Dr Robert Sabat, Interdisciplinary Group of Molecular Immunopathology, Dermatology/Medical Immunology, Campus Charité Mitte, University Hospital Charité, Charitéplatz 1, Berlin D-10117, Germany, Tel.: +49 30 450 518 009, Fax: +49 30 450 518 964, e-mail: firstname.lastname@example.org
Abstract: Psoriasis is a chronic skin disease that affects about 1.5% of the Caucasian population and is characterized by typical macroscopic and microscopic skin alterations. Psoriatic lesions are sharply demarcated, red and slightly raised lesions with silver-whitish scales. The microscopic alterations of psoriatic plaques include an infiltration of immune cells in the dermis and epidermis, a dilatation and an increase in the number of blood vessels in the upper dermis, and a massively thickened epidermis with atypical keratinocyte differentiation. It is considered a fact that the immune system plays an important role in the pathogenesis of psoriasis. Since the early 1990s, it has been assumed that T1 cells play the dominant role in the initiation and maintenance of psoriasis. However, the profound success of anti-tumor necrosis factor-α therapy, when compared with T-cell depletion therapies, should provoke us to critically re-evaluate the current hypothesis for psoriasis pathogenesis. Recently made discoveries regarding other T-cell populations such as Th17 and regulatory T cells, dendritic cells, macrophages, the keratinocyte signal transduction and novel cytokines including interleukin (IL)-22, IL-23 and IL-20, let us postulate that the pathogenesis of psoriasis consists of distinct subsequent stages, in each of them different cell types playing a dominant role. Our model helps to explain the varied effectiveness of the currently tested immune modulating therapies and may enable the prediction of the success of future therapies.
Psoriasis vulgaris (a.k.a. psoriasis in the rest of this review) is a common, chronic, relapsing skin disease. It is characterized by macroscopic (clinical) and corresponding microscopic (histological) skin alterations and leads to considerable impairment of the quality of life of the affected patients. Moreover, special forms of psoriasis (e.g. arthropathic form) can be accompanied by severe extra-cutaneous changes. In addition to the disease frequency and patients’ disability, researchers have concentrated on psoriasis because of the postulated pathogenesis commonality of this disorder with other chronic immune-mediated inflammatory diseases. In fact, similar reactions are hypothesized to contribute to the initiation and maintenance of diseases such as rheumatoid arthritis and Crohn’s disease, and it is hoped that an improved understanding of the pathogenesis of psoriasis will aid in the understanding of other chronic inflammatory diseases and lead to novel treatment options.
Psoriasis is a worldwide occurring disease. In the Caucasian population, the prevalence is about 1.5% (1,2). It means that in Europe alone, more than 7 million people are afflicted with this disease. In other ethnic groups such as the Japanese, the prevalence of psoriasis is much lower (3,4). Epidemiological studies revealed that a distinct group of disorders is quite frequently associated with psoriasis, e.g. rheumatoid arthritis, colitis, diabetes, metabolic syndrome and hypertension. In contrast, atopic dermatitis and allergies are less frequently seen to be associated with psoriasis compared with normal rates of occurrence (5). Interestingly, in contrast to atopic dermatitis, the reported incidence of psoriasis has not increased in the last 20–30 years. Men and women are affected by psoriasis at the same rate. The first manifestation of the disorder usually occurs around the age of 20 or between 50 and 60. However, it must be emphasized that psoriasis can manifest itself at any age. According to Henseler and Christophers, psoriasis can be differentiated into two subgroups: type I, which begins before age 40, and type II, which begins after age 40 (6). Type I psoriasis, which accounts for approximately 75% of all psoriasis patients, is associated with a more severe course of disease, limited success of treatment, increased prevalence of certain human leucocyte antigen (HLA)-types and stronger hereditary ties. Although the inheritance pattern is currently still unclear, genetic disposition appears to play an important role in the susceptibility to develop psoriasis. This view is based on three observations. First, the likelihood of developing psoriasis is raised when first-grade relatives suffer from the disease. The risk is about 20% if one parent has psoriasis, and is about 75% if both parents are affected. If one monozygotic twin suffers from psoriasis, the probability is more than 55% that the other will be affected too (4,7). Second, psoriasis is associated with certain HLA-types (HLA-Cw6, HLA-B13, HLA-B17, HLA-Bw57 and HLA-DR4). People with HLA-Cw6, for example, have a 10-fold higher risk of disease (4,5). Thirdly, several psoriasis susceptibility loci have been described. The PSORS1 in the major histocompatibility complex (MHC) region on chromosome 6 (6p21) appears to be associated with most cases of psoriasis. Interestingly, some genes from this region are associated also with other immune diseases (rheumatoid arthritis, colitis and diabetes) (3,8).
An infection with β-haemolytic streptococci often precedes the first manifestation of psoriasis (9). The course of the disease is then chronic whereby the disease length, intensity and recurrence are very different between patients and sometimes even in the same patient. Consecutive new exacerbations can be triggered by mechanical irritation (so-called Koebner reaction; new lesions emerge at locations that are mechanically irritated), medications (e.g. β-receptor blockers, lithium, chloroquine, non-steroidal anti-inflammatory agents, tetracyclines and interferons) and infections (e.g. viral or bacterial infections) (10–12). Interestingly, a positive Koebner reaction predicts subsequent disease activity (10).
Psoriatic lesions are of different shape, sharply demarcated, red and slightly raised lesions with silver-whitish scales (Fig. 1). The scales are only lightly attached and can be easily peeled off in toto. Upon peeling several layers at once, point bleeding can occur in the now apparent dermal papillae. Frequent locations for psoriasis plaques are the extensor side of the extremities, the sacral region and the head. The lesions often have a small point form at the onset. During the course of the disease, they grow and can take on a geographical shape or, in very severe cases, cover the whole body (3–5). It is also common to observe nail changes in psoriasis (13). Additionally, more than 10% of psoriasis patients have arthritis (14).
Already at the onset of a psoriatic plaque, histological alterations include dermal oedema, dilatation of vessels of the papilla in the dermis and perivascular cell infiltration composed of T cells, dendritic cells (DC) and monocytes/macrophages (15). Later, the density of infiltrates increases, and CD8+ T cells and neutrophilic granulocytes are found particularly in the epidermis. Neutrophilic granulocytes form very characteristic Munro’s microabscesses in the epidermis (15,16). Other prominent changes are found in the epidermis: acanthosis (raised number of keratinocytes and the thickening of the spinous layer), loss of the granular layer, parakeratosis (dysfunction of the cornification process with nucleus-containing keratinocytes in the cornified layer) and hyperkeratosis (thickening of the cornified layer). In the chronic stage, the epidermal changes come to the fore. At the same time, an increasing amount and dilatation of capillaries in the dermal papillae, surface vessels that facilitate a renewed immigration of the immune cells, is observed (4,8).
Psoriatic skin changes are well known since biblical times. The first documented description is found in the Old Testament in the third book of Moses. That means that humankind has dealt with this disease for at least 3000 years. Since then, the hypothesized causes of the disease have naturally evolved. Until the late 1970s, the cause of the disease was considered to be due to a dysfunctionally increased proliferation and altered differentiation of the keratinocytes (17,18). The typical microscopic changes of the epidermis offer a good indication for this. In the 1980s and 1990s, three observations were made that allowed researchers to assume that activated T cells have a dominant pathogenic role in the initiation and persistence of psoriasis (19,20). First, therapeutic success was found with medications that inhibit T-cell functions. The first of these pharmaceutical products was cyclosporin A, a substance that diminishes T-cell proliferation and cytokine production (21). An improvement of psoriasis was also observed after treatment with other T-cell modulating drugs: anti-CD4 antibodies (22,23) and a fusion protein composed of interleukin (IL)-2 and fragments of diphtheria toxin (24). Second, psoriatic lesions healed when patients received a haematopoietic stem cell transplant, necessary because of an unrelated disease, from a non-psoriatic donor. On the other hand, if an otherwise healthy patient receives bone marrow from a donor with psoriasis, the patient also frequently develops psoriasis (25,26). Thirdly, in a severe combined immunodeficiency mouse that received a transplant of uninvolved skin taken from a patient with psoriasis, psoriasis-like alterations of the transplanted skin could be observed if autologous blood immune cells were activated in vitro and injected into this skin (27–29). Interestingly, when skin grafts from healthy individuals were transplanted and autologous blood immune cells were injected, no conversion to psoriasis-like plaques was found (29).
