratories, West
Grove, Pa.) were maintained and bred according to institutional
guidelines in the animal facilities of the University Medical Center,
Freiburg, Germany.
|
Northern and Western blot analyses
Total RNA of kidney from 28-day-old mice was prepared as
described (26)
. Total RNA (5 µg) was separated in a
formaldehyde agarose gel and processed as described previously
(27)
. Filters were hybridized with ctsl cDNA
(22)
and a 280 bp cDNA fragment from
glyceraldehyde-3-phosphate dehydrogenase (G3PD; 28
).
Lysosomes were enriched from kidney homogenates as described previously
(29)
. Soluble lysosomal protein (400 µg) was separated
by SDS-PAGE, blotted and probed with a rabbit polyclonal antiserum
against rat CTSL using the ECL immunodetection system (Amersham, Little
Chalfont, U.K.) as described (27)
.
Detection of CTSL enzyme activity
CTSL proteolytic activity in lysosome enriched fractions
(29)
of kidney was determined using the substrate
Z-Phe-Arg-4-methyl-coumarin-7-amide (20 µM; Bachem, Bubendorf,
Switzerland) in the presence or absence of the cathepsin B-specific
inhibitor CA-074 (20 nM) as described (1
, 30)
. Reaction
mixtures were preincubated at 37°C for 30 min. One unit corresponds
to the enzyme activity liberating 1 µmol aminomethyl coumarin per
min. Protein concentrations were determined according to Lowry et al.
(31)
.
Skin harvesting for analysis of hair follicle morphogenesis and
cycling, histological analysis of skin sections, and histomorphometry
Newborn mice (3 to 5 animals per experimental group) were used
to study the neonatal skin and hair phenotype as well as hair follicle
development and cycling. Full-thickness back skin was harvested on days
0, 2, 6, 14, 17, 20, and 28 postpartum (p.p.) perpendicular to the
paravertebral line to obtain longitudinal hair follicle sections. Skin
sections were embedded and frozen in liquid nitrogen as described
(32)
. The angling of the sections was kept identical. For
identification of defined stages of hair follicle morphogenesis and
cycling, histochemical detection of endogenous alkaline phosphatase
activity was used, since this allows one to visualize the morphology of
the dermal papilla as a morphological marker for staging of hair
follicles. Cryostat sections (8 µm) of full-thickness back skin were
stained for alkaline phosphatase activity according to standard
protocols (33)
and subsequently counterstained with
Meyer’s hematoxylin. The developmental stages of hair follicles
(morphogenesis or cycling) were classified in more than 50 hair
follicles per animal as described (34
35
36)
.
Skin organ culture
Punch biopsies (4 mm) were prepared under sterile conditions
from full-thickness back skin following described proto cols (37
, 38)
with minor modifications. Per experimental group, five to
six randomized skin punches from three mice were placed onto gelatin
sponges (Gelfoam, Upjohn Co., Kalamazoo, Mich.) and cultured as
described (36)
. After 4–5 days in culture skin punches
were harvested and subsequently embedded and frozen in liquid nitrogen
as described (32)
.
In situ hybridization
For in situ hybridizations adolescent mice were
perfused with 8% paraformaldehyde and processed according to standard
protocols. Hybridization was performed according to standard protocols
using DIG nucleotide detection kit (Boehringer Mannheim, Mannheim,
Germany). The 425 bp ctsl probes where derived from the
3'region of a murine ctsl cDNA and labeled with a DIG RNA
labeling kit (Boehringer Mannheim).
Immunohistochemistry and quantitative histomorphometry of
proliferation and apoptosis
For estimation of apoptotic cells in skin sections, a
commercially available TUNEL kit (ApopTaq, Oncor, Gaithersburg, Md.)
was used as described (39)
. For double immunofluorescence
detection of TUNEL-positive cells and cells positive for the
proliferation marker Ki67, sections were incubated with
digoxigenin-dUTP in the presence of TdT, followed by incubation with
the rabbit anti-mouse Ki67 antibody (Dianova, Hamburg,
Germany). Subsequently, TUNEL-positive cells were visualized by
anti-digoxigenin FITC-conjugated F(ab)2
fragments, and Ki67 immunoreactivity was detected by goat anti-rabbit
TRITC-conjugated antibody. Nuclei were counterstained using Hoechst
33242 dye (Sigma, St. Louis, Mo.). For quantitative histomorphometry,
the number of Ki67-positive cells in the basal layer of the
interfollicular epidermis was counted in relation to the total number
of interfollicular basal cells. Three to five microscopic fields
(magnification: x400) on three individual cryostat sections per animal
were analyzed. Statistical significance was estimated using the
Wilcoxon-Mann-Whitney U test. Differences were judged as
significant if P < 0.05.
