Resolution of plaque-type psoriasis: what is left behind (and reinitiates the disease)

Psoriasis is a chronic skin disease that results from a multifaceted interaction of immunological, environmental, and genetic factors [1, 2]. According to a recent systematic review, the global prevalence of psoriasis ranges from 0.09 to 5.1% [3]. The most common form of psoriasis, occurring in 85-90% of all patients, is plaque-type psoriasis characterized by well-demarcated, scaly and erythematous, infiltrated plaques [1, 4]. On the microscopic level, keratinocytes proliferate rapidly at high turnover rates, which leads to incomplete terminal differentiation. Abnormal differentiation causes thickening of the epidermis (acanthosis), retention of keratinocyte nuclei in the stratum corneum (parakeratosis), and loss of the granular layer. Neutrophils accumulate in the epidermis to form Munro’s microabscesses. Psoriatic lesions are highly vascular and also densely infiltrated by T cells and dendritic cells (DC) [2, 5, 6] which are key players in its pathophysiology.

Various external factors like trauma and injury, infection, or medication can stress or damage keratinocytes [4]. Stressed or dying cells release nucleic acids along with the antimicrobial peptide (AMP) LL-37. LL-37 causes the loss of tolerance to self-nucleic acids and forms complexes with self-DNA/-RNA. These complexes can activate plasmacytoid dendritic cells (pDCs) via toll-like receptor 7 (TLR7) [7]. pDCs mainly produce type 1 interferons (IFN-α, IFN-β), which then activate myeloid dendritic cells (mDCs) to produce key psoriatic inflammatory cytokines IL-12 and IL-23 [8]. mDCs can also be directly activated by nucleic acid-LL37 complexes via TLR8 [9] and then migrate to lymph nodes where they promote the differentiation and activation of T cells via TNFα, IL-12, and IL-23. Activated T cells enter the circulation and move to inflamed skin via adhesion molecules on the endothelial cells of blood vessels. At the site of inflammation, different subsets of T cells release their effector molecules IFN-γ, IL-17, and IL-22. These cytokines, in concert with the pro-inflammatory signals such as IL-23 produced by mDCs, act on keratinocytes. Keratinocytes produce AMPs, cytokines, and chemokines to attract other immune cells like neutrophils and macrophages (Fig. 1) [2, 5, 10].

In chronic psoriatic lesions, mDCs produce IL-12 and IL-23, which in turn constantly stimulate T cell subsets Th1, Th22, and Th17 to release IFNγ, IL-22, and IL-17. CD8+ T cells are also present in psoriasis, produce the same range of cytokines as CD4+ cells, and reside predominantly within the epidermis [5, 10].Over the last decade, it has been shown that the main source of IL-17 in psoriasis is not T cells, but rather innate cells like neutrophils and mast cells [11‐14]. Neutrophils, for example, are a rich source of IL-17, AMPs, and elastase, all of which help keratinocytes hyperproliferate. Recent studies also implicate innate lymphoid cells (ILCs) as an additional source of IL-17 in psoriasis [15, 16]. IL-17 acts directly on keratinocytes, the main IL-17 receptor (IL-17R)-expressing cell type, but has only limited effects on their gene expression [17]. Synergism of IL-17 and TNFα is key to the full-blown effect on mRNA expression of pro-psoriatic genes (such as DEFB4, S100A7, IL19, IL17C, CXCL8, CCL20, LCN2) in keratinocytes [18]. In psoriasis, a combination of the inflammatory cytokines IL-17, TNFα, IL-22, and IFNγ drives keratinocyte hyperproliferation and cytokine and chemokine release [5]. Keratinocyte-produced IL-1β and IL-18 act on DCs and T cells, and chemokines such as CXCL-5 and CXCL-8, as well as AMPs, constantly attract neutrophils [4]. Keratinocytes also produce vascular endothelial growth factor (VEGF), which recruits and favors proliferation of endothelial cells, thereby promoting angiogenesis and creating highly vascular psoriatic plaques (Fig. 2) [10]. Together, structural abnormalities of the epidermis, inflammatory cellular skin infiltration, and increased dermal angiogenesis result in full-blown chronic psoriasis.