Many new insights have been gained over the last few years that, in our opinion, should change the view of the pathogenesis of psoriasis as a T1-mediated skin disease. These new results include: (i) the observations of a profound success of anti-tumor necrosis factor (TNF)-α therapy in psoriasis patients, (ii) the knowledge regarding the role of several types of immune cells such as DCs, Th17 cells, γδ T cells, natural killer (NK) T cells and regulatory T cells, (iii) the knowledge about the effects of signal transduction activation in keratinocytes (e.g. STAT3) and (iv) the findings regarding new mediators such as IL-22, IL-23 and IL-20. All of these new insights prompted us to propose a novel model of psoriasis pathogenesis. Our model is based on numerous research results, although at some points the facts are tied together by conjecture. These conjectures are clear identified as such in the text. According to our model, the onset of the disease is similar to an immune reaction, which is composed of three phases: a sensitization phase, a silent phase and an effector phase (Fig. 2). During the sensitization phase, DCs process and present antigens and subsequently induce the development of skin infiltrating effector/memory Th17 and T1 cells. It is important to note that the sensitization phase is not accompanied by any skin alterations. The sensitization phase is then followed by a silent phase of variable length. Each cycle of the reoccurring effector phase can be differentiated into three subsequent stages: (i) skin infiltration of immune cells, (ii) immune cell activation in the skin and (iii) keratinocyte response. The ‘keratinocyte response’ leads to an overshot reaction of these tissue cells that is reminiscent of an overshot regeneration process (like wound-healing). After a successful treatment, the effector phase transposes into the silent phase (Fig. 1). After a while, a renewed effector phase follows. In our model, different cell types contribute significantly to the onset of psoriasis at different time points. This should be taken into consideration when developing new therapies. Currently, most patients are treated during the ‘keratinocyte response’ stage of our model, because this is the only stage of psoriasis pathogenesis in which skin alterations are visible. We hypothesize that in this stage T cells, macrophages and DCs are initially mainly responsible for the alteration of keratinocytes. However, over time, macrophages, DCs and later tissue cells also play a dominant role. In the final stages of the ‘keratinocyte response’, T cells take on only a minor role.
In subsequent sections of this review we describe in detail every stage of our model.
Sensitization phase – antigen processing and presentation
The initial step of every specific immune reaction to an antigen (Ag) is performed by professional antigen presenting cells (APCs) such as DCs or macrophages. At the onset of psoriasis, the initial step may comprise the recognition and uptake of (exogenous) Ag by tissue-guarding, immature DCs, the migration of these cells to the T-cell areas of secondary lymphatic organs, the processing of the Ag and the presentation of selected Ag fragments (peptides) on the DCs’ cell surface MHC class II (MHC II) molecules to T cells. In this process, the composition of the antigen-processing proteases in DCs as well as the DCs’ MHC type determines the nature of the presented peptides and thereby, in the end, the specificity of the effector and memory T cells that is generated in the sensitization phase.
After the Ag uptake, DCs undergo a maturation process if they encounter inflammatory conditions (because of cytokines like TNF-α or IL-1β) or microbial products that stimulate DCs via Toll-like receptors (30,31). They upregulate the functional CCR7 chemokine receptor on their surface and migrate into the T-cell area of the regional lymph nodes via afferent lymphatic vessels following a gradient of CCL21 and CCL19 (ligands for CCR7) (32–34) thereby carrying the previously uptaken Ag with them and drastically reducing their capacity to further absorb Ags (35,36). Moreover, several changes occur regarding the MHC II pathway. In immature DCs, most MHC II resides intracellularly. Upon DC maturation, a redistribution of MHC II occurs in the cells leading to high expression on the cell surface. Some studies provided evidence that the intracellular accumulation of MHC II in immature DCs is caused by the inefficient formation of mature complexes of MHC II and antigenic peptides that are caused by a minimal cellular lysosomal activity. In fact, it has been shown that in immature DCs, the (at least partially intact) Ag as well as the MHC II, the latter being associated with the chaperonic invariant chain fragment p10, resided within the cell in specialized endosomal/lysosomal compartments, co-localized with HLA-DM, inactive hydrolytic proteases, and protease inhibitors such as cystatin C (37–41). The invariant chain is known to stabilize the conformation of freshly synthesized MHC II, to protect it from premature peptide loading by blocking its binding groove with its class-II-associated invariant-chain peptide (CLIP) portion, and to retain the MHC II in the endosomal compartments with the help of two sorting signals in its cytoplasmic portion (42). Increased protease activity upon DC maturation may be achieved by the decreased production of protease inhibitors and the activation of the vacuolar ATPase (38,40), and leads to splitting of both the invariant chain and the Ag. The splitting of the invariant chain occurs in a defined sequence of steps (43). The limiting step is the last cleavage whereby the endosomal retention signal sequence is split-off from the p10 fragment to leave the MHC II peptide-binding, groove-blocking CLIP fragment. In DCs, this last cleavage is thought to be carried out by the cysteine endopeptidase cathepsin S (44,45). Like the processing of the invariant chain, the degradation of the uptaken Ag upon DC maturation is a process in which numerous proteases, including cathepsins and legumain, play a role (42). The end result is approximately 12–25 amino acid long peptides. These peptides are exchanged for CLIP, provided a certain affinity for the MHC II concerned exists. This most probably takes place with the help of the non-classical MHC II molecule HLA-DM (43). The assembled peptide-loaded MHC II can then be transported to the surface of the DC. The additional mechanism for the intracellular accumulation of MHC II in immature DCs may be the increased turnover of cell surface MHC II peptide complexes because of their increased endocytosis (46). In fact, the half-life of the MHC II at the cell surface is about 10 h in immature DCs, and increases to more than 100 days after DC maturation (47).
Interestingly, the synthesis of further MHC α and β chains, after a transient upregulation, is inhibited in mature DCs (47,48). This has been shown to be caused by the reduced expression of the transcriptional regulator CIITA that is caused by the epigenetic modulation of the promoters of the encoding gene (49). Like the downregulated Ag uptake, the reduced MHC II synthesis may also prevent the presentation of newly encountered Ags. This prevention, in turn, in connection with the low turnover of cell surface MHC II, may cause DCs to retain a long-lasting, selective memory of the special, previously uptaken Ag as proven by its capacity to stimulate T cells even several days later (47,48). MHC II peptide complexes on the DC surface can be recognized by CD4+ T cells via the T cell receptor (TCR) given the fact that this TCR is specific for the MHC–peptide complex. The increased cell surface expression of peptide-loaded MHC II is also accompanied in DCs by strengthened expression of adhesion and co-stimulatory molecules (see below).
Although the presentation of peptides derived from Ags from the extracellular environment typically occurs via the MHC II pathway as described above, DCs are endowed with the capacity to cross-present those Ags on MHC I molecules leading to the additional, initial activation of CD8+ T cells (50).
At this point, there remains at least one question that is crucial for the understanding of the pathogenesis of psoriasis, namely which Ag is responsible for initial T-cell priming in the majority of patients.
The answer should be revealed when the reactivity of T cells in psoriatic lesions can be identified. To this purpose, the Fry group isolated T cells from psoriatic lesions of acute guttate and chronic plaque psoriasis and established T-cell lines from these cells. They could show that these T-cell lines were reactive against isolates from streptococci (51,52). A streptococcal origin of the primary Ag(s) would be in line with the fact that first manifestation and relapses and aggravations of psoriasis are often linked to infections with streptococci, especially β-haemolytic Streptococcus pyogenes (9,11). At this point, it should be emphasized that neither the portion of psoriasis patients with skin infiltrated T cells against streptococcal Ags nor the proportion of such T cells within the whole T-cell infiltrate is currently known. In addition to streptococcal Ags, other Ags, even a few endogenous Ags, were postulated to trigger psoriasis. Interestingly, it has been shown in experimental models that DCs can break tolerance to endogenous antigens via bystander activation under certain circumstances. In this case, it would make no sense to look for exogenous ‘cross-reactive’ antigens as a trigger of psoriasis.