Quantitative histomorphometry of skin thickness
Epidermal and dermal thickness was assessed from
n = 3 mutant and wild-type animals, each on day 14 p.p.; three to five microscopic fields of three to five routinely
stained cryostat sections (8 µm) from each animal were analyzed
morphometrically. Statistical significance was estimated using the
Wilcoxon-Mann-Whitney U test. Differences were judged as
significant if P < 0.05.
For morphometric analyses of epidermal thickness from 3-month-old ctsl-/ctsl- mice and wild-type controls (5 animals per experimental group), 2 µm paraffin sections of neck and tail skin were studied. The epidermal thickness was analyzed rectangular to the basal lamina, excluding the stratum corneum because of its frequent embedding artifacts, using an automatic image analyzing system with random systematic sampling (Kontron, Eching, Germany). Statistical significance was estimated using the Wilcoxon-Mann-Whitney U test. Differences in neck (P<0.01) and tail skin (P<0.05) were judged as significant.
Sequence analysis of furless ctsl gene
Genomic DNA of fs mice was obtained from The Jackson
Laboratory. Six genomic DNA fragments covering all exons of the
ctsl gene were polymerase chain reaction (PCR) amplified
from DNA of fs and 129Sv/J control mice. Amplification
primers were removed and PCR-amplified DNA fragments were directly
sequenced. Both strands of the entire open reading frame (exons 1–8)
of the ctsl gene of fs and of 129Sv/J control
mice were analyzed using an AmpliTaq Cycle sequencing kit and an ABI
373 DNA sequencer. Detailed information about oligonucleotides used for
PCR amplification and sequencing are available on request.
Transfection of
ctsl-/ctsl- mouse
embryonic fibroblasts
Cstl-/ctsl-
fibroblasts were generated from day 12.5 embryos and immortalized by
transfection with the SV40 large T antigen expression vector pMSSVLT
(40)
. The wild-type ctsl cDNA (22)
and the fs mutant cDNA constructed by subcloning of
PCR-amplified exon 5 into the wild-type cDNA were inserted into the
expression vector pMPSVEH (41)
.
Ctsl-/ctsl-
embryonic fibroblasts were stably cotransfected with wild-type,
fs mutant, or expression plasmid vector and a hygromycin
resistance plasmid as described (42)
. CTSL enzyme activity
was measured in lysosomal fractions of
ctsl-/ctsl-
fibroblasts stably transfected with plasmid expression vector,
wild-type ctsl expression plasmid, or fs mutant
ctsl-allele expression plasmid, respectively. The
CTSL-specific activity was measured with
Z-Phe-Arg-4-methyl-coumarin-7-amide as substrate in the presence of the
cathepsin B-specific inhibitor CA-074 (± SD;
n=3).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The targeting vector used for disruption of the
ctsl gene (Fig. 1A
) is based on a 6 kb DNA
fragment covering part of the promoter region, exons 1–5, and the
respective introns of the ctsl gene. The open reading frame
was interrupted in exon 3 by insertion of a linker, which contains stop
codons in all reading frames. A neomycin cassette (23)
was
inserted into the linker in opposite orientation to the ctsl
gene. The truncated open reading frame codes for the leader peptide and
part of the CTSL proregion. Hence, the introduced mutation is most
likely a null mutation.
Four targeted ES cell clones were identified among 47 colonies
screened by Southern blot analysis with 5' and 3' external probes (Fig. 1B
). Chimeric animals were generated by injection of
targeted ES cells into C57BL/6J blastocysts. Chimeric males were mated
with C57BL/6J females. Transmission of the mutant allele through the
germline was confirmed by Southern blot analysis (data not shown).
Heterozygous animals did not show differences in phenotype or fertility
as compared to wild-type littermates (data not shown). Genotyping of
331 offspring from heterozygote crosses (Fig. 1C
) revealed a
frequency of 28.4% of homozygous mutant mice
(ctsl-/ctsl-)
resembling the expected Mendelian frequency and excluding embryonal
lethality.