A more detailed understanding of the pathophysiology of psoriasis has revolutionized anti-psoriatic treatment and led to the successful development of various new drugs targeting key inflammatory cytokines (Table 1).

Nowadays, targeted therapy, in particular with antibodies against IL-17 or IL-23, permits complete clinical clearance of skin lesions in a high percentage of cases, at rates similar to those observed after PUVA treatment [31‐35] but with an apparently more favorable risk profile. Severity of psoriasis is assessed with tools such as the psoriasis area and severity index (PASI), which scores both involved body area and clinical appearance of psoriatic lesions. In treated patients, complete clearance is observed in approximately 30 to 50% of patients and reduction in severity of symptoms by 90% (PASI90) in approximately 50 to 70% of patients [25, 27‐29]. However, if treatment is stopped, skin lesions eventually reappear within months. Conspicuously, patients treated with psoralen plus UVA (PUVA) photochemotherapy or anti-IL-23 antibodies have the longest disease-free periods. In fact, median time to recurrence was 8 months for PUVA treatment in the study by Yones et al. [32], compared to similar sustained remission times after stopping anti-IL-17 or anti-IL-23 antibody treatment. For instance, after stopping secukinumab, the median time to relapse was 20 to 24 weeks [36]; after stopping ixekizumab, 20 weeks [37]; after stopping tildrakizumab, 20 weeks (as measured by PASI75) [26]; and after stopping guselkumab, approximately 15 weeks (as measured by PASI90) [38].

As early as 1994, Vallat et al. [39] provided a possible explanation for the longer remissions obtained with phototherapy than with other treatments such as then state-of-the-art cyclosporine. They showed that bath PUVA strongly suppressed both immunological and epidermal activation in psoriasis. Keratinocyte proteins such as keratin 16, filaggrin, and involucrin that were abnormally expressed in active psoriasis were more normally expressed after PUVA treatment. PUVA strongly suppressed epidermal and dermal CD4+ and CD8+ T cell infiltration of the skin, with virtual elimination of IL-2 receptor bearing activated T cells in some patients [39]. The authors proposed that these changes could be the cellular basis for the more sustained disease remission seen after PUVA treatment versus that seen after simple immune suppression by cyclosporine. The therapeutic mechanisms of action of phototherapy in psoriasis have been recently reviewed, emphasizing phototherapy’s dual action on both players in the disease (i.e., keratinocytes and skin-infiltrating immune cells) [40‐42].

Phototherapy induces apoptosis and immunosuppression. However, it is not clear whether these two effects act simultaneously or independently [43]. In psoriasis, PUVA [44, 45] or UVB [46‐48] may act by upregulating anti-inflammatory cytokines (IL-4 and IL-10) and downregulating pro-inflammatory cytokines (IL-8, IL-17, IL-22, IL-23, TNFα, and IFNγ). In addition, UV therapy can induce regulatory T cells (Tregs) [49, 50] and Langerhans cell emigration [51, 52]. However, it has also been shown that phototherapy leads to direct apoptosis of keratinocytes. Indeed, a study using a computational model of psoriatic epidermis combined with in vivo and in vitro experimentation revealed that apoptosis occurs in both stem and transit-amplifying cells and seems to be sufficient to explain UVB-induced plaque resolution [53].