Sensitization phase – generation of effector and memory T cells
The conversion from naïve to effector and memory T cells that occurs in secondary lymphatic organs under the guidance of DCs is necessary for T cells to obtain their functions and their ability to immigrate into tissues. The naïve T cells permanently circulate between blood and secondary lymphatic organs, such as lymph nodes and tonsils. In T-cell areas of secondary lymphatic organs, they congregate with mature DCs. A naïve CD4+ T cell whose TCR has a high enough affinity for the respective MHC II–peptide complex, sticks to the respective DC and forms a so-called immunological synapsis. In order to become activated, the T cell now needs three signals: the first signal has already been delivered by interaction between TCR and MHC II–peptide complex. The second signal is given by the so called co-stimulatory molecules. The third signal for T-cell activation is delivered by soluble mediators (Fig. 3).
The most important co-stimulatory molecules expressed by DCs are those of the B7 family: CD86 (B7-2), CD80 (B7-1), B7h, PD-L1 and PD-L2 as well as CD40 (53,54). CD86 and CD80 interact with the T-cell molecules CD28 and CD152 (cytotoxic T-lymphocyte-associated protein 4, CTLA-4). CD28 shows a large constitutive expression in T cells, particularly in naïve T cells. Engagement of CD28 reduces the required number of triggered TCRs for T-cell activation and allows activation of T cells by low affinity ligands (55). Additionally, it is important for stable IL-2 production and IL-2Rα (CD25) expression, and it prevents the induction of unresponsiveness and apoptosis of T cells after TCR stimulation (55,56). Hence, CD86 functions as a major co-stimulatory molecule in DCs and is critical for full activation, particularly of naïve T cells. CD152 does not appear to be expressed on resting T cells, but is transported from intracellular clathrin vesicles to the cell surface following TCR stimulation (57). Engagement of CD152 inhibits TCR- and CD28-mediated signal transduction, increases the threshold for activation of these cells and represses the cell cycle (58). Therefore, it counteracts acute T-cell responses but is also essential for the formation of memory T cells (otherwise all cells would differentiate into effector cells with short life times). The counterpart for B7h (DC) on T cells is ICOS, for PD-L1 and PD-L2 on T cells is PD-1, and for CD40 on T cells is CD154 (CD40 ligand). The immunological synapsis is stabilized by adhesion molecules expressed by DCs and T cells (59). Among the most important are (i) the interactions between intercellular adhesion molecule 1 (ICAM-1; CD54) (DC) and lymphocyte function-associated antigen-1 (LFA-1; CD11a/CD18; αL:β2) (T cells) and (ii) the interactions between lymphocyte function-associated antigen 3 (LFA-3; CD58) (DC) and CD2 (T cells).
The third signal mentioned above essentially influences the mode of action of the effector and memory T cells, which evolve from these activated T cells (Fig. 3). According to what we actually know, naïve CD4+ T cells can be polarized into four different directions: Th1, Th2, Th17 and regulatory T cells. If IL-12 (p35/p40) is present during activation of naïve T cells, they are polarized into Th1 cells through activation of the transcription factors STAT4 and T-bet. The generation of Th1 cells is supported by gamma interferon (IFN-γ), which increases the expression of the specific receptor chain for IL-12 (IL-12Rβ2) (60). During repeated activation, Th1 cells primarily produce IFN-γ, IL-22, IL-26, GM-CSF and TNF-β. If IL-4 is present, naïve T cells are polarized into Th2 cells through activation of the transcription factors STAT6 and GATA-3 (60). These T cells secret IL-4, IL-5 and IL-13 during repeated activation. The activation of naïve T cells with mature DCs in the presence of IL-6 and transforming growth factor (TGF)-β upregulated the receptor for IL-23 (p19/p40) and together with this cytokine induced Th17 cell development (61,62). Th17 cells primarily produce IL-6, IL-17 and IL-22. Th17 development is independent of STAT1, STAT4 and STAT6 signalling. The Th1-cell inducing cytokines IL-12 and IFN-γ actively suppress the development of Th2 and Th17 cells. Conversely, the Th2-cell inducing cytokine IL-4 inhibits the development of Th1 and Th17 cells (61,62). Regulatory T cells are thought to be responsible for limiting T-cell immune responses, partly by production of immunosuppressive cytokines, such as IL-10 and TGF-β. However, there is still a rather broad discussion on the generation and maintenance of regulatory T cells (60).
Preliminary results describe distinct patterns of cell-surface molecule expression allowing phenotypical distinction between Th1, Th2, Th17 and regulatory T cells. CCR5 and CXCR3 expression are thought to characterize Th1-cells whereas CCR3 and CCR4 expression are assigned to Th2-cells (63–65). However, Th17 T cells may, also be CCR4+. Regulatory T cells seem to be contained in the bright positive CD25 fraction of CD4+ T cells (66).
Independent of its polarization into a Th1, Th2, or Th17 cell, a successfully activated naïve T cell develops into either an effector T cell (TE), an effector memory T cell (TEM) or a central memory T cell (TCM) (67). Effector T cells immediately home into inflamed tissue, display their effector functions and die. TEM cells recirculate between blood and peripheral tissues and rapidly produce effector cytokines upon restimulation, but they have a poor proliferative capacity. In contrast, TCM cells mainly recirculate between blood and lymph nodes and have limited effector functions, but proliferate and become effectors following secondary stimulation. For peripheral blood CD4+ and CD8+ T cells, the expression of the lymph node homing receptors CCR7 and CD62 L in combination with CD45RA expression may be able to characterize these subsets phenotypically: in CD4+ T cells CD45RA+/CCR7+/CD62 L+ cells represent the naïve compartment, CD45RA−/CCR7−/CD62 L cells represent the TEM compartment, and CD45RA−/CCR7+/CD62 L+ cells represent the TCM compartment. Among CD8+ T cells, effector T cells seem to add to the CD45RA+/CCR7−/CD62 L− fraction (68,69).
A subset of successfully activated T cells expresses the so-called cutaneous lymphocyte-associated antigen (CLA). CLA, a defined carbohydrate epitope (so-called sialyl-Lewis(x)), guides leucocytes to inflamed skin. With help from fucosyltransferase VII, CLA can be placed on top of P-selectin glycoprotein ligand-1 (PSGL-1) (70). PSGL-1 is a cell surface molecule expressed constitutively on all peripheral T cells (70). Mice transgenically deficient in fucosyltransferase VII show a reduced percentage of skin homing T cells and other leucocytes (71). Very recent results show that CD43, a sialomucin also constitutively expressed on T cells, can also be crowned with the CLA epitope (72). Interestingly, CLA-modified PSGL-1 binds both E- and P-selectin, whereas CLA modified CD43 only binds E-selectin. However, the mechanisms by which DCs induce the expression of CLA on a defined subset of activated T cells are still widely unknown. DCs may induce the expression of tissue-specific homing receptors on T cells congruent to the organ from which the DC comes. An exception may be made for tonsillar DCs that present streptococcal Ags corresponding to the localization of streptococcal infection in the upper respiratory tract and the tonsils. These DCs, however, seem to generate some skin-homing T cells as, like in the skin-draining lymph nodes, 5–10% of tonsillar T cells express CLA. Most interestingly, TCR spectra-types and sequencing of TCR rearrangements provided evidence of shared T-cell clonality of tonsillar and skin lesional CLA+ T cells, but not CLA-tonsillar and peripheral T cells in psoriatic patients (73). In contrast to the generation of skin-homing T cells in tonsils, the generation of such cells could occur directly in skin-draining lymph nodes because of a cutaneous presence of streptococcal Ags that may reach the skin to some extent during streptococcal infection (74).
Effector phase – skin infiltration of T and other immune cells
If psoriatic plaques develop, the first microscopically visible events are the superficial perivascular infiltration of lymphocytes and monocytic cells and the dilation of the blood vessels in the dermal papillae (15). The passage of leucocytes from the blood vessels into tissue occurs in five steps (Fig. 4). In these processes, endothelial cells play a decisive role.
In the first step, leucocytes roll along the blood vessel wall. Rolling reduces the flow velocity of the leucocytes and is mediated by the interaction between P- and E-selectin expressed by endothelial cells and selectin ligands expressed by leucocytes (PSGL-1, CLA-modified PSGL-1 and CLA-modified CD43). In healthy controls, endothelial cells of cutaneous postcapillary venules express low amounts of P-selectin molecules (75). In the uninvolved skin of patients with psoriasis, endothelial cells express both P- and E-selectin, and its expression is much more enhanced in psoriatic skin (75). One of the actions of TNF-α is the early release of so-called Weibel-Palade bodies containing P-selectin. Additionally, TNF-α induces the synthesis and surface expression of E-selectin molecules.
Very important for the tissue homing is the production of chemokines, which exist not only as soluble mediators but also bound to glycosaminoglycans on endothelial cells at the site of inflammation (76,77). In the second step of skin infiltration, the immune cells rolling along the blood vessel wall recognize chemokines presented by endothelial cells and are activated. It was postulated that the chemokine CCL27 (ligand for CCR10) is highly relevant for migration of T cells into psoriatic skin (78) although other chemokines seem to play a predominant role (76,77,79–81). Reiss et al. show that the inhibition of CCL27 has no effect on lymphocyte recruitment into skin in the psoriatic mouse model, but only the additional blockage of the interaction between CCL17 and CCR4 abrogates skin recruitment in this model (80). Moreover, most skin-infiltrating lymphocytes in allergic delayed-type hypersensitivity express CCR4, but only about 10% express CCR10 (77). In psoriasis lesions, the majority of CD4+ T cells and about half of the CD8+ T cells expressed CCR4 and CCR6 (81). Interestingly, CCL17, the ligand for CCR4, was expressed by TNF-α or IFN-γ-stimulated dermal endothelial cells but not by activated keratinocytes (81,82). CXCR3 is expressed by about one-third of CD4+ and CD8+ T cells infiltrating psoriatic skin lesions. CXCR3 may be particularly important for the migration of CD8+ T cells into the epidermis, as its ligands CXCL9 and CXCL10 are expressed by TNF-α or IFN-γ-stimulated keratinocytes (82).
The next step of skin infiltration is characterized by the formation of tight adhesions between endothelial cells and immune cells (Fig. 4). This is achieved by integrins expressed on immune cells and their ligands expressed on endothelial cells. Chemokines induce the integrin-dependent adhesion of immune cells to endothelial cells and cause the rapid arrest of the former under physiological flow (76). CD11a/CD18 (LFA-1) seems to be the most important integrin for skin homing. It binds to CD54 and CD102 (ICAM-2) expressed by endothelial cells. Another integrin system, consisting of very late antigen (VLA)-4 (α4:β1) expressed by T cells and vascular cell adhesion molecule 1 (CD106) expressed by endothelial cells, does seem to be less important in this process. In fact, the vast majority of intra-epidermal T cells in psoriatic skin lesions expressed CD11a/CD18, whereas only 58% of CD4+ and 85% of CD8+ T cells expressed VLA-4 (81). The endothelium-presented chemokines triggered instantaneous extension of bent CD11a/CD18, which led to an increased affinity of CD11a/CD18 – CD54 binding (83). Although VLA-4 affinity is not altered upon chemokine signalling, subsequent VLA-4 clustering at the leucocyte–substrate contact zone results in enhanced leucocyte avidity to CD106 (84). Resting endothelial cells only express CD102. TNF-α and IFN-γ, but not IL-6 or GM-CSF, induced the expression of CD54 on endothelial cells (85,86).
The passage of immune cells through the endothelial wall is called diapedesis and is probably performed by pores formed between endothelial cells (Fig. 4). It seems that this step is also dependent on integrins and, within the integrin system, more on CD11a/CD18 than on VLA-4 (87). In fact, the intensity of endothelial cell expression of CD54, the CD11a/CD18 ligand, correlated with the degree of dermal inflammation (88). A small amount of peripheral T cells also migrates into non-inflamed skin (see TEM cells). This is made possible by constitutive expression of P-selectin and CD102 on resting dermal endothelial cells. Interestingly, Lowes et al. showed that in non-lesional psoriatic skin, the number of T cells is higher than in skin from healthy participants (89). In psoriatic skin lesions, CD8+ T cells primarily home into the epidermis, while CD4+ cells are mainly present in the dermis (90,91). The reasons for these different anatomical homing patterns may be caused by the varied expression of chemokine receptors (such as CXCR3, see above) and integrins (such as CD103) on CD4+ and CD8+ T cells. In fact, 80% of epidermal CD8+ T cells expressed the integrin CD103 (αE:β7). In contrast, only 5% of the dermal T cells and less than 1% of peripheral blood lymphocytes derived from psoriatic patients or healthy participants were CD103-positive (92,93). CD103 binds epidermal E-cadherin (94). Interestingly, in vitro culture experiments showed that CD103 was preferentially expressed on CD8+ T cells after stimulation with anti-CD3 monoclonal antibodies. Co-culture with TGF-β and IL-4 upregulated the CD103 expression on T cells, whereas IL-12 downregulated it (92,93).
Other immune cells such as NK cells, monocytes/macrophages, DCs and neutrophilic granulocytes also use a mechanism similar to the one described above for T cells that migrate into the skin. However, they are often guided to the skin by other chemotactic factors and can also use other adhesion molecules. In fact, skin-derived NK cells have been shown to express high levels of the chemokine receptors CXCR3 and CCR5, and intermediate amounts of CXCR1, CCR6 and CCR8. In line with that, prompt migration of these cells could be induced in vitro with CXCL10 and CCL5, and to a lower extent with CCL20 and CCL4 (95). The complement component C5a is apparently the most important of these chemotactic factors that capture the myeloid DCs in psoriatic lesions (96). Langerhans cells immigrate into psoriatic lesions most likely because of CCL20 (ligand for CCR6). Keratinocytes produce CCL20 after stimulation and CCL20 is upregulated in lesional psoriatic skin (97). For homing of neutrophilic granulocytes, interaction between the integrin MAC-1 (CD11b/CD18; αM:β2), expressed by neutrophils, and CD90, expressed by endothelial cells, does play an important role (98). In the end, neutrophils migrate into the stratum corneum and form microabscesses characteristic for psoriatic skin lesions. CXCL8 and complement factors (C5a) may also play an important role in this immigration.
Mostly as a consequence of the infiltration, significantly more immune cells are found in psoriatic lesions than in healthy skin. Interestingly, the numbers of T cells, macrophages and DCs are similar (89,99).
Not surprisingly, CD4+ and CD8+ T cells of psoriatic skin lesions have activation and memory phenotypes (90,91,100). T cells from the lesional epidermis expressed HLA-DR (86%), CD69 (59%), CD25 (55%) (all activation markers), and were CD45RA negative (91%) (memory phenotype). T cells from lesional dermis showed similar characteristics except for CD69 (17%) (101). CD69 is a surface molecule that is expressed by T cells for only a short time after activation. Regarding cytokine phenotypes, Schlaak et al. already described in 1994 that T-cell clones generated from lymphocytes taken from psoriatic skin partially exhibited a T1 cytokine secretion profile (102). The authors demonstrated that such clones produced large amounts of IL-2 and IFN-γ but only a little or no IL-4, IL-10 and TNF-α (102). A later study demonstrated that about 40% of skin CD4+ and CD8+ T cells were able to produce IFN-γ, IL-2 and TNF-α after ex vivo stimulation (103). Recently, our group generated quantitative data on mRNA expression of 20 cytokines demonstrating that of all T-cell cytokines investigated, IL-22 and IL-17 had the highest expression in lesional psoriatic skin (104,105) (R. Sabat, unpublished data). The expression of IL-22 was approximately 10 times higher than that of IFN-γ (105). Interestingly, we also found IL-22 in the blood of patients with psoriasis (in contrast to IFN-γ), and the level of it in the blood correlated with the severity of disease (106). IL-22 is a member of the IL-10 IFN family (107). It can generally be produced by activated T and NK cells, but not by other immune cells or tissue cells (105,108). The main source of IL-22 in psoriatic lesions should be the effector/memory Th17 and Th1 cells (108–110). As soon as the late 1990s, it was determined that the majority of the CD4+ and CD8+ T-cell clones derived from lesional psoriatic skin expressed IL-17 mRNA, suggesting that skin-infiltrating T cells produce this cytokine (111). In fact, Chan et al. very recently demonstrated elevated IL-17 expression in psoriatic skin lesions (112). Our measurements showed that this elevated expression was in the same magnitude as IL-22 (104,105) (R. Sabat, unpublished data). All these data suggest that the majority of CD4+ T cells in lesional psoriatic skin may be IL-22 and IL-17-producing T cells. Interestingly, Cargill et al. very recently found that single-nucleotide polymorphisms in the IL-23 receptor gene are associated with psoriasis, supporting the role of Th17 T cells in psoriasis pathogenesis (113).
Since more than 15 years, it has been known that the number of macrophages is increased in psoriatic skin lesions (114–116). We recently found that the numbers of macrophages (CD68+) and T cells in diseased skin are very similar (R. Sabat, unpublished data). The number of DCs in psoriatic skin lesions is apparently only about half the amount of T cells (117) (R. Sabat, unpublished data). In normal skin, two populations of DCs in particular can be found; namely the Langerhans cells that are present in the epidermis and the dermal DCs that are located in the dermis (118). Psoriatic lesional skin contains unaltered numbers of Langerhans cells (119) and at least two other DC populations. In contrast to normal skin, there are also inflammatory dendritic epidermal cells (CD11c+/CD1a+/Lin−/CD123−) and small amounts of plasmacytoid DCs (CD11c−/CD1a−/ Lin−/CD123+) (117,119,120). It should also be mentioned that a high number of CD11c+ APCs can be found in diseased skin. This population consists of DCs (89) and probably mainly macrophages (121).
In psoriatic lesions, the amount of NK T cells is also increased in comparison with normal skin (122). NK T cells are a separate line of T cells, which are known for (i) their expression of NK-cell receptors, (ii) their extremely restricted TCR repertoire (mostly Vα24 and Vβ11) and (iii) their activation by glycolipids presented via CD1d (123). NK T cells can be CD4+ or CD4-/CD8. Approximately 5% of the cells infiltrated in the psoriatic lesions are NK cells (95).
Effector phase – immune cell activation in the skin
The immune cells may be activated after migration into the skin. In the dermis and epidermis, different APC populations such as macrophages and various types of DCs can stimulate T cells, and vice versa. T cells that have migrated into the epidermis may additionally be activated by keratinocytes. T cells may proliferate as a consequence of activation in the skin (124,125). Psoriatic lesions even demonstrate some characteristics of lymph nodes. The activation of T cells in the skin can be negatively influenced by resistant intraepithelial T cells and regulatory T cells.
Already in 1994, Nestle et al. demonstrated that DCs from psoriatic plaques were much more effective stimulators of spontaneous T-cell proliferation than blood-derived DCs or DCs from skin of healthy individuals. Antibody-blocking studies demonstrated involvement of HLA-DR, CD28 ligands and LFA-1 in this stimulation (126). New studies hinted at an important role of plasmacytoid DCs during the development of psoriasis. In a xenograft model of human psoriasis, blocking IFN-α signalling or inhibiting the ability of plasmacytoid DCs to produce IFN-α prevented the T-cell proliferation and development of psoriasis (127). Interestingly, Eriksen et al. very recently showed that psoriatic T cells have increased and prolonged responses to IFN-α with respect to the level of STAT activation, when compared with infiltrating T cells from the skin of non-psoriatic donors. This increased IFN-α-induced signalling led to an enhanced binding of STAT4 to the IFN-γ promoter and an enhanced IFN-γ production, as well as to inhibition of T-cell growth (128).
An important question for the understanding of the psoriasis pathogenesis is how monocytes/macrophages and DCs activate the immigrated T cells in the skin. We assume that cytokines (in particular IL-23 and IL-6) produced by these APCs play a dominate role. In fact, myeloid DCs (CD83+) and in particular monocytes/macrophages (CD68+) separated from psoriatic skin lesions expressed IL-23, and IL-23 (a Th17-cell-inducing cytokine) expression was increased in psoriasis. In contrast, IL-12 (a Th1-cell-inducing cytokine) expression was not upregulated (121). Another question is whether a certain antigen is required for the activation of T cells in the skin, and if so which one.
Some studies propose a mechanism for the reactivation of psoriasis-relevant T cells that were primarily directed against streptococcal components. Although some spreading of streptococcal components into the skin has been reported (74), these Ags do not seem to persist in psoriatic lesions and do not seem to play a role in the activation of the immigrated T cells. This is also supported by the fact that antibiotic treatment or tonsillectomy seldom improves the course of the psoriasis (129,130). An attractive basis for the resolution of this discrepancy may be the concept of cross-reactivity of the T cells, primarily directed against streptococcal Ags, with auto-Ags. This may be possible in principle as a certain number of potentially autoreactive T cells seem to persist in the organism despite the negative T-cell selection process in the thymus. The model of cross-reactivity requires that one or several microbial structures are shared by structures of the human tissue which is, in the case of psoriasis, the skin [molecular mimicry (131)].
Indeed, initial database searches identified high structural similarities between streptococcal M proteins, which are major streptococcal virulence factors, and type I keratins (132). Both types of proteins share the α-helical coiled-structure formed by hepta-peptide repeat patterns. More recently, several studies demonstrated responses of peripheral blood T cells from psoriatic patients but not healthy participants to several synthetic peptides corresponding to shared sequence motifs present in M proteins and type I keratins. One example is sequences comprising the ALEEAN motif common in both streptococcal M6 and K17 (133,134). Among them, Gudmundsdottir et al., using the 146-K17 peptide, demonstrated that out of 17 patients and 17 control persons tested, 13 patients responded to this peptide compared with four controls, and the responses were significantly stronger in the patient group (133).
Although all type I keratins share significant sequence similarity, K17 seem to be a preferred candidate for the psoriasis relevant auto-Ag. The auto-Ag is expected to be present in skin (and may be in the synovial compartments), but not in uninvolved tissues. Indeed, K17 has been shown to be present in skin, but not in buccal mucosa. Under normal conditions, however, cutaneous K17 expression is missing except for the hear follicles, nail beds, sweat glands and sebaceous glands. Its expression is induced in suprabasal keratinocytes in psoriatic lesions (135,136). In vitro, K17 expression could be induced in the keratinocytes HaCat cell line by IFN-γ, and to a lesser extent, by TGF-α, whereas other pro-inflammatory cytokines such as IL-1, IL-6, IL-8 and IL-18 had no inducing effect (137,138). Regarding the expressional regulation of K17, one would argue that the K17 expression is not the cause but rather a secondary effect induced by activated T cells. However, there may be one argument supporting a primary involvement of K17: psoriasis often starts with the affection of the scalp, where K17 is physiologically present.
In the epidermis, migrated T cells can also be activated by keratinocytes. The keratinocytes have been shown to express MHC class II molecules and CD54 after exposure to IFN-γ (139). Such activated keratinocytes were able to induce T-cell proliferation by bacterial-derived super-Ags (such as staphylococcal enterotoxin A and B), which could be significantly and partially inhibited by mAbs against LFA-1 and by mAb against MHC class II, respectively, but not by mAbs against the CD28 ligands (140). Interestingly, these IFN-γ-treated keratinocytes are not able to support T-cell proliferation to alloantigens (140). This can be taken to mean that such cells impair antigen presentation via MHC II. However, results from Griffiths et al. showed that the presence of intra-epidermal lymphocytes was not correlated with keratinocyte HLA-DR expression (88). Additionally, the restricted clonality of the CLA+ T cells that are present within the lesional skin in psoriatic patients (73), as well as the disease’s familiarity coupled to certain prevailing MHC types argues against the involvement of super-Ags and for the involvement of classical antigenic proteins of the streptococci. It should be noted, however, that bacterial super-Ags possibly contribute to skin inflammation through direct activation of keratinocytes, probably by binding to MHC II (141).
Keratinocytes may also play a role in the activation of NK T cells through CD1d. CD1d is expressed by keratinocytes in normal skin at a relatively low-level and confined to upper-level keratinocytes immediately beneath the lipid-rich cornified layer. In psoriatic plaques, there is an overexpression of CD1d. CD1d could also be rapidly induced on keratinocytes in normal skin by physical trauma that disrupted barrier function. Keratinocytes also displayed enhanced CD1d following exposure to IFN-γin vitro. Combining CD1d+ keratinocytes with human NK T-cell clones resulted in the clustering of NK-T cells, and, while no significant proliferation ensued, NK T cells became activated to produce large amounts of IFN-γ (142).
The activation of T cells in the skin should be inhibited by two subpopulations of T cells. In normal mouse skin, and also in low amounts in human skin, intraepithelial T cells can be found. The biology of these cells is poorly understood. They express the γδ TCR and often show less TCR variability, in comparison with T cells that express the αβ TCR. This distinction is especially evident in mice and several other mammals as well as in humans albeit to a lesser degree. Recent studies in laboratory animals have indicated the capacity of γδ T cells to play major roles in the maintenance of the epidermal barrier, the regulation of cutaneous inflammation, and the protection against cutaneous neoplasms (143). Interestingly, the intraepithelial T cells are able to inhibit αβ T cells and prevent dermatitis in the mouse model (144). In his presentation at the 4th International Congress ‘Psoriasis from Gene to Clinic’ in London in 2005, Hayday hypothesized that following a chronic activation of γδ T cells in psoriasis, the cells become ‘exhausted’ which then leads to a lack of inhibition of αβ T cells in the skin. The second subpopulation of T cells that can inhibit Th17 and Th1 cells is the regulatory T cells. New data suggests that these cells are dysfunctional in patients with psoriasis. In fact, regulatory CD4+ T cells from peripheral blood from patients with psoriasis have been shown to be deficient in their suppressor activity (145). In this study, regulatory T cells were also isolated from the site of inflammation, the psoriatic plaques, and were analysed. At calculated ratios of regulatory T cells to effector T cells found to be present in the skin, the psoriatic regulatory T-cell population demonstrated decreased suppression of effector T cells.
The skin-infiltrating monocytes/macrophages and DCs are apparently activated by IFN-γ produced by T, NK T and NK cells, and possibly by heat shock protein produced by keratinocytes. These cells then may begin to produce TNF-α, IL-6, IL-18, IL-19, IL-20 and IL-23. Interestingly, the biological activity of IL-1 is also not enhanced in psoriatic lesions. Dermal macrophages in the papillary dermis could be the main source of TNF-α in psoriatic skin. After staining sections of a psoriatic lesion, Nickoloff et al. found also TNF-α concentrated in keratinocytes and in intra-epidermal Langerhans cells, although it was not found after staining of endothelial cells, mast cells or dermal DCs (146). In a recent study, Lowes et al. showed TNF-α staining in CD11+ APCs (89). As already noted above, this cell population probably comprises DCs and, in particular, macrophages. It should also be mentioned that some other manuscripts describe the keratinocytes as a mention-worthy cellular source of TNF-α in psoriatic lesions (75,147).
Effector phase – keratinocyte response
Keratinocytes can be activated during the initiation of psoriatic lesions mainly by mediators produced by T1 cells (IFN-γ and IL-22). However, we postulated that over time the mediators of Th17 cells (IL-6, IL-17 and IL-22), followed by those of macrophages and DCs (TNF-α, IL-6, IL-18, IL-19 and IL-20) (Wolk et al., unpublished data) as well as lastly mediators produced by keratinocytes such as TGF-α, nerve growth factor (NGF), IL-19, IL-20, and by stromal cells in the dermis such as keratinocyte growth factor (KGF), insulin-like growth factor 1, and fibroblast growth factor 10, become increasingly important (Fig. 5). Activation of keratinocytes leads to (i) increased proliferation of these cells and (ii) alteration of their maturation. Moreover, activated keratinocytes produce numerous varied mediators that can cause further immigration of immune cells, activate stromal cells in the dermis, and induce angiogenesis.
The keratinocyte proliferation in psoriatic lesions is raised almost 50-fold. Until today, it has not been possible to identify the mediator(s), which is clearly responsible for this massive increase. Immune cells could partly be responsible for the proliferation. Hancock et al. could show that activated and non-activated T cells release factors that could increase keratinocyte proliferation (148). The same group also found that suppressive molecules were produced preferentially by monocyte cultures. Bata-Csorgo et al. demonstrated that CD4+ T cells, cloned from lesional psoriatic skin and stimulated by immobilized anti-CD3 plus fibronectin, promoted psoriatic uninvolved keratinocyte proliferation via soluble factors (149). The search for the T-cell mediator that causes this development has been disappointing to date. Of the T-cell-produced mediators that have been investigated, mediators such as IFN-γ and TGF-β appear to inhibit the proliferation of keratinocytes and others appear to have no effect on proliferation or they increased proliferation of keratinocytes at only high concentrations (IL-6, IL-8) (148,150). As has been documented in many studies, the strong inhibitory impact of IFN-γ on keratinocyte proliferation suggests that this cytokine does not play a role in any of the marked changes in the keratinocytes in psoriasis. At this point, it should be interjected that a discrepancy between the presence of IFN-γ in psoriatic lesions and the increased proliferation of keratinocytes does not necessarily exist. For this discrepancy, there may be two explanations. First, the expression of IFN-γ in psoriatic lesions is only slightly increased in contrast to other cytokines (104,105). Second, the psoriatic keratinocytes show abnormal signalling in the IFN-γ pathway. Supporting the second explanation, Jackson et al. demonstrated that psoriatic keratinocytes showed a reduced induction of IRF-1 and STAT1α activation after stimulation with IFN-γ, compared with normal keratinocytes (151). Nevertheless, IFN-γ can play a role in psoriasis, particularly in the early stages of the effector phase: (i) by increasing the immigration of immune cells into the skin (by induction of numerous chemokines in keratinocytes such as CCL2, CXCL2, CXCL9, CXCL10 and CXCL11), (ii) by activating monocytes, macrophages, DCs and endothelial cells, and (iii) by upregulating MHC I, MHC II, CD1d and CD54 in keratinocytes.
Coming back to the increased proliferation of keratinocytes, it was interesting that products from neutrophilic granulocytes were identified that could be responsible for the increased proliferation of keratinocytes. Human leucocyte elastase induced proliferation in murine keratinocytes in concentrations that can be found on the skin surface of psoriatic lesions (152). Moreover, daily topical application of human leucocyte elastase on hairless mice induced a concentration-dependent epidermal hyperproliferation. Histologic analysis revealed marked vasodilation but, interestingly, no inflammatory infiltrates (152). However, anti-CXCL8 therapy in psoriatic patients significantly reduced the amount of neutrophilic granulocytes in the diseased skin, but had no impact on the macroscopic skin alteration (153). This contradicts an important role of neutrophilic granulocytes in the induction of the keratinocyte response.
The impressive therapeutic effects of the neutralization of TNF-α in psoriasis let us assume that this cytokine strongly, although indirectly, increases the proliferation of keratinocytes. As mentioned previously, macrophages in the papillary dermis could be the main source of TNF-α in the psoriatic skin (146). If this is the case, macrophages appear to be the cell type mainly responsible for the psoriatic skin alterations. Interestingly, two independent studies very recently demonstrated in two different murine models that activated macrophages are essential for chronic psoriasis-like skin inflammation (154,155). TGF-α could play a role as an intermediate in the TNF-α-elicited increase of keratinocyte proliferation. Keratinocytes themselves produce TGF-α and this mediator even induces its own gene expression (156). TNF-α has been shown to induce TGF-α in keratinocytes (157) and TGF-α messenger RNA and protein were much more abundant in lesional psoriatic epidermis than in normal-appearing skin of psoriatic patients or in normal epidermis (158,159). TGF-α is known to bind to the epidermal growth factor receptor and to stimulate proliferation of keratinocytes and accelerates epidermal regeneration (160). In contrast to TGF-α, mRNA levels of TGF-β1, which inhibits epithelial cell growth (161), are not significantly different in normal, uninvolved, and lesional psoriatic epidermis (158). Apparently, during the course of a psoriatic lesion, TGF-α and other factors, which are themselves produced by keratinocytes, play a role in keratinocyte proliferation.
NGF may be one such mediator. NGF is able to increase the proliferation of keratinocytes through high-affinity NGF receptors (trk) (162,163). Even if the expression of this receptor appeared to be decreased by keratinocytes of non-lesional and lesional skin in patients with psoriasis (164), it is still possible that NGF contributes to the increased proliferation of these cells. Two facts speak in favour of this: (i) NGF is expressed at high levels by keratinocytes in lesional and non-lesional skin (165), and (ii) high-affinity nerve growth factor receptor blockers improve psoriasis-like skin alterations in the severe combined immunodeficient mouse–human skin model (166). At this point, it must be emphasized that the effects of neuropeptides such as NGF are not limited to their impact on the proliferation of keratinocytes in psoriasis, but can also influence the angiogenesis, T-cell activation and proliferation of cutaneous nerve cells.
Also, the major mediators that lead to altered maturation of keratinocytes in psoriasis have not yet been clearly identified. In normal skin, the maturation of keratinocytes from the basal layer to the cornified layer takes approximately 28 days. The last step of this maturation process is the terminal differentiation. This is a particular apoptotic process that results in the formation of the mechanically resistant cornified layer (167). The manifold morphological changes that take place during the transition from granular to cornified cells include the synthesis of numerous proteins that are encoded by the epidermal differentiation complex on chromosome 1 (1q21), the aggregation of the keratin intermediate filament network into macrofibrils, the synthesis of extracellular lipids, and the dissolution of the nucleus and other organelles (168). In psoriatic lesions, the maturation is shortened to 5 days. This shortened maturation is associated with massively disrupted terminal differentiation of keratinocytes and is mainly reflected by parakeratosis. Psoriatic lesions are also characterized by the absence of a granular layer. Some markers, specific for granular layer in normal skin such as involucrin and transglutaminase are expressed in the spinous layer, while other granular markers such as filaggrin, are either absent or found in the parakeratotic scales (169). Additionally, the expression of proteins such as corneodesmosin, that is present in normal skin in the cornified layer, is significantly increased and also observed in the spinous layer in psoriatic lesions (170). Furthermore, there is apparently a downregulation of keratin K1 and K10, which are typically necessary for terminal differentiation. Moreover, the production of extracellular lipids is reduced and finally the retention of partially differentiated keratinocytes causes typical psoriatic scales.
We assume that IL-20 and IL-22 are the key mediators that alter the terminal differentiation of psoriatic keratinocytes. As stated above, the main source of IL-22 in psoriatic lesions should be the effector/memory Th17 and Th1 cells (108–110). IL-22 affects tissue cells and not immune cells (105,108). Keratinocytes are one of the most important targets of these cytokines (105). IL-22 operates through a receptor complex, which is composed of IL-22R1 and IL-10R2 (171). IL-20 can affect keratinocytes via two different receptor complexes (IL-20R1/IL-20R2 and IL-22R1/IL-20R2). This sharing of identical receptors chains may not be associated with binding competition or mutual limitation of biological effects of the different mediators (172). IL-22 regulates three functions of keratinocytes: (i) production of antimicrobial proteins, (ii) differentiation and (iii) migration (105,106,173).
In 2004, we could show that IL-22 strongly increases the expression of the antimicrobially acting β-defensin 2 and β-defensin 3 (105). It was the first discovered effect of IL-22 on keratinocytes. More recently, we were able to show that IL-22 also induces the expression of S100A7, S100A8 and S100A9 (106). The modes of action of β-defensins and S100A proteins are different; S100A proteins act through zinc sequestration, whereas β-defensins destabilize the microbial membrane (174). This use of different mechanisms suggests that IL-22 kills microbial pathogens very efficiently and is therefore a very potent player in the antimicrobial defense of the epidermis. Despite the troubled integrity of the epidermis, cutaneous infections are not common in psoriasis. Interestingly, the genes encoding S100A7, S100A8 and S100A9 are located inside the epidermal differentiation complex.
More importantly for the psoriasis pathogenesis, IL-22 regulates the terminal differentiation of keratinocytes. In fact, our data showed that IL-22 reduced the expression of profilaggrin (FLG), K1, K10, calmodulin-like 5 (CALML5), keratinocyte differentiation-associated protein (KDAP) and kallikrein 7 (KLK7) (106). Profilaggrin (FLG) is known to be processed during terminal differentiation by several proteases providing the N-terminal peptide and several copies of filaggrin. The N-terminal peptide translocates into the nucleus where it contributes to the nuclear dissolution (175). Filaggrin binds to the cytoplasmic keratin intermediate filament network and aggregates it into macrofibrils (176). The keratins contained in the macrofibrils are K1 and K10. Calmodulin-like 5 is a calcium-binding protein that interacts with transglutaminase 3, a key enzyme in the terminal differentiation of keratinocytes (177). Transglutaminases are known to crosslink many proteins in the cornified layer (apart from keratins and filaggrin, these include loricrin and involucrin), and contribute to the formation of the cornified cell envelope (178,179). A serine protease involved in the physiological detachment of corneocytes from the stratum corneum is KLK7. It cleaves two adhesive proteins from the extracellular part of corneodesmosomes, namely corneodesmosin and DSC1 (180). The degradation of these proteins at the epidermal surface is necessary for the physiological desquamation.
The last group of IL-22 sensitive genes is comprised of genes that encode proteins responsible for cellular migration. IL-22 enhanced the expression of MMP1 and MMP3, and reduced the expression of DSC1 (106). Metalloproteinases constitute a family of structurally related zinc-dependent neutral endopeptidases, which degrade the extracellular matrix (181). MMP1 and MMP3 are inducible and secreted, and their expressions are upregulated in psoriatic lesions. MMP1 cleaves fibrillar type I collagen and is needed to initiate keratinocyte migration. MMP3, a protease with a wide range of substrate specificities, degrades old multicellular actin networks thereby playing a role not only in wound contraction but also in immune cell infiltration of the skin (182). DSC1 is an adhesive desmosomal protein.
Interestingly, two groups very recently described that repeated cutaneous IL-23 application induced epidermal hyperplasia with parakeratosis in mice (112,183). Moreover, Zheng et al. demonstrated that these alterations were partially dependent on IL-22 (183).
We postulate that IL-20 (and possibly IL-19 as well) has effects similar to those of IL-22 because it acts through a similar receptor (see above). We have demonstrated that IL-20 is expressed in vitro in particular not only by keratinocytes stimulated with pro-inflammatory cytokines but also by activated monocytes (104,108). This is in line with in situ hybridization data from the Kragballe group that found IL-20 particularly in the suprapapillary epidermis in psoriatic plaques close to macrophages (184). IL-20 affects keratinocytes among other cellular targets (104,108). Interestingly, TNF-α increases the sensitivity of keratinocytes to IL-20 (185). Single-nucleotide polymorphisms of the IL-20 gene are associated with psoriasis (186,187). Even if the effects of IL-20 on keratinocytes have not yet been described in detail, the observations that (i) IL-20 is strongly expressed in psoriatic skin (104,184,188,189) and (ii) overexpression of IL-20 in transgenic mice causes skin thickening with aberrant differentiation, let us assume that IL-20 has an important role in psoriasis (185,190). Interestingly, the skin aberrations in IL-20 transgenic mice are without immune cell infiltration. This lets us postulate that IL-20 could be a distal mediator of psoriasis pathogenesis.
IL-20 and IL-22 activate STAT3 in keratinocytes (104,106). Here it may be interesting to mention that mice that transgenically express a constitutively active STAT3 variant in keratinocytes have been shown to develop psoriasis-like skin alterations (191). Furthermore, a transgenic mice strain engineered with a deleted STAT3 gene in keratinocytes developed relatively normal skin and hair follicles, although the hair cycle and wound healing were severely compromised (191).
Other characteristics of psoriatic lesions are the dilation and increased number of dermal blood vessels. During dilation, complement products and TNF-α are allowed to play an important role. Angiogenesis is probably dependent on vascular endothelial growth factor (VEGF) and angioprotein 2. VEGF is a selective endothelial cell mitogen that also enhances microvascular permeability. TGF-α induced VEGF mRNA expression in cultured epidermal keratinocytes. Moreover, the hyperplastic epidermis of psoriatic skin expresses strikingly increased amounts of VEGF, and two VEGF receptors, kdr and flt-1, are overexpressed by papillary dermal microvascular endothelial cells (192). Angiopoietin 1, angiopoietin 2 and their receptor (Tie2) are also upregulated in involved psoriatic skin compared with uninvolved psoriatic skin and healthy skin. Angiopoietin 1 is expressed by stromal cells in the highly vascularized papillary dermis and angiopoietin 2 is expressed by endothelial cells in the vicinity of the proliferating epidermis that abundantly expressed VEGF. VEGF was shown to enhance angiopoietin 2 and Tie2 expression in dermal microvascular endothelial cell cultures (193). Angiopoietin 1 induces Tie2 signalling as a receptor activator and maintains blood vessel formation, whereas angiopoietin 2 destabilizes vessels by blocking Tie2 signalling as an antagonist of angiopoietin 1 and acts with VEGF to initiate angiogenesis. In addition to keratinocytes and stromal cells of the dermis, macrophages could also play a certain role in angiogenesis. By the release of proteases, growth factors (basic fibroblast growth factor, GM-CSF, TGF-α, IGF-I, PDGF and VEGF), and other cytokines, activated macrophages have the capability to influence each phase of the angiogenic process (194). Apparently, the direct effect of T cells on the new angiogenesis is minimal. Actually, all IFNs dose- and time-dependently inhibited the proliferation of endothelial cells in vitro (195).
In this review, we postulate that psoriasis is an immunologically induced, overshot, regeneration-like reaction of the skin in which different cells play a dominant role at different stages. According to this hypothesis, the pathogenesis of psoriasis can be basically subdivided into three phases: (i) the sensitization phase, (ii) the silent phase and (iii) the effector phase (Fig. 2). In the first phase, specific effector Th17 and Th1 cells evolve from naïve T cells under the influence of DCs in secondary lymphatic organs such as the lymph nodes or tonsils (Fig. 3). In this process, DCs should play a dominant role because they do not only determine the specificity and the homing receptors but also the characteristic functional phenotype of effector T cells that determines their future action. As long as the sensitization phase is not associated with an infection, it is clinically unnoticeable and not characterized by any skin alterations. Afterwards, a silent phase of variable duration occurs. The third phase begins with the skin infiltration of various immune cells (monocytes/macrophages, various subpopulations of DCs, various subpopulations of T cells and neutrophilic granulocytes) (Fig. 4). The infiltration is made easier by the fact that endothelial cells of the dermal blood vessels in uninvolved skin from psoriasis patients already express P- and E-selectin. Immigration is apparently initiated by banal trauma or by invasion of a few microorganisms (without clinical manifestation of infection) that leads to the activation of local tissue macrophages, DCs and mast cells. Finally, through the products of these cells and complement components, endothelial cells are activated, which then actively support the infiltration of immune cells into the skin. Later in this phase, the immigrated immune cells activate each other so that the T cells stimulate monocytes/macrophages and DCs through IFN-γ and these cells then activate the T cells through IL-23 and IL-6. At the same time, keratinocytes are activated whereby they actively support the infiltration of further immune cells through the production of chemokines. As this phase continues, the biology of the keratinocytes changes resulting in a massively increased proliferation of these cells and their altered terminal differentiation. Apparently, these changes are not induced by IFN-γ but rather by other T-cell cytokines (e.g. IL-22, IL-17) and by mediators of macrophages/DCs (e.g. TNF-α, IL-6). Still later in this phase, the mediators of macrophages/DCs as well as the keratinocyte autocrine mediators (e.g. IL-20, TGF-α) and stromal cell mediators of the dermis gain dominance in the activation of keratinocytes (Fig. 5). In conclusion, the reaction of the keratinocytes mimics an overshot wound-healing process. This model of psoriasis pathogenesis is summarized in Fig. 6. A successful therapy transposes the effector phase into the silent phase. After a while, a renewed effector phase follows (Fig. 1). Without an adequate treatment existing lesions are likely to persist permanently in most cases.
We assume that two conditions must be met in order to bring about the onset of psoriasis: the presence of activated immune cells with a specific function/phenotype in the skin, and the hyper-reactive dysfunction of keratinocytes. The presence of activated immune cells in the skin is presumed to be dependent on the following factors: (i) the generation of specific T1 and Th17 cells with skin-homing receptors, (ii) the increased readiness of the endothelial cells of dermal blood vessels in uninvolved skin to support the infiltration of immune cells in the skin, (iii) the deficiency of mechanisms in involved skin to inhibit the activation of immune cells. The preferred generation of T1 and Th17 cells is probably contingent upon the orientation of the immune system. In fact, patients with psoriasis also commonly suffer from other inflammatory diseases (e.g. rheumatoid arthritis, colitis). The increased readiness of the endothelial cells of uninvolved skin could be because of a genetic defect of these cells and/or could be because of a hyper-reactivity of the tissue-localized immune cells such as macrophages, DCs and mast cells that create a permanent subinflammation condition in the skin without adequate cause. The lack of mechanisms to inhibit immune cells in involved skin could either be due to a functional deficiency of regulatory cells (such as regulatory T cells or intraepithelial T cells), a hyper-reactivity of the immune cells (that may be due to a reduced effectiveness of intracellular negative regulators of cellular activation), and/or to a hyper-reactivity of the keratinocytes and other tissue cells. It is most probable that more than one of these causes exists simultaneously. Some of them have been experimentally confirmed (such as the deficiency of regulatory cells and the hyper-reactivity of DCs in psoriatic lesions). We are convinced that a hyper-reactive dysfunction of keratinocytes also exists and is a prerequisite for the initiation and maintenance of psoriasis. This opinion is based on the existence of macroscopic (clinical) as well as microscopic (histological) differences between chronic psoriasis and skin manifestations of chronic graft-versus-host disease (GVHD). In both diseases, the first phase (the sensitization phase) and the beginning of third phase (the effector phase) of our model seem to be similar (generation of specific effector T cells, immune cell skin infiltration, production of IFN-γ by T1 cells). However, the results of the third phase are completely different. In GVHD, the skin does not show hyper-proliferation of keratinocytes or enhanced desquamation. This means that although the beginning of the immunological reaction may be similar, the keratinocytes react differently in these two diseases. This may be contingent upon genetically determined reaction patterns of keratinocytes that should be different in both diseases. Moreover, some gene products from the psoriasis-susceptibility loci regulate the differentiation of keratinocytes (such as the genes from epidermal differentiation complex in PSORS4) (8).
Our model of psoriasis explains the differences between the observed therapeutic successes of various immune treatments. It should be taken into account in which phase of the pathogenesis these patients find themselves when they are being treated. According to our model, patients who already have existing macroscopic skin alterations are currently in the last stage of the effector phase: in the ‘keratinocyte response’ stage. Consistent with this, therapies that influence the beginning of the immune reaction have a weak-efficiency over a short period. These are applications that have the goals of (i) preventing the generation of effector cells (CTLA-4-Ig), or (ii) inducing the emergence of oppositionally polarized effector T cells (IL-10, and IL-4). Such therapies can be useful only after a long treatment, if indeed the already existing ‘pathological’ effector T cells are no longer present. According to our model, also therapies that influence the first stage of the effector phase (skin infiltration of T and other immune cells) should have a weak efficiency over a short period (196). In fact, despite numerous attempts, no psoriasis therapy with medications that interact with chemokines has been successful (197).
We postulate that therapies that act on the ‘keratinocyte response’ stage will have the greatest success already after a short treatment. According to our model, at the beginning of this stage the ‘keratinocyte response’ is induced by T cells and macrophages/DCs. As it continues, the macrophages/DCs and tissue cells are the main driving force. Consistent with this, inhibition of T-cell function or T-cell depletion is only successful in a portion of the patients (198). In contrast, anti-TNF-α therapy appears to cause an improvement in the skin alterations of all patients (198). The anti-TNF-α therapy is apparently so effective for two reasons; it leads to partial depletion of macrophages/DCs and it neutralizes TNF-α, a cytokine that strengthens many inflammatory processes of this immunologically induced, overshot, regeneration-like reaction.
In the coming years, research will tell us to what extent our model corresponds to the real pathogenesis of psoriasis.
We would like to thank Dr Wolf-Dietrich Döcke for very helpful discussions and Sascha Rutz for the assistance in the figure production. We also thank the German Ministry of Education and Research (Bundesministerium für Bildung and Forschung) for their generous support.