Ctsl gene expression in mutant mice was tested in kidney, a
tissue with high ctsl-expression in wild-type mice
(43)
, by Northern and Western blot analyses. Neither
ctsl transcripts were detectable in RNA from homozygous
mutant animals (Fig. 1D
) nor was CTSL protein present in
lysosomal protein extracts (Fig. 1E
). Using the synthetic
substrate Z-Phe-Arg-4methyl-coumarin-7-amide, which is cleaved by
cathepsins L and B, in the presence of the cathepsin-B-specific
inhibitor CA-074 (30)
, no CTSL-activity was detected in
kidney from
ctsl-/ctsl-
mice (Fig. 1F
). These data indicate that the
ctsl gene was successfully inactivated.
Ctsl-/ctsl- mice show retarded hair growth
and develop periodic hair loss
Up to weaning, the mortality of
ctsl-/ctsl-
mice is elevated to 15% as compared to 6% in wild-type littermates;
thereafter, mutant mice exhibit a normal mortality for an interval of
more than 50 wk;
ctsl-/ctsl-
mice are fertile. At the day of birth, vibrissae, which are the first
hair follicles to develop during fetal life, have not penetrated the
epidermis in
ctsl-/ctsl-
mice (data not shown) in contrast to wild-type littermates
(44)
.
Between days 7 and 9 p.p. the skin of mutant mice has a shiny and
squamous appearance, and the first emergence of fur is delayed by 2
days (Fig. 2A
). Thereafter, fur development macroscopically proceeds
apparently normally until approximately day 21 p.p. At this time
ctsl-/ctsl-
mice start to lose their fur, beginning at the head and progressing
toward the tail region of the back. Between approximately day 28 and
30 p.p.,
ctsl-/ctsl-
mice are almost nude (Fig. 2B
). Thereafter, a new coat
starts to grow, i.e., during the onset of the first growth phase of the
hair follicle cycle, the anagen phase (Fig. 2C
; 44
, 45
). At 7 wk of age, a new wave of spatially restricted hair
loss occurs (not shown). Mature
ctsl-/ctsl-
mice are always partially devoid of hair (Fig. 2D
).
|
Because of the stringent timing and control of hair follicle
development and cycling in neonatal mouse skin (34
, 47)
,
the loss of pelage hair starting at day 21 p.p. suggests that CTSL
deficiency disturbs normal telogen development, i.e., the programmed
transition of hair follicles from catagen, the apoptosis-driven
regression phase of the hair cycle (39)
, to telogen, a
period of relative hair follicle resting.
Hair follicle morphogenesis and progression through the hair cycle
are delayed in
ctsl-/ctsl-mice
To investigate the periodic loss and regrowth of pelage hair in
ctsl-/ctsl-
mice in detail, hair follicle morphogenesis and follicle cycling were
analyzed by quantitative histomorphometry (34
35
36
, 41
, 47)
. Staging of hair follicle morphogenesis (34
, 48)
in cryosections of day 6 p.p. back skin confirmed the
macroscopic observation of a delayed appearance of the first fur in
ctsl-/ctsl-
mice (Fig. 2A
).
In the skin of wild-type animals, 49% of hair follicles have reached
stage 8 of follicle morphogenesis, which is characterized by
penetration of the hair shaft through the epidermis (47
;
Fig. 3A
, C
). In contrast, 84% of hair follicles from
ctsl-/ctsl-
mice are still in stage 7 of morphogenesis, with its substantially
shorter hair shafts not yet emerging through the skin (Fig. 3B
, C
). Densities of hair follicles were found
not to differ significantly between wild-type and
ctsl-/ctsl-
skin (data not shown).
|
After the completion of morphogenesis, hair follicles enter into the
first hair cycle at day 17 p.p. The murine hair cycle is composed
of three major phases: the anagen or growth phase, the catagen or
regression phase, and the telogen or resting phase (for review, see ref
49
). At day 17 p.p., the same stage of catagen
development was observed in
ctsl-/ctsl- and
wild-type skin (data not shown), indicating that hair follicles of both
genotypes initiate the first hair cycle at the same time by
synchronized entry into catagen. However, quantitative histomorphometry
of day 20 p.p. back skin revealed that catagen completion is
significantly delayed in
ctsl-/ctsl-
mice (Fig. 3D
). About 35% of
ctsl-/ctsl-
hair follicles were still in catagen VI and only 8% of mutant
follicles had completed catagen development (catagen VIII; 49
, 50
). In contrast, in wild-type back skin only 3% of hair
follicles were still in catagen VI and 
39% had already completed
catagen (catagen VIII; Fig. 3D
). At this time the overall
thickness of wild-type skin is clearly reduced, which is characteristic
for entry into telogen (data not shown; 34
, 49
), whereas
mutant skin still maintains the thickness of late catagen stages (data
not shown; 51
).
This delay in hair cycle progression could be caused by
alterations of systemic parameters, such as changes in the expression
level or affinity of hormone-like growth factors, or by immunological
alterations (21)
in mutant mice that may reach the skin
via the circulation. To narrow down the molecular mechanisms underlying
the observed delay in catagen completion, full-thickness skin organ
cultures of day 17 p.p. back skin were initiated. After 5 days in
culture, 88% of hair follicles in wild-type skin biopsies had reached
stages VI or VII of catagen, whereas in
ctsl-/ctsl-
skin organ cultures only 20% had reached stage VI; no follicles in
stage VII were observed, and 50% of follicles were still in catagen
III (Fig. 3E
). These data strongly suggest that the delay in
catagen completion in mutant mice is independent of systemic factors
and most likely caused by the deficiency of CTSL in the skin itself.
Around day 20 p.p. during hair follicle regression, the
bottom of the hair shaft starts to be transformed to a so-called club
hair (catagen V-VII) for retention in the hair canal during subsequent
hair cycles (34
, 46
, 49)
. Histological analyses of the
proximal bulb region of day 20 p.p. hair follicles revealed a
pathological disintegration of the developing club hair in
ctsl-/ctsl-
follicle bulbs (data not shown). The secondary hair germ (epithelial
cell layers in the direct vicinity of the dermal papilla) as well as
the outer root sheath display a marked hyperplasia in mutant mice, and
the hair canal is distended (data not shown). Together, the abnormally
widened hair canal and the malformation of the club hair may cause the
hair shaft to lose its normal mooring in the hair canal, causing it to
slide toward the skin surface with subsequent pathological hair loss at
the end of catagen. The end of catagen of the first hair cycle is
reached around day 21 (34
, 50)
and coincides exactly with
the start of the first macroscopically observed hair loss in
ctsl-/ctsl-
mice (see above).
Ctsl-/ctsl- mice show
premature entry into anagen
At day 28 p.p., 69% of hair follicles in wild-type skin
(Fig. 4A
) are in telogen, which is characterized by minimal skin
thickness and reduced follicle length seen at any time during
synchronized hair follicle cycling (52
, 53)
. At the same
time, all
ctsl-/ctsl-
hair follicles have already prematurely entered anagen (anagen V or VI)
of the first genuine hair cycle (Fig. 4B
). At the transition
from telogen to anagen, equivalent hair follicle densities in wild-type
and ctsl-/ctsl-
skin were observed (data not shown). This premature entry into the
synchronized wave of anagen development may be caused by the hair loss
in late catagen, since experimental depilation also triggers immediate
initiation of anagen and hair shaft removal from the hair canal has
been invoked as one of the endogenous signals of anagen induction
(46
, 53
, 54)
.
|
Ctsl is specifically expressed in epithelial tissues
Ctsl expression in the skin was monitored by RNA in
situ hybridization in adult wild-type mice. High ctsl
transcription levels were identified in epithelial cells of the
epidermis (Fig. 4C
) as well as in the epithelial sheaths of
hair follicles (Fig. 4D
), whereas low ctsl
expression was observed in the underlying mesenchyme. This result
points to a potentially essential function of CTSL in epithelial cells.
Hyperproliferation of basal keratinocytes causes epidermal
thickening
Beyond the pathology of the fur,
ctsl-/ctsl-
mice develop defined abnormalities of their interfollicular skin.
Quantitative histomorphometry of day 14 p.p. back skin revealed a
drastic thickening of both the epidermis (Fig. 5A
, B
, C
) and the dermis
(ctsl-/ctsl-:
299.7±18.4 µm (SE) vs. wild-type: 182.1±14.9
µm; P<0.05). The epidermal thickening develops
postnatally, since no significant differences between mutant and
wild-type mice were detected at day 3 p.p. (Fig. 5F
,
left panel). The epidermal thickening reflects an increase in the
number of epidermal cell layers, especially of the stratum granulosum
and the stratum corneum (data not shown). Hyperplasia, acanthosis, and
hyperkeratosis were also observed in the epidermis of back and tail
skin of 3-month-old mutant mice (epidermis of back skin: 19.3±4.7 µm
(SD) in
ctsl-/ctsl- vs.
11.7±0.9 µm in wild-type mice, P<0.01; epidermis of tail
skin: 33.7±1.3 µm (SD) in
ctsl-/ctsl- vs.
30.4±1.7 µm in wild-type mice, P<0.05).
|
Immunohistochemical analyses of
ctsl-/ctsl-
skin with the markers CD4, CD8, ICAM-I, MHC class II, and histochemical
Giemsa staining of mast cells did not reveal significant differences
between mutant and wild-type animals (data not shown), which makes it
highly unlikely that abnormal inflammatory responses cause the observed
skin thickening. Since the thickness of a stratified epithelium like
the epidermis reflects the established balance between cell division in
the basal layer, terminal differentiation, and loss of cells by cell
death, we then estimated the equilibrium between keratinocyte
proliferation and apoptosis in the hyperplastic epidermis of
ctsl-/ctsl-
mice. Cryosections of day 14 p.p. back skin were
immunohistochemically analyzed with an antibody for the proliferation
marker Ki67, which stains all cells that have entered the cell cycle
(55)
. Apoptosis was identified by an in situ
end-labeling technique, the TUNEL assay (39)
. Whereas the
number of apoptotic cells in the epidermis of
ctsl-/ctsl-
mice appears to be unaltered, quantitative histomorphometry revealed a
threefold elevation of Ki67-positive cells in the basal layer of the
epidermis in
ctsl-/ctsl-
mice when compared to wild-type epidermis (Fig. 5A
, B
, C
). Furthermore, a marked elevation of the
number of Ki67-positive outer root sheath cells was observed in mutant
hair follicles (Fig. 5B
).
To investigate whether systemic effects might have caused this
hyperproliferation-induced skin thickening, full-thickness back skin
biopsies of day 3 p.p.
ctsl-/ctsl-
mice were cultured in vitro for a period of 4 days. At the
start of the culture, no difference in epidermal thickness was detected
between wild-type and
ctsl-/ctsl-
skin (Fig. 5F
, left panel). However, after 4 days in culture
skin of
ctsl-/ctsl-
mice (Fig. 5E
, 5F
) displayed more than
twice the epidermal thickness of wild-type controls (Fig. 5D
, 5F
). Thus development of the epidermal
thickening is most likely an autonomous effect of the CTSL deficiency
in the cells of the skin itself, based on a hyperproliferation of basal
keratinocytes in
ctsl-/ctsl-
mice.
The ctsl gene is mutated in furless mice
The phenotype of
ctsl-/ctsl-
mice is remarkably similar to that of the spontaneous mouse mutant
furless (fs) described by Green in 1954 (56)
.
Recently, the ctsl gene has been mapped to the vicinity of
the fs locus on chromosome 13 (57
58
59)
,
suggesting that fs may be a mutant allele of
ctsl. To test this hypothesis, fs mice (FSB/GnEi,
Jackson Laboratories) and
ctsl-/ctsl-
mice were mated. Eight litters arose from these crosses. All offspring
born from matings of homozygous fs/fs and
ctsl-/ctsl-
mice and about half of the offspring from homozygous x
heterozygous crosses
(ctsl-/ctsl-xfs/+
and ctsl-/+xfs/fs)
developed the furless phenotype (Table 1
). This result indicates that fs and ctsl are
allelic, and raised the possibility that the fs mutant
reflects a loss of function mutation in the ctsl gene.
|
Southern blot analyses revealed no differences between
fs and wild-type genomic DNA at the ctsl locus
(data not shown). This result excluded larger deletions, insertions, or
rearrangements in the ctsl gene of furless mice.
To identify a putative point mutation in the ctsl gene of
fs mice, sequence analyses of the entire open reading frame
of the ctsl gene of fs were performed. A G-to-A
transition was identified in exon 5, resulting in an arginine for
glycine substitution at position 149 of CTSL (Fig. 6A
). This amino acid exchange is located 11 amino acid
residues carboxyl-terminal of the active site cysteine residue of mouse
CTSL. The Gly149 residue is conserved in multiple cysteine proteinases
(Fig. 6B
; 60
). Furthermore, the corresponding
glycine residue in the CTSL-related cysteine proteinase cathepsin K is
also substituted by arginine in a patient with the lysosomal disease
pycnodysostosis, in which cathepsin K is deficient (16)
.
This suggests that the alteration identified may cause an inactivation
of CTSL in fs mice.
|
To test this hypothesis, the G149R mutation was introduced into the
ctsl cDNA (22
, 61)
, and wild-type and
fs CTSL were stably expressed in fibroblasts established
from
ctsl-/ctsl-
embryos (for details, see Materials and Methods). No CTSL activity was
detected with the fluorogenic substrate
Z-Phe-Arg-4-methyl-coumarin-7-amide in fibroblasts transfected with the
mutant cDNA, whereas the CTSL activity was readily detectable in
fibroblasts transfected with the wild-type ctsl cDNA (Fig. 6C
). These data demonstrate that the G149R mutation
abolishes the enzymatic activity of CTSL and thereby most likely
causesthe phenotype of furless mice.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. To identify
specific in vivo functions of the lysosomal cysteine
proteinase CTSL, the gene was disrupted by homologous recombination.
The CTSL-deficient mice presented in this study display a
characteristic periodic loss and regrowth of hair as well as epidermal
hyperplasia, acanthosis, and hyperkeratosis.
Similar phenotypic alterations have been described more than 40 years
ago in mice homozygous for the recessive mutation fs. A
deficiency of CTSL is identified as the molecular defect underlying the
fs phenotype and a missense mutation is detected in the
ctsl gene of fs. The human ctsl gene
has been mapped to chromosome 9 q21–22 (61
, 62)
. This
region of chromosome 9 shows synteny homologies with a region of mouse
chromosome 13, to which the mouse ctsl gene has been
assigned (57
, 59)
. Searching the OMIM (Online Mendelian
Inheritance in Man; NCBI) database, however, did not uncover a human
disorder with conspicuous similarity to fs at chromosomal
location 9 q21–22.
CTSL-like cysteine proteinases have been identified recently in humans
and mice. Human cathepsin L2/V (12
, 63)
is expressed in
thymus and testis, whereas CTSL-like mouse cathepsin J
(64)
is exclusively expressed in placenta. As of today,
neither a murine ortholog of cathepsin L2/V nor a human ortholog of
cathepsin J has been identified. Furthermore, additional CTSL-like
cysteine proteinases may be present in the human and/or murine genome,
e.g., CTSL-like genomic sequences have been identified on human
chromosome 10 q (65)
. Hence, direct conclusions from
the in vivo functions of murine CTSL defined in this study
to those of human CTSL should not be drawn, since it is conceivable
that CTSL-like cysteine proteinases are able to compensate for
different functions in different organisms.
Molecular mechanisms controlling morphogenesis, differentiation, and
growth of the skin and its appendages, i.e., hair follicles, sweat
glands, etc., are only partially understood at present (46
, 47
, 54
, 66)
. Numerous spontaneous mouse mutants with skin and hair
abnormalities have been described (67)
. The molecular
basis of several of these mutants has been identified by targeted
disruption of growth factor and growth factor receptor genes.
Transforming growth factor 
, keratinocyte growth factor/fibroblast
growth factor 7, and fibroblast growth factor 5 (FGF5) -deficient mice
elucidated the molecular defects of mouse mutants waved-1
(wa-1), rough (ro), and
angora (ag), respectively (68
69
70)
.
Moreover, epidermal growth factor receptor was shown to be mutated in
waved-2 (wa-2; 71
). The pathological
phenotype of these mutants is mostly restricted to hairs and hair
follicles, with the interfollicular skin remaining largely unaffected
(68
69
70
71)
. In contrast, CTSL-deficient mice not only
exhibit major alterations of hair follicle morphogenesis and cycling
with periodic loss and regrowth of hair, but also develop a
pathological phenotype of the interfollicular epidermis. To our
knowledge the ctsl knock-out presents the first evidence
that papain-like lysosomal cysteine proteases are specifically involved
in skin homeostasis.
The unifying theme of hyperkeratosis, acanthosis and hyperplasia of the
epidermis on the one hand and the hair follicle alterations on the
other hand in the
ctsl-/ctsl-
mouse mutant are hyperproliferation of basal epidermal and hair
follicle keratinocytes. One characteristic feature of numerous mouse
mutants with hyperplastic skin phenotypes is the presence of dermal
inflammatory cell infiltrates, e.g., the autosomal recessive mutant
flaky skin (fsn) presents with progressive thickening
of the epidermis—most notably of the stratum corneum—and a mixed
inflammatory cell infiltrate in the dermis. Due to these observations,
fsn is considered a model of certain subtypes of the human
cutaneous disease psoriasis (67
, 72)
. In contrast,
ctsl-/ctsl-
mice do not exhibit inflammatory responses in the skin as shown by
immunohistochemistry, excluding the possibility that their skin
pathology is secondary to inflammation. Furthermore, systemic factors
could be excluded as a potential cause of the
ctsl-/ctsl-
phenotype by in vitro reproduction of the skin and hair
pathology in organ culture.
The in vitro reproduction of an altered hair follicle cycling and epidermal thickening in ctsl-/ctsl- skin organ culture indicate that either the proteinase deficiency in basal keratinocytes and hair follicle epithelial cells themselves or in the adjacent dermal fibroblasts cause the observed epidermal hyperproliferation. Proliferation vs. terminal differentiation of keratinocytes and hair follicle epithelial cells in the skin and its appendages can be viewed as competitive. Therefore, the delay in hair follicle morphogenesis and cycling as well as the epidermal hyperplasia may be explained by a later onset of terminal differentiation of epithelial cells due to their hyperproliferation. In the absence of CTSL the balance between proliferation and terminal differentiation seems to be shifted toward proliferation, causing the epidermal and hair follicle hyperplasia.
We have shown recently that
ctsl-/ctsl-
mice are immunocompromised. They exhibit reduced numbers of
CD4+ T cells due to impaired positive selection
in the thymus. CTSL was shown to be essential for proteolytic
degradation of the major histocompatibility complex class II-associated
invariant chain (Ii) in cortical thymic epithelial cells but not in
bone marrow-derived antigen-presenting cells (21)
. The
generation of cathepsin S-deficient mice now indicates that this
proteinase, which also belongs to the family of papain-like cysteine
proteinases, is essential for Ii degradation in bone marrow-derived
cells (18
, 19)
. Epidermis, hair follicle epithelium, and
cortical thymic epithelium are classified as stratified squamous
epithelia (73
, 74)
, since these epithelia express a
specific set of keratin genes. The pathological phenotype of
ctsl-/ctsl-
mice seems to be confined to these three epithelia, which are of a
common ectodermal ontogenic origin (75)
. We have been able
to show that degradation intermediates of Ii accumulate in cortical
thymic epithelial cells (21)
. These data indicate that
degradation of Ii in the endosomal/lysosomal compartment of thymic
cortical epithelial cells is altered due to the deficiency of CTSL. By
analogy, these results may suggest that proteolytic processing of as
yet unknown CTSL in vivo substrates in basal keratinocytes
and hair follicle epithelial cells is diminished considerably and may
not be compensated in these epithelial cells, in contrast to adjacent
connective tissue cells, for example.
Two central control mechanisms have been shown to play an essential
role in developmental processes affecting epidermis homeostasis and
formation of skin appendages (76)
. First, transcription
factors control epidermal gene expression and are of central importance
to coordinate keratinocyte specificity and epidermal differentiation.
Second, a precise balance between proliferation and differentiation is
necessary to maintain sensitivity to environmental changes. A number of
key players in these processes have been identified including growth
factors, their receptors, and extracellular matrix or cell adhesion
molecules (76)
. The data presented in this study give evidence that a ubiquitously expressed lysosomal cysteine proteinase
has essential functions in skin homeostasis and hair formation by
controlling epidermal cell proliferation. CTSL may influence
extracellular matrix turnover in the skin, e.g., by activation of
extracellular matrix degrading metalloproteinases, which in turn may
alter proliferation of epidermal and hair follicle epithelial cells. On
the other hand, absence of CTSL may more directly alter auto- or
paracrine mechanisms; it may be involved in proteolytic processing of
paracrine growth factors or their respective receptors modulating
proliferation rates of epidermal and hair follicle epithelial cells.
Furthermore, the in vivo functions of the CTSL inhibitors
stefins, cystatins, and kininogens have to be considered: alterations
in the balance between these inhibitors and cysteine proteinases may
contribute to tumor progression (77)
. Investigations aimed
at differentiation between these possibilities and hence identification
of CTSL in vivo substrates are in progress.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Current address: Department of Dermatology, Boston University, Boston MA 02118, USA.
3 Current address: Hoffmann-La Roche, Basel, Switzerland.
4 Current address: Hautklinik, UKE, Universität Hamburg, Hamburg, Germany.
Received for publication January 13, 2000.
Revision received April 7, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