The dual action of phototherapy (and of PUVA in particular) on keratinocytes and the immune system may be responsible for the sustained resolution of psoriatic disease after cessation of treatment. Having said that, the effect of phototherapy in psoriasis is limited to the site of exposure [31, 54], which suggests a crucial direct effect on keratinocytes (i.e., apoptosis) beyond the local and systemic effects on the immune system. We speculate that a dual effect of anti-IL-17 or anti-IL-23 antibody treatment on keratinocytes and immunocytes may also be responsible for the long-lasting, sustained response seen after treatment cessation. In contrast, other anti-psoriatic treatments such as topical steroids, vitamin D₃ analogs, or anthralin [55‐57], systemic traditional treatments such as cyclosporine or methotrexate, and even first-generation biologics (i.e., TNF antagonists) are usually associated with rather quick relapses within several weeks after treatment is stopped.

Nowadays, psoriasis is considered as systemic disease often associated with comorbidities such as metabolic syndrome, hypertension, elevated blood lipid levels, and cardiovascular disease [58]. Intriguingly, psoriatic plaques appear in most patients only at certain body locations, in particular at predilection sites such as the extensor surfaces of the extremities (knee and elbows) and/or other mechanically stressed locations such as the sacral region. Even more intriguingly, despite a general genetic predisposition, more than 35 psoriatic susceptibility genes are known so far [59], certain patients develop life-long psoriasis only on fixed, circumscribed, and sometimes solitary body sites. Psoriatic lesions can resolve entirely after treatment and always without scarring. However, when treatment is stopped, lesions recur most often at the exact same body sites. This raises an important question: what kind of residual “molecular scars” are left behind on a clinically cleared psoriatic lesion that eventually cause its reappearance. What those molecular scars look like after treatment with topical and traditional systemic agents is obscured by insufficient data, but what they look like after treatment with anti-TNFα, anti-IL-17, and anti-IL-23 antibodies is clearer thanks to recent studies of skin biopsy samples.

The fact that psoriasis recurs most often at the very same body sites where the initial lesion occurred suggests that critical cells may be left behind in the skin in situ after treatment is stopped. In this regard, Cheuk and coworkers [76] reported the results of an elegant study in 2014, addressing the question whether psoriatic lesions harbor tissue-resident memory T cells (TRMs) that could possibly drive disease recurrence at the same, i.e., initial site of lesions. They sampled narrow band-UVB-treated, anti-TNFα or anti-IL-12/23-treated healed psoriatic skin and studied gene expression and ex vivo cytokine production by T cells. Interestingly, they found that approximately half of CD8+ T cells in the epidermis of active lesions expressed TRM markers CD103 and CD49a. This suggested an expansion of TRMs at diseased body sites of psoriasis compared with healthy skin. In resolved lesions, epidermal CD4+ T cells had returned to normal levels, while numbers of dermal CD4+ and CD8+ T cells were still elevated. Although epidermal CD8+ T cells were decreased, the number of CD49a-expressing CD8+ T cells was still higher than in non-lesional or healthy skin (that had not had visible psoriatic manifestation). Transcriptional analysis of isolated T cells from resolved psoriatic lesions showed that epidermal T cells had higher expression of Th17-associated genes (RORC, IL17A, IL22) than did dermal T cells. Upon ex vivo stimulation, dermal T cells from those lesions had a cytokine production profile similar to that found in healthy skin. Interestingly, the percentage of IL-22+ epidermal CD4+ T cells from resolved lesions was similar to that in active lesions. Epidermal CD8+ T cells from resolved lesions produced more IL-17A after ex vivo stimulation than did those from healthy skin, indicating that these cells are still functional even after long-term anti-psoriatic treatment. Epidermal IL-17-producing CD8+ T cells from healthy skin, lesional psoriasis, and resolved psoriasis expressed CD103, suggesting that these are in fact TRMs. In resolved lesions, an epidermal subpopulation of CD8+ T cells co-expressed TRM marker CD103 and Th17 markers IL-23R and CCR6, indicating responsiveness to IL-23 signaling. Moreover, Cheuk et al. [77] found two subsets of CD8+ TRMs in the epidermis, IL-17-producing CD49- and IFNγ-producing CD49+ TRMs. Intriguingly, IL-17-producing CD49− TRMs from the epidermis showed higher IL-17 production than did CD49− TRMs isolated from dermis. This suggested that distinct functional subsets exist in epidermis and in dermis. In active psoriasis, some CD49− TRMs as well as some CD49+ TRMs expressed both IL-17 and IFNγ. However, after isolation and ex vivo stimulation, these cells did not produce IL-17. Thus, the investigators speculated that T cell receptor (TCR) activation might be needed in addition to inflammatory signals promoting IL-17 response [77].

Most recently, Matos and co-workers [78] used immunofluorescence staining and high-throughput sequencing (HTS) of the CDR3 domain of the T cell receptor as a different approach to assess density and clonality of T cells in order to study residual T cells in resolved psoriatic lesions. TCR sequencing indicated that there were more T cells in resolved psoriatic lesions than in the skin of healthy controls and non-lesional skin of psoriasis patients; in fact, levels in resolved lesions were similar to those in active lesions. While diverse T cell clones were present in active lesions before treatment, only 7% of T cell clones remained detectable in clinically healed lesions. This suggests that, when inflammation subsides, the majority of T cell clones leave the skin. Increased numbers of oligoclonal T cell populations were found not only in resolved lesions but also in non-lesional skin. When T cells were isolated from resolved lesions to study their cytokine production ex vivo, expanded T cell clones from both active and clinically healed lesions produced IL-17A. TCR sequencing also showed that putative pathogenic T cell clones were αβT cells (implying that their TCR is composed of α- and β-chains) and not γδT cells. Furthermore, unique αβ TCR sequences that were identified in the skin of psoriasis patients were absent from the skin of healthy controls and patients with other skin diseases. These findings implied that, in psoriasis, increased T cell response might be due to the presence of a common antigen. Pathogenic T cell clones producing IL-17A and residing in low numbers in clinically healed lesions could in fact be TRMs. They may be resistant to elimination by therapy and bear the potential to restart the psoriatic inflammatory feedback loop at the site of residence [78].

In 2018, Gallais-Sérézal and colleagues studied how resolved psoriatic tissue responded to activation of putative pathogenic TRMs. Using sampled skin biopsies from healthy controls and resolved lesions from psoriasis patients, they produced skin explant cultures and then stimulated them with anti-CD3 antibody OKT-3 to activate resident T cells. The skin environment after T cell activation was analyzed by NanoString transcriptional methodology, and IFNγ-related signaling pathways were found to be upregulated in both healthy and resolved psoriatic skin. However, in resolved psoriasis alone (and not in healthy skin), IL-17 signaling was upregulated in the epidermal compartment. This implied that renewed activation of resident T cells in resolved lesions leads to keratinocyte activation, which could translate into chemokine release and recruitment of circulating leukocytes. Indeed, IL-17 signaling-related tissue-response correlated with clinical relapse after therapy, thus highlighting the key role of pathogenic resident T cells in psoriasis [79]. The same group of investigators studied TRMs in non-lesional skin from psoriasis patients to determine whether these resident T cells are indeed pathogenic and have the ability to induce psoriasis in never-lesional skin. Indeed, ex vivo T cell activation of skin explants led to psoriasis-typical tissue responses as determined by NanoString analysis. Additionally, CD103+ CD8+ T cells and CCR6+ CD4+ T cells were enriched in non-lesional skin of psoriasis patients compared with that in healthy controls. Moreover, CCR6+ CD4+ T cells were capable of producing IFNγ and IL-17. These results suggested that pathogenic TRMs do exist in non-lesional skin [80].

Although pathogenic TRMs have certainly gained more attention in recent years, some progress has been made in analyzing the presence and function of DCs and Langerhans cells in resolved psoriasis as well. While DC numbers were found to be high in active psoriatic lesions and absent from resolved skin, the data on Langerhans cells are controversial [81, 82]. Langerhans cells from active lesions expressed psoriasis-associated genes (IL23, IL1B, IