ARTICLE
Cell death is an essential strategy of dynamic balance in the living
system. Homeostasis is maintained by the balance between cell proliferation
and cell death. There are two distinct forms of cell death, called necrosis
and apoptosis. The term apoptosis was introduced in 1972 by Kerr et
al. to describe a distinct form of cell death with characteristic
morphologic features that differs from necrosis [1]. Necrosis is a passive
and pathological form of cell death resulting from acute cellular injury,
in which the cells tend to swell and lyse. Apoptosis, by contrast, represents
an active and physiological process, by which the nucleus and cytoplasm
shrink and often fragment: phagocytosis by macrophages and other cells
is the final event in the apoptotic process.
The term programmed cell death is derived from developmental biology.
For example, programmed cell death is responsible for the elimination
of larval tissues during amphibian and insect metamorphosis, as well as
for the elimination of tissue between digits during the formation of fingers
and toes. It is now clear that apoptosis is the mechanism of programmed
cell death, which is genetically programmed to occur during development
and differentiation, mediating active changes in organ structure and function
[2].
Apoptosis is the major mechanism by which homeostasis of a number of
physiological systems in the body can be regulated. Furthermore, recent
studies have suggested that the failure of cells to undergo apoptotic
cell death might be involved in the pathogenesis of a wide variety of
human diseases, including cancer, autoimmune diseases, and viral infections
[3]. There is accumulating evidence in the skin that apoptosis occurs
not only in the pathological conditions of the skin, but is a ubiquitous
process that is important in regulating epidermal growth [4-6]. We herein
review the basic concept of apoptosis and its relevance to skin biology.
Morphology
Morphologically, necrosis is typified by cellular and organella swelling,
blebbing, vacuolization, and lysis. In contrast, the characteristic morphological
feature of apoptosis is cell shrinkage. The most easily recognized features
are changes that occur within the nucleus, in which the chromatin becomes
pyknotic and packed into smooth masses applied against the nuclear membrane.
The nucleus also breaks up into small pieces (karyorhexis) and the cells
emit processes that often contain pyknotic nuclear fragments. Finally,
the cell itself shrinks and breaks up into membrane-enclosed fragments
called apoptotic bodies [7, 8]. Within tissues, apoptotic cells or apoptotic
bodies are recognized and rapidly phagocytosed by neighboring cells including
epidermal keratinocytes or macrophages, and then degraded within their
lysosomes, resulting in effective removal of apoptotic bodies. Removal
occurs before lysis, which prevents the release of potentially toxic and
immunogenic intracellular contents from the apoptotic cells into the surrounding
tissue: therefore, cell death by apoptosis does not invoke an inflammatory
response although recent studies have suggested that phagocytosis of apoptotic
cells also stimulates the macrophages to express an antiinflammatory or
suppressive phenotype; whereas necrosis is associated with loss of cell
membrane integrity, resulting in leakage of cytoplasmic contents and induction
of an inflammatory response. Apoptosis usually affects scattered individual
cells rather than cell groups or a whole tissue, unlike necrosis. The
early cellular events in apoptosis can run their course very fast, even
in a minute. The duration of apoptosis from initial cell shrinkage through
to removal of apoptotic bodies requires as little as 1-3 hrs in lymphocytes,
but up to 48-72 hrs in epidermal keratinocytes [9].
Biochemistry
The apoptotic process is accompanied by major changes in cellular biochemistry
involving the activation of catabolic enzymes (Fig.
1). The first demonstrated biochemical hallmark of apoptosis was
intranucleosomal DNA cleavage of genomic DNA, which is generally referred
to as oligonucleosomal DNA fragmentation. The enzyme responsible for the
DNA-degradation is a putative Ca2+/Ma2+-dependent
endonuclease that fragments the genome into approximately 200 base pairs
multimers [2]. Therefore, DNA from tissue cultured cells or lymphocytes
undergoing apoptosis displays a characteristic series of bands (so called
nucleosomal ladder) after agarose gel electrophoresis [10]. Recent studies
have indicated that intranucleosomal DNA cleavage dose not occur in all
cell types and some cell types undergo apoptosis without endonuclease
activation. Thus, endonuclease may not be central to the apoptotic process
and rather fulfills the function of cleaning up after cell death.
Recent attention has been focused on the possibility that intracellular
proteases might play a critical role in the initiation of apoptosis. Numerous
studies using both molecular cloning approaches and in vitro systems
have identified a class of highly specific cellular proteases, named caspases,
that appear to be have important roles in apoptotic execution [11, 12].
They are related to mammalian interleukine-1ß converting enzyme
(ICE) and to cell death abnormal (CED)-3, the production of a gene that
is absolutely necessary for appropriate suicide in the Caenorhabditis
elegance (C. elegance). So far, three different caspase family members
have been identified. All caspases are synthesized as proenzymes which
are proteolytically processed to form active heterodimeric enzymes. Despite
a notable similarity in structure, different members of the caspase family
possess distinctive activation requirements, substrate specificities,
and inhibitory profiles. Some caspases are endowed with the capacity of
autoactivation. Moreover, caspases can activate others following an ordered
sequence. For example, caspase-8 is responsible for the activation of
caspase-1, which then activates caspase-3 during Fas-induced apoptosis.
Gene knockout experiments have demonstrated an essential role for caspase-1
in Fas-induced apoptosis. Overall, caspases are considered to be executors
of a common cell-death pathway that is triggered in response to a variety
of stimuli.
Molecular regulation of apoptosis
Many of the factors that influence commitment or cellular susceptibility
to apoptosis are involved directly in the reception and transduction of
the apoptotic signal. According to recent understanding, the process of
apoptosis can be subdivided into at least three different phases [13].
During the initiation phase, apoptosis can be affected by a variety of
extrinsic and intrinsic signals including Fas/tumor necrosis factor receptor
(TNFR), cytokines, calcium, hormones, growth factors, radiotherapy, UV,
cytotoxic drugs, and viruses. The factors that induce or inhibit apoptosis
are listed in Table I.
These triggers can activate inducers of apoptosis controlled by several
regulators, then the decision to die is defined, and finally leads to
the activation of effectors of apoptosis. It is assumed that the execution
phase of apoptosis defines the decision to die at the point of no return
of the apoptotic cascade. During the execution phase the central executioner
of apoptosis is activated. It is at this level that the different private
pathways converge into a few common pathways and that cellular processes
still have a decisive regulatory function [13]. Effectors cause the cellular
and biochemical changes such as DNA fragmentation seen in apoptosis. This
degradation phase is similar in all cell types. As mentioned above, it
is characterized by the action of catabolic enzymes, mostly specific protease
(caspases) within the limits of a near-to-intact plasma membrane. Thus,
although diverse apoptotic stimuli can provoke cell death by an unknown
mechanism, it is controlled under the Bcl-2 family of dimerizing proteins,
whereas members of TNFR family bypass regulation by Bcl-2 family members
by directly activating caspases. Considerable evidence exists that an
increase in reactive oxygen species constitutes an intracellular signal
that can lead to apoptosis. Apoptosis can be induced in a number of cell
systems by reactive oxygen species. The decrease in antioxidant enzymes
could lead to an increase in cellular reactive oxygen species responsible
for signalling apoptosis. Antioxidants inhibit apoptosis induced by a
variety of stimuli [13a].
Apoptosis is an active, genetically controlled process. In this regard,
a number of evolutionarily conserved genes regulate a final common cell
death pathway that is conserved from insects to mammals. The best evidence
that apoptosis is of genetic origin is derived from studies of programmed
cell death during the development of C. elegans, in which programmed cell
death (PCD) has been divided into four distinct stages, each controlled
by a specific set of genes identified as ced [14]. During embryonic
and larval development, 131 of 1,090 cells in C. elegans are eliminated
in a well-characterized, spatially and temporally invariant program. For
example, two of the genes, ced-3 and ced-4 are required
for apoptosis to occur, and ced-9, homologous to Bcl-2, is required
for suppressing apoptosis.
As mentioned above, many molecules or genes involved in the regulation
or induction of apoptosis have been identified. We here describe Fas and
Bcl-2, which have been most investigated in the field of skin research.
Fas
Fas (also known APO-1 or CD95) is a 45-kDa glycosylated type 1 transmembrane
receptor that is a member of TNF/nerve growth factor (NGF) receptor superfamily.
Superfamily members include TNFR1, TNFR2, the low-affinity NGFR, CD27,
CD30, CD40, OX40, and 4-1BB [12, 13]. Fas L, the ligand for Fas, is a
40-kDa glycosylated type 2 transmembrane protein that belongs to the TNF
family. Current model indicates that binding of Fas L to Fas at the cell
surface causes the association of FADD (Fas-associated protein with death
domain or MORT1) and other proteins to the Fas cytoplasmic tail (death
domain), via a homotypic death domain-death domain interaction.
For the TNFR pathway, the TNFR associates with TRADTD (TNFR1-associated
death domain), which in turn recruits FADD to the cell membrane. Caspase-8
(originally called FLICE) is then recruited, which in turn may induce
self-activation of the protease domain. The activated ICE-like proteases
can cleave a number nucleoprotein substrates, resulting in DNA fragmentation.
The precise roles and relative importance of the various ICE family members
have been difficult to define because of a functional redundancy among
ICE-like proteases.
Fas can be expressed on a variety of both lymphoid and nonlymphoid cells,
including liver, ovary, heart, lung, kidney, and skin. In contrast, the
expression of Fas L is more limited than the expression of Fas. Although
Fas L expression was initially confined to activated T cells, recent studies
indicated that Fas L is expressed widely in adult tissues, in particular
neutrophils and activated lymphocytes, in immune privileged tissues such
as the eye and testis, and in certain tumor cells. Moreover, a variety
of cell types can express Fas L in response to different stimulatory conditions
such as HIV-infected macrophages. Resting T cells do not constitutively
express Fas L, whereas activated T cells express Fas L. Thus, when a Fas-expressing
activated T cell comes in contact with another T cell expressing Fas L
on its surface, it undergoes apoptosis. This Fas-mediated apoptosis provides
a mechanism that eliminates the expanded lymphoid populations that are
no longer needed. Thus, this "fratricide" of activated T cells is thought
to be involved in the regulation of the size of the pool of activated
T cells. Cytotoxic T cells also use Fas L to kill Fas-expressing target
cells, which is a critical function of the immune system. Nearby Fas-expressing
lymphocytes and nonlymphoid cells also undergo Fas-mediated apoptosis
as a consequence of Fas L binding. Fas-Fas L interactions also play a
dominant role in preventing potentially harmful immune reactions in immunologically
"privileged" sites like eye and testis [15]. Such immunologically privileged
sites constitutively express Fas L, which causes Fas-mediated apoptosis
of infiltrating Fas-expressing T cells, thus protecting these tissues
from an immune attack. Expression of Fas L on certain tumor cells also
induces Fas-mediated apoptosis of tumor-specific cytotoxic T cells expressing
Fas, thus providing malignant cells with resistance to tumor immunity.
The important role of Fas-mediated cell death in autoimmunity has been
convincingly demonstrated in studies of mice with lpr (lymphoproliferation)
and gld (generalized lymphoproliferative disease) mutations, which
are complementary mutations of Fas and Fas L genes, respectively. Both
the lpr and the gld mice are unable to mediate Fas/Fas L-dependent
apoptosis, leading to the accumulation of T cells with an unusual surface
phenotype and a variety of autoimmune reactions and the eventual development
of an lymphoproliferative autoimmune disease resembling lupus erythematous
in these mice [16]. These observations suggest that Fas/Fas L interactions
control the peripheral lymphocyte life span and thereby participate in
peripheral elimination of autoreactive lymphocytes. Nevertheless, Fas-mediated
apoptosis is not the only mechanism of activation-induced cell death of
T cells: activation-induced cell death of CD8+ T cells, and
perhaps even of some CD4+ T cells, may be caused by TNF-TNFR
interactions and independent of Fas.
A soluble splice variant of Fas (sFas) has been identified in human
serum in various conditions including autoimmune diseases. Although sFas
has been initially shown to inhibit apoptosis induction in vitro
and thought to be generated by alternative splicing rather than proteolytic
cleavage, recent studies have suggested that sFas release may be correlated
with the amount of tissue damage. Fas L also exist in a soluble form in
addition to a membrane-bound form. Cells expressing Fas L use metalloproteinases
to cleave Fas L from their membrane surfaces, thereby generating soluble
Fas L and potentially attenuating their own capacity to deliver death
signals to Fas-expressing cells.
Specific viral infections have been shown to lead to increased Fas and/or
Fas L expression and increased sensitivity to Fas/Fas L-dependent apoptosis
[17]. These changes can result in extensive cell death and tissue damage.
For example, liver damage due to hepatitis B and C viruses, and in part
T lymphocyte cell death during HIV infection are among them. Viral clearance
is probably achieved by the cooperation of at least two mechanisms. First,
viral antigen-specific cytotoxic T cells (CTL) recognize and deliver apoptotic
signal mediated by both Fas L and perforin to their target infected cells.
Because Fas L on the CTL can dock with Fas on healthy cells in the vicinity
of infected cells, it can also trigger their suicide. This bystander effect
may explain why hepatitis viruses can cause extensive liver damage despite
relatively few liver cells infected with the viruses. Thus, the propensity
of CTL to destroy bystander cells depends partly on the relative efficiency
with which CTL are brought into proximity to target cells by their receptors.
Second, they also produce IFN-gamma and TNF-alpha, which have been shown
to abolish viral gene expression and its replication, and thereby curing
the infection. On the other hand, viruses have a variety of strategies
to blunt the antiviral immune responses and inhibition of apoptosis is
critical to efficient replication and establishment of latency in many
pathogenic viruses: certain viruses have evolved ways to resist Fas-mediated
cell death and thus promote their survival. For instance, the Epstein
Bar (EB) virus encodes homologs of mammalian Bcl-2, and EBV LMP-1 interacts
with members of TRAF (TNFR-associated factor) family, inhibits apoptosis
of infected B cells and induces the infected cell to increase its own
expression of Bcl-2. Adenovirus-encoded proteins can promote persistent
adenovirus infections by clearing Fas from the cell surface of the infected
cells and reducing killing by CTL that express Fas L. Other viruses can
elaborate a protein that prevents ICE-like proteases from carrying out
the apoptotic program. In addition, Fas/Fas L-mediated apoptosis has been
shown to provide a mechanism that enables the clearance of greatly increased
populations of CTL which are found during viral infections. At the end
of an immune response against viral infections, activated T cells downregulate
Bcl-2 and Bcl-xL expression (see below) and are destined to undergo apoptosis.
This may protect against overstimulation of the immune system.
Bcl-2 family
The Bcl-2 is a proto-oncogene that was originally found as a result
of its location at the site of a translocation between chromosomes 14
and 18 and is present in most human follicular lymphomas [18]. Although
initially viewed as an oncogene, Bcl-2 has little mitogenic effect. Instead,
its oncogenic potential has been attributed to its ability to inhibit
apoptosis. Bcl-2 prolongs the survival of cells in the absence of required
growth factors by blocking apoptosis, even in the presence of a variety
of stimuli such as chemotherapeutic agents, irradiation, TNF, heat shock,
and transfection with p53 or c-myc. Furthermore, the introduction
of genes that inhibit Bcl-2 can induce apoptosis in a wide variety of
tumor cell types, which suggests that many tumors continually rely on
Bcl-2 to prevent cell death. IL-2 prevents activated T cell apoptosis
by upregulating expression of Bcl-2. Removal of IL-2 from activated T
cells in vitro leads to reduced Bcl-2 expression and apoptosis.
Alternatively, the overexpression of Bcl-2 increases the viability of
IL-2 dependent cells, upon IL-2 withdrawal. The observation can be extented
to other cytokine-dependent cells, such as IL-3, IL-4, IL-6, and GM-CSF.
Bcl-2 can also protect T cells from a variety of apoptotic signals, including
glucocorticoids, gamma-irradiation, phorbol esters, and ionomycin. Bcl-2
has been shown to suppress Fas-induced apoptosis in some cell types but
not in others.
Recently a number of Bcl-2 family members have been identified. Bcl-2,
Bcl-xL, Bcl-w and Mcl-1 inhibit apoptosis, whereas others, such as Bax,
Bik, Bak, Bad, and Bcl-xs activate apoptosis. Because many of these proteins
are coexpressed in the same cells, the ratio of antiapoptotic (e.g.
Bcl-2) vs pro-apoptotic protein (e.g. Bax) levels has been
suggested to determine the inherent susceptibility of a given cell to
respond to apoptotic signal [12, 19]. Although most studies on the function
of these proteins have relied on simple overexpression, the levels of
expression of these proteins do not always predict the ability of a given
cell to resist apoptotic stimuli. Nevertheless, the control of apoptosis
in T cells by Bcl-xL appears to be mediated by a simple increase or decrease
in expression. For example, at the end of an immune response, the majority
of the expanded T cell population is removed by undergoing apoptosis,
which results from downregulation of Bcl-2 and Bcl-xL expression. Recent
studies have indicated an additional role of Bcl-2 in regulation of cell
cycle progression: Bcl-2 deficient T cells demonstrate accelerated cell
cycle progression and increased apoptosis following activation; and Bcl-2
overexpressing peripheral T cells exhibit delayed entry to S phase and
diminished IL-2 production upon activation. Bcl-2-xL has a similar inhibitory
effect on cell cycle entry in activated T cells.
Apoptosis in the skin
In the skin, cells dying by apoptosis have been found in a wide variety
of conditions, such as inflammatory dermatoses and skin tumors [4-6] (Table
II). Evidence is accumulating that apoptosis plays an important
role not only in the pathogenesis of skin diseases, but is also involved
in the homeostatic mechanisms in healthy skin. In this respect, terminal
differentiation of keratinocytes is thought to be a special form of apoptosis,
because there are similarities between terminally differentiating keratinocytes
and apototic cells; for example, granular keratinocytes show signs of
endonuclease activation and DNA fragmentation [20]. Thus, it is likely
that the proliferation of keratinocytes is regulated by apoptotic cell
death to maintain a constant thickness of the epidermis.
The growing literature on the expression of Fas, Fas L, and the Bcl-2
family proteins in the skin during a variety of disease conditions provides
clues about the role of apoptosis in regulating homeostasis in the skin.
The results of these studies are summarized in Table
III. In interpreting these results, one must appreciate that the
mere presence of Fas and Fas L is only a first determinant of apoptosis
and that the susceptibility of a given cells to die in response to cell
death signals such as Fas/Fas L binding and cytokine deprivation can be
modified by many different proteins such as the Bcl-2 family proteins.
Keratinocyte
At light microscopic level, apoptotic keratinocytes are characterized
by a condensed and basophilic nucleus, and eosinophilic homogenization
of the cytoplasm which sometimes contains irregular basophilic materials.
Such individually dying cells are traditionnally referred to by several
histological terms, which include dyskeratotic cells, Civatte bodies,
colloid bodies, dark cells, satellite cell necrosis, or sunburn cells.
These cells represent distinctive subtypes of apoptotic keratinocytes.
Apoptotic keratinocytes are most frequently seen in association with the
lichenoid tissue reaction which is a histological pattern found in a heterogeneous
group of dermatoses that have in common basal keratinocyte damage and/or
vacuolar change intimately associated with the infiltrate of T cells [4,
5, 21-23]. Included in the lichenoid tissue reaction are lichen planus,
lupus erythematosus, erythema multiforme, fixed drug eruption, and graft-vs-host
disease. In acute experimental GVHD, intraepithelial apoptosis is the
predominant form of cellular injury, which correlates with the onset of
lymphocyte infiltration [24]: nevertheless, it has also been shown that
some of the epidermal damage can be observed prior to histological evidence
of lymphocytic infiltration, indicating that keratinocyte apoptosis may
occur even in the absence of direct lymphocyte-target cell interactions,
although most of apoptotic keratinocytes in lichenoid tissue reactions
are postulated to result from cell-mediated immune reactions against the
epidermis [25]. There are several mechanisms by which keratinocytes undergo
apoptosis [12, 26]. First, activated cytotoxic T cells express Fas L,
which binds to Fas expressed on keratinocytes and results in apoptosis.
Indeed, Fas antigen is expressed on keratinocytes in the lesional epidermis
in lichenoid skin diseases, and anti-Fas antibody can trigger apoptosis
in the IFN-gamma-treated cultured keratinocytes [27]. Second, apoptosis
can also be induced via the release of effector cell granules,
which include perforins and multiple serine proteases, called granzyme.
Granzyme B can activate some of the caspase family members by proteolysis.
Although Fas/Fas L and perforin/granzyme can independently trigger the
cell death program, the processes leading to apoptosis are similar in
both pathways (Fig. 2).
CD8+ CTL and NK cells use both the perforin/granzyme and Fas/Fas
L pathways, whereas Th1-type CD4+ T cells preferentially use
the Fas/Fas L pathway. In addition, because TNF can induce apoptosis by
ICE-like protease-dependent and -independent pathways, it is logical that
TNF released from mast cells and activated T cells is sufficient stimulus
for elicitation of keratinocyte apoptosis. Recently, human epidermal keratinocytes
have been shown to have the ability to produce granzyme B, perforin, and
Fas L, which can be used to protect the epidermis from immune-mediated
damage or invading pathogens [28]. This is a significant departure from
current dogma, which views the keratinocytes merely as a victim of the
immune-mediated damage.
The data appear to suggest that Fas L-bearing keratinocytes during the
lichenoid tissue reaction could induce Fas-mediated death upon neighboring
Fas-bearing T cells and contribute to the resolution of injurious immune
reactions mediated by the T cells. This mechanism is analogous to the
recently established role of Fas L in mediating immune privilege in the
eye and testis and immune escape in malignancies. If this is the case,
the lichenoid tissue reaction could be viewed either as a tissue-sparing
strategy by keratinocytes, contributing to the elimination of potentially
harmful autoaggressive T cells, or as a self-protection mechanism by immune
cells, contributing to the elimination of abnormal keratinocytes. The
relative balance of autoaggressive T cells and keratinocytes would determine
the outcome of the lichenoid skin disease (Fig.
3). Toxic epidermal necrolysis is likely to represent the most
devastating extreme.
Psoriasis can be viewed as a hyperproliferative disorder of keratinocytes
mediated by T cells. In contrast to lichenoid skin diseases, there is
no microscopic evidence for the presence of apoptotic keratinocytes in
psoriasis, despite Fas expression on the lesional keratinocytes. One possible
explanation is that Bcl-xL, shown to block apoptosis, is overexpressed
on keratinocytes within lesional plaques [29]. In this respect, increased
epidermal thickness in psoriasis can be explained by abnormality in the
apoptotic cell death pathway. This indicates that the difference in the
behavior of keratinocytes between lichenoid skin diseases and psoriasis
can be in part caused by changes of the keratinocyte expression pattern
of pro- and anti-apoptotic genes. Alternatively, psoriatic keratinocytes
may be endowed with the superior ability to produce granzyme B and Fas
L, thereby being extremely resistant to killing by Fas-bearing T cells.
Thus, the susceptibility of the epidermis to undergo immune-mediated damage
may be dependent on its ability to express Fas L and the density of its
Fas membrane expression.
Melanocytes and melanocytic
tumors
Melanocytes are neural crest-derived cells. Neural tissues highly express
Bcl-2. In normal skin, melanocytes, like neural tissue, constitutively
express Bcl-2 [30]. Normal melanocytes are long-lived post-mitotic cells
that do not produce any mitogens that stimulate their own growth. Expression
of Bcl-2 on melanocytes may therefore be needed to escape from apoptosis.
Likewise, expression of Bcl-2 can be commonly observed in melanocyte nevi
and malignant melanoma [30, 31]. However, no differences in Bcl-2 expression
were found among various subtypes of benign and malignant melanocytic
proliferation [30], suggesting that Bcl-2 can not be considered as a marker
of malignancy in melanocytic neoplasma, as has been demonstrated from
lymphomas. The expression of Bcl-2 with melanocytic nevi tends to diminish
when neuroid changes are present: this finding may help explain the clinical
life cycle of melanocytic nevi [32]. Melanocytes and melanoma cell lines
are relatively resistant to the induction of apoptosis induced by UV radiation
or cytokines. Although Bcl-2 expression by melanoma cells could contribute
in part to the resistance of melanomas to chemotherapy and radiation,
additional factors yet to be defined would be required for the resistance
because no differences in Bcl-2 expression were found between melanocytic
nevi and malignant melanoma. Thus, the functional role of Bcl-2 in the
resistance remains inconclusive. Nevertheless, Jansen et al. demonstrated
that Bcl-2 antisense oligonucleotide treatment improves the chemosensitivity
of human melanomas grown in severely combined immunodeficiency mice [33].
Thus, reduction of Bcl-2 expression in melanomas may be a novel and rational
approach to improve chemosensitivity and treatment outcome. Malignant
melanomas also express Fas L, but not Fas [34]. No Fas L expression is
found in normal melanocytes, indicating that the upregulation of Fas L
occurs during tumorgenesis. Fas L expression may be a more general strategy
used by tumor cells to escape immune rejection because Fas L expressing
melanoma cells can kill Fas-bearing activated T lymphocytes.
Skin tumors
For a long time the dysregulation of growth which leads to neoplasm
was explained largely in terms of increased cell proliferation. The control
of cell numbers in both normal and neoplastic conditions depends on factors
influencing the balance between cell growth and death. It has become clear
that the proliferation of neoplasm seems to be partially controlled by
apoptosis. Apoptosis can be found in a wide variety of both benign and
malignant skin tumors, including basal cell carcinoma (BCC), squamous
cell carcinoma (SCC), pilomatricoma, keratocanthoma, and Merkel cell tumor.
BCC is known to be typically slow-growing tumor, often taking months to
years to reach significant proportions, in spite of the numerous mitotic
figures histopathologically. Since Kerr et al. firstly proposed
that clinically slow growth rate in BCC may be due to prominent apoptotic
cell death [35], several explanations for the clinical behaviors have
been described. The histological examination reveals that apoptotic cells
outnumber mitotic cells in BCC. The factors responsible for the occurrence
of apoptosis in BCC are now controversial: while previous reports demonstrated
that Bcl-2 is highly expressed in BCC [36-38], others reported that spontaneous
apoptosis decreases in association with an increase in Bcl-2 expression
[39], a finding in contrast to a high rate of apoptotic cell death in
BCC. Bcl-2 expression may be insufficient to protect BCC cells against
apoptotic stimuli. Verhaegh et al. speculated that BCC may be a
neoplastic transformation resulting from extended cell survival, rather
than from proliferative pathways. In contrast to BCC, in SCC Bcl-2 is
only detectable in the basal cells [36-38] and a significantly higher
number of apoptotic cells can be observed than in BCC. Contrary to BCC,
increased proliferation rather than decreased cell death seems to contribute
to tumorgenesis of SCC. Lymphocyte-mediated apoptosis could be a mechanism
responsible for the spontaneous regression of skin tumors. Despite numerous
lymphocytes infiltrating in the vicinity of tumors, spontaneous regression
can be occasionally seen in many skin tumors. BCC cells express Fas-L,
but not Fas, which may allow tumor expansion by killing Fas-bearing activated
T cells. Intralesional injection of IFN-alpha has been found to be highly
effective in inducing BCC regression. In the IFN-alpha-treated patients,
BCC cells express not only Fas L but also Fas, whereas the peritumor infiltrate
that mainly consists of CD4+ T cells predominantly expresses
Fas but not Fas L. Thus, it is possible than Fas-Fas L interactions within
the tumors rather than those between BCC cells and T cells might act as
an alternative mechanism for IFN-alpha-induced regression of BCC [40].
Apoptosis can be observed in pilomatricoma and regressing kerathoacanthoma.
In pilomatricoma, Bcl-2 is expressed on basophilic cells, but not on transitional
cells. In keratoacanthoma, Bcl-2 is expressed on the basal layer of the
neoplasm in its proliferative stage, whereas the basal cells rarely express
Bcl-2 in its involuting stage. These findings might represent the mechanism
responsible for the biological behavior of these tumors.
UV irradiation
A hallmark event of UV exposure is the occurrence of sun burn cells
in the epidermis. These cells have been considered as keratinocytes undergoing
apoptosis. Recent studies have proved that UV irradiated keratinocytes
display DNA fragmentation characteristic of apoptosis. Although little
is known regarding the mechanisms that regulate this process, Fas system
or tumor suppressor gene p53 is shown to be involved in UV-irradiated
apoptosis of keratinocytes [41]. UV irradiation induces both Fas and Fas
L expression on keratinocytes at mRNA and protein levels [42]. Addition
of neutralizing Fas L antibodies inhibits UV-induced apoptosis of IFN-gamma
treated keratinocytes. These findings suggest that the Fas system contributes
to keratinocytes apoptosis in UV-irradiated skin. UV light may act to
induce Fas L in skin tumors such as BCC thereby enabling them to escape
from an immune attack by CTL, whereas UV-induced Fas L on the psoriatic
keratinocytes may act to kill intraepidermal T cells, thereby improving
the lesions [43]. p53 is another major regulatory factor contributing
to UV-induced apoptosis in keratinocytes. After in vitro UV irradiation,
p53 protein levels were noted to increase prior to the induction of apoptosis
in human keratinocytes [44]. In mice exhibiting different p53 genotype,
a correlation has been found between the decrease in numbers of sunburn
cells and the decrease in copy numbers of the p53 genes. Furthermore,
p53-knockout mice do not develop apoptosis in the epidermis after UV irradiation.
Thus, p53 plays a role in the induction of UV irradiated apoptosis of
keratinocytes. Although the functional role of sunburn cells remains obscure,
UV damaged keratinocytes may die as sun burn cells to escape the risk
of becoming UV induced skin cancer.
Hair cycle and hair loss
The hair follicles undergo a cycle of growing, regressing, and resting
phases (anagen, catagen, telogen, respectively). The mechanism responsible
for the involution of the hair follicle during catagen has not been satisfactorily
explained. In normal catagen, apoptotic cells are scattered in the outer
root sheath and are engulfed quickly, initially not by macrophages but
by nearby epithelial cells [4, 5]. Lidner et al. showed biological
evidence that catagen is associated with endonuclease activation and that
physiological and pathological catagen is characterized by an up-regulation
of ICE expression and an apparent inversion of the Bcl-2/Bax ratio in
epithelial follicle regions that undergo involution during catagen [45].
Thus, apoptosis is a central element in the regulation of hair follicle
regression (catagen).
Because alopecia is frequently induced by chemotherapeutic agents which
are much more strongly tied to induction of apoptosis than had been thought
[46], it may be that the hair follicle is one of those tissues extremely
susceptible to various apoptotic stimuli.
Other diseases
Keloids are collagenous lesions acquired as a result of abnormal wound
healing. Appleton et al. demonstrated that proliferation, apoptosis,
and necrosis occur simultaneously in keloids and that these processes
are distinctly compartmentalized. As keloid matures, apoptosis and necrosis
result in selective removal of certain cellular populations resulting
in the characteristic avascular fibrotic collagenous lesion [47]. Scleroderma
is an autoimmune disorder characterized by degenerative fibrotic skin
lesions. Sgonc et al. demonstrated that endothelial cells are clearly
the first cells to undergo apoptosis in the skin of avian scleroderma
and that apoptotic endothelial cells can only be detected in early inflammatory
disease stages of human scleroderma [48].
There are two distinct forms of cell death,
called necrosis and apoptosis.
Necrosis is a passive and pathological form of cell death.
Apoptosis, by contrast, represents an active and physiological process.
Programmed cell death is responsible for the elimination of larval
tissues during amphibian and insect metamorphosis, as well as for the
elimination of tissue between digits during the formation of fingers and
toes.
Programmed cell death is genetically programmed.
Failure of cells to undergo apoptotic cell death might be involved
in the pathogenesis of a wide variety of human diseases.
Necrosis is typified by cellular and organella swelling, blebbing,
vacuolization, and lysis.
The characteristic morphological feature of apoptosis is cell shrinkage.
Apoptotic cells or apoptotic bodies are recognized and rapidly phagocytosed
by neighboring cells including epidermal keratinocytes or macrophages.
Cell death by apoptosis does not invoke an inflammatory response.
Necrosis is associated with loss of cell membrane integrity, resulting
in leakage of cytoplasmic contents and induction of an inflammatory response.
Apoptosis usually affects scattered individual cells rather than cell
groups or a whole tissue, unlike necrosis.
Biochemical hallmark of apoptosis was intranucleosomal DNA cleavage
of genomic DNA.
Intracellular proteases might play a critical role in the initiation
of apoptosis.
Some cell types undergo apoptosis without endonuclease activation.
Caspases appear to have important roles in apoptotic execution.
Caspase-8 is responsible for the activation of caspase-1, which then
activates caspase-3 during Fas-induced apoptosis.
The process of apoptosis can be subdivided into at least three different
phases.
Initiation phase, a variety of extrinsic and intrinsic signals including
Fas/tumor necrosis factor receptor (TNFR), cytokines, calcium, hormones,
growth factors, radiotherapy, UV, cytotoxic drugs, and viruses.
Execution phase of apoptosis defines the decision to die at the point
of no return.
Degradation phase: the apoptotic mechanism is controlled under the
Bcl-2 family of dimerizing proteins, whereas members of TNFR family bypass
regulation by Bcl-2 family members by directly activating caspases.
Decrease in antioxidant enzymes could lead to an increase in cellular
reactive oxygen species responsible for signalling apoptosis.
Evolutionarily conserved genes regulate a final common cell death
pathway that is conserved from insects to mammals.
Current model indicates that binding of Fas L to Fas at the cell
surface causes the association of FADD (Fas-associated protein with death
domain or MORT1).
Resting T cells do not constitutively express Fas L, whereas activated
T cells express Fas L. Thus, when a Fas-expressing activated T cell comes
in contact with another T cell expressing Fas L on its surface, it undergoes
apoptosis.
Fas-Fas L interactions also play a dominant role in preventing potentially
harmful immune reactions in immunologically "privileged" sites like the
eye and testis.
Fas L on certain tumor cells also induces Fas-mediated apoptosis
of tumor-specific cytotoxic T cells expressing Fas, thus providing malignant
cells with resistance to tumor immunity.
Activation-induced cell death of CD8+ T cells, anne perhaps
even of some CD4+ T cells, may be caused by TNF-TNFR interactions
and independent of Fas.
sFas release may be correlated with the amount of tissue damage.
Viral infections have been shown to lead to increased Fas and/or
Fas L expression and increased sensitivity to Fas/Fas L-dependent apoptosis.
Because Fas L on the CTL can dock with Fas on healthy cells in the
vicinity of infected cells, it can also trigger their suicide.
Certain viruses have evolved ways to resist Fas-mediated cell death
and thus promote their survival.
At the end of an immune response against viral infections, activated
T cells downregulate Bcl-2 and Bcl-xL expression and are destined to undergo
apoptosis. This may protect against overstimulation of the immune system.
Bcl-2 prolongs the survival of cells in the absence of required growth
factors by blocking apoptosis.
IL-2 prevents activated T cell apoptosis by upregulating expression
of Bcl-2.
Bcl-2 can also protect T cells from a variety of apoptotic signals,
including glucocorticoids, gamma-irradiation, phorbol esters, and ionomycin.
Bcl-2 family members Bcl-2, Bcl-xL, Bcl-w and Mcl-1 inhibit apoptosis,
whereas others, such as Bax, Bik, Bak, Bad, and Bcl-xs activate apoptosis.
The ratio of antiapoptotic (e.g. Bcl-2) vs pro-apoptotic
protein (e.g. Bax) levels determine the inherent susceptibility
to apoptotic signal.
Apoptosis plays an important role not only in the pathogenesis of
skin diseases, but is also involved in the homeostatic mechanisms in healthy
skin.
Terminal differentiation of keratinocytes is thought to be a special
form of apoptosis, because there are similarities between terminally differentiating
keratinocytes and apototic cells.
Dyskeratotic cells, Civatte bodies, colloid bodies, dark cells, satellite
cell necrosis, or sunburn cells represent distinctive subtypes of apoptotic
keratinocytes which are most frequently seen in association with the lichenoid
tissue reaction.
Keratinocyte apoptosis may occur even in the absence of direct lymphocyte-target
cell interactions.
Activated cytotoxic T cells express Fas L, which binds to Fas expressed
on keratinocytes and results in apoptosis.
Apoptosis can also be induced via the release of effector
cell granules.
CD8+ CTL and NK cells use both the perforin/granzyme and
Fas/Fas L pathways, whereas Th1-type CD4+ T cells preferentially
use the Fas/Fas L pathway.
Human epidermal keratinocytes have been shown to have the ability
to produce granzyme B, perforin, and Fas L.
There is no microscopic evidence for the presence of apoptotic keratinocytes
in psoriasis, despite Fas expression on the lesional keratinocytes.
Susceptibility of the epidermis to undergo immune-mediated damage
may be dependent on its ability to express Fas L and the density of its
Fas membrane expression.
Normal melanocytes are long-lived post-mitotic cells that do not
produce any mitogens that stimulate their own growth. Expression of Bcl-2
on melanocytes may therefore be needed to escape from apoptosis.
The expression of Bcl-2 with melanocytic nevi tends to diminish when
neuroid changes are present: this finding may help explain the clinical
life cycle of melanocytic nevi.
Bcl-2 antisense oligonucleotide treatment improves the chemosensitivity
of human melanomas grown in severely combined immunodeficient mice.
Apoptosis can be found in a wide variety of both benign and malignant
skin tumors, including basal cell carcinoma (BCC), squamous cell carcinoma
(SCC), pilomatricoma, keratocanthoma, and Merkel cell tumor.
Apoptotic cells outnumber mitotic cells in BCC.
In SCC Bcl-2 is only detectable in the basal cells and a significantly
higher number of apoptotic cells can be observed than in BCC.
BCC cells express Fas-L, but not Fas, which may allow tumor expansion
by killing Fas-bearing activated T cells.
Apoptosis can be observed in pilomatricoma and regressing kerathoacanthoma.
Sun burn cells in the epidermis have been considered as keratinocytes
undergoing apoptosis.
UV irradiation induces both Fas and Fas L expression on keratinocytes.
UV light may act to induce Fas L in skin tumors such as BCC thereby
enabling them to escape from an immune attack by CTL, whereas UV-induced
Fas L on the psoriatic keratinocytes may act to kill intraepidermal T
cells, thereby improving the lesions.
After in vitro UV irradiation, p53 protein levels were noted
to increase prior to the induction of apoptosis in human keratinocytes.
p53-knockout mice do not develop apoptosis in the epidermis after
UV irradiation.
In normal catagen, apoptotic cells ares scattered in the outer root
sheath and are engulfed quickly, initially not by macrophages but by nearby
epithelial cells.
Proliferation, apoptosis, and necrosis occur simultaneously in keloids.
Endothelial cells are clearly the first cells to undergo apoptosis
in the skin of avian scleroderma and that apoptotic endothelial cells
can only be detected in early inflammatory disease stages of human scleroderma.
Abbreviations
ICE interleukin-1ß converting enzyme
TNF tumor necrosis factor
TNFR tumor necrosis factor receptor
PCD programmed cell death
NGFR nerve growth factor
FADD Fas-associated protein with death domain
MORT-1 mediator of receptor-induced toxicity
TRADD TNFR-associated death domain
lpr lymphoproliferation
gld generalized lymphoproliferative disease
CTL cytotoxic T cell
LMP-1 latent membrane protein
TRFA TNFR-associated factor
CED cell death abnormal
CONCLUSION
Apoptosis was first described as a morphologically distinct type of cell
death seen in lichenoid tissue reaction and some skin tumors. Within the
past few years, a large body of evidence on the molecular and cellular
mechanisms involved in apoptosis has been accumulating: although numerous
molecules that can initiate, execute, or inhibit the apoptotic process
have been identified, our understanding of how these molecules act remains
largely limited. It is therefore difficult to translate the knowledge
gained from these studies into clinical settings.
Thus, efforts aimed at treating disease by manipulating this process
are at relatively early stages. Because targeted induction or inhibition
of apoptosis is an ideal way to treat a particular disease, we need to
learn much more about how to promote or inhibit the apoptotic process
in selective tissues. The development of such therapeutic approaches should
remain a high priority.
REFERENCES
1. Kerr JFR, Wyllie AH, Currie AR. Apoptosis. A basic biological phenemenon
with wideranging implications in tissue kinetics. Br J Cancer 1972;
26: 239-57.
2. Sen S. Programmed cell death: concept, mechanism and control. Biol
Rev 1992; 67: 287-319.
3. Thompson CB. Apoptosis in the pathogenesis and treatment of disease.
Science 1995; 267: 1456-62.
4. Weedon D. Apoptosis. Adv Dermatol 1990; 5: 243-56.
5. Paus R, Rosenbach T, Haas N, Czarnetzki BM. Patterns of cell death:
the significance of apoptosis for dermatology. Exp Dermatol 1993;
2: 3-11.
6. Raskin CA. Apoptosis and cutaneous biology. J Am Acad Dermatol
1997; 36: 885-96.
7. Kerr JFR, Winterford CM, Harmon BV. Apoptosis: its significance in
cancer and cancer therapy. Cancer 1994; 73: 2013-26.
8. Manjo G, Joris I. Apoptosis, oncosis, and necrosis. Am J Pathol
1995; 146: 3-15.
9. Hakke AR, Palokowska RR. Cell death by apoptosis in epidermal biology.
J Invest Dermatol 1993; 101: 107-12.
10. Zhivotovsky B, Wade D, Nicotera P, Orrenius S. Role of nucleuses
in apoptosis. Int Arch Allergy Immunol 1994; 105: 333-8.
11. Nicholson DW, Thornberry NA. Caspases: killer proteases. TIBS
1997; 22: 299-306.
12. Nagata S. Apoptosis by death factor. Cell 1997; 22: 355-64.
13. Penninger JM, Kroemer G. Molecular and cellular mechanisms of T
lymphocyte apoptosis. Adv Immunol 1998; 68: 51-144.
13a. Powis G, Gasdask JR, Baker A. Redox signaling and the control of
cell growth and death. Adv Pharmacol 1997; 38: 329-59.
14. Ellis RE, Yuan Y, Horvitz HR. Mechanisms and functions of cell death.
Annu Rev Cell Biol 1991; 7: 663-98.
15. Griffith TS, Brunner T, Fletcher SM, Green DR, Ferguson TA. Fas
ligand-induced apoptosis as a mechanism of immune privilege. Science
1995; 270: 1189-92.
16. Watanabe-Fukunaga R, Adachi M, Matsuzawa A, Kasugai T, Kitamura
Y, Itoh N, Suda T, Nagata S. Lymphoproliferation disorder in mice explained
by defects in Fas antigen that mediates apoptosis. Nature 1992;
356: 314-7.
17. Galle PR, Hofmann WJ, Walczac H, Schaller H, Otto G, Stremmel W,
Krammer PH, Runkel L. Involvement of the CD95 (Apo1/Fas) receptor and
ligand in liver damage. J Exp Med 1995; 182: 1223-30.
18. Tsujimoto Y, Gorham J, Cossman J, Jaffe E, Groce CM. The t (14:18)
chromosome translocation involved in B cell neoplasma result from mistakes
in VDJ joining. Science 1985; 229: 1390.
19. Debra TC, Stanley JK. Bcl-2 family: regulators of cell death. Annu
Rev Immunol 1998; 16: 395-419.
20. McCall C, Cohen JJ. Programmed cell death in terminally differentiating
keratinocytes: role of endogenous endonuclease. J Invest Dermatol
1991; 97: 111-4.
21. Langley RG, Walsh N, Nevill T, Thomas L, Rowden G. Apoptosis is
the mode of keratinocytes death in cutaneous graft-versus-host
disease. J Am Acad Dermatol 1996; 35: 187-90.
22. Inachi S, Mizutani H, Shimizu M. Epidermal apoptotic cell death
in erythema multiforme and Stevens-Johnson syndrome. Arch Dermatol
1997; 133: 845-9.
23. Shimizu M, Higaki Y, Higaki M, Kawashima M. The role of granzyme
B-expressing CD8-positive T cells in apoptosis of keratinocytes in lichen
planus. Arch Dermatol Res 1997; 289: 527-32.
24. Gilliam AC, Whitaker-Menezes D, Korngold R, Murphy G. Apoptosis
is the predominant form of epithelial target cell injury in acute experimental
graft-versus-host disease. J Invest Dermatol 1996; 107:
377-83.
25. Shiohara T, Moriya N, Nagashima M. Induction and control of lichenoid
tissue reactions. Springer Semin Immunopathol 1992; 13: 369-85.
26. Berke G. The CTL's kiss of death. Cell 1995; 81: 9-12.
27. Sayama K, Yonehara S, Watanabe Y, Miki Y. Expression of antigen
on keratinocytes in vivo and induction of apoptosis. J Invest
Dermatol 1994; 103: 330-4.
28. Berthou C, Michel L, Soulie A, Jean LF, Flageul B, Dubertret L,
Sigaux F, Zhang Y, Sasportes M. Acquisition of granzyme B and Fas ligand
proteins by human keratinocytes contributes to epidermal cell defense.
J Immunol 1997; 159: 5293-300.
29. Wrone-Smith T, Johnson T, Nelson B, Boise LH, Thompson CB, Nunez
G, Nickoloff BJ. Discordant expression of Bcl-x and Bcl-2 by keratinocytes
in vitro and psoriatic keratinocytes. Am J Pathol 1995;
146: 1079-88.
30. Van den Oord JJ, Vandeghinste N, De Ley M, De Wolf-Peeters C. Bcl-2
expression in human melanocytes and melanocytic tumors. Am J Pathol
1994; 145: 294-300.
31. Saenz-Santamaria MC, Reed JA, McNutt S, Shea CR. Immunohistochemical
expression of Bcl-2 in melanomas and intradermal nevi. J Cutan Pathol
1994; 21: 393-7.
32. Morales-Ducret CRJ, van de Rijn M, Smoller BR. Bcl-2 expression
in melanocytic nevi. Arch Dermatol 1995; 131: 915-8.
33. Jansen B, Schlagbauer WH, Brown BD, Bryan RN, van Elsas A, Muller
M, Wolff K, Eichler HG, Pehamberg H. Bcl-2 antisense therapy chemosensitizes
human melanoma in SCID mice. Nat Med 1998; 4: 232-4.
34. Hahne M, Rimoldi D, Romero SP, Schreier M, French LE, Schneider
P, Bornald T, Fontana A, Lienard D, Cerottini JC, Tschopp J. Melanoma
cell expression of Fas (Apo-1/CD95) ligand: implications for tumor immune
escape. Science 1996; 274: 1363-6.
35. Weedon D. Apoptosis: its nature and implications for dermatopathology.
Am J Dermatopathol 1979; 1: 133-44.
36. Cerroni L, Kerl H. Aberrant Bcl-2 protein expression provides a
possible mechanism of neoplastic cell growth in cutaneous basal-cell carcinoma.
J Cutan Pathol 1994; 21: 398-403.
37. Morales-Ducret CRJ, van de Rijn M, LeBurn DP, Smoller BR. Bcl-2
expression primary malignancies of the skin. Arch Dermatol 1995;
131: 909-12.
38. Verhaege MEJM, Sanders CJG, Arends JW, Neumann HAM. Expression of
the apoptosis-suppressing protein Bcl-2 in non-melanoma skin cancer. Br
J Dermatol 1995; 132: 740-4.
39. Wikonkal NM, Berg RJ, van Haselen CW, Horkay I, Remenyik E, Begaany
A, Hunyadi J, van Vloten WA, de Gruijl FR. Bcl-2 vs p53 protein
expression and apoptotic rate in human nonmelanoma skin cancers. Arch
Dermatol 1997; 133: 599-602.
40. Buechner SA, Werni M, Harr T, Hahn S, Itin P, Erb P. Regression
of basal cell carcinoma by intralesional interferon-alpha treatments is
mediated by CD95 (Apo-1/Fas)-CD95 ligand-induced suicide. J Clin Invest
1997; 100: 2691-6.
41. Gniadecki R, Hansen M, Wulf HC. Two pathways for induction of apoptosis
by ultraviolet radiation in cultured human keratinocytes. J Invest
Dermatol 1997; 109: 163-9.
42. Leverkus M, Yaar M, Gilchrest BA. Fas/Fas ligand contributes to
UV-induced apoptosis in human keratinocytes. Exp Cell Research
1997; 232: 255-62.
43. Gutierrez-Steri C, Wrone-Smith T, Sun X, Krueger JG, Coven T, Nickoloff
BJ. Sunlight-induced basal carcinoma tumor cells and ultraviolet B-irradiated
psoriatic plaques express Fas ligand (CD95L). J Clin Invest 1998;
101: 33-9.
44. Henseleit U, Zhang J, Wanner R, Hasse I, Kolde G, Rosenbach T. Role
of p53 in UVB-induced apoptosis in human HaCaT keratinocytes. J Invest
Dermatol 1997; 109: 722-7.
45. Linder G, Botchkarev VA, Bochkarev NV, Ling G, van der Veen C, Paus
R. Analysis of apoptosis during hair follicle regression (catagen). Am
J Pathol 1997; 151: 1601-17.
46. Cese Raffaele, Cazzanigra S, Morelli D, Sfondrini L, Bignotto M,
Menard S, Colnaghi M, Balsari A. Apoptosis of hair follicle cells during
doxorubicin-induced alopecia in rats. Lab Invest 1996; 75: 601-9.
47. Appleton I, Brown NJ, Willoughby DA. Apoptosis, necrosis, and proliferations:
possible applications in the etiology of keloid. Am J Pathol 1996;
149: 1441-7.
48. Sgonc R, Gruschwitz MS, Dietrich H, Recheis H, Gershwin ME, Wick
G. Endothelial cell apoptosis is primary pathogenic event underlying skin
lesions in avian and human scleroderma. J Clin Invest 1996; 98:
785-92.
|
Version imprimable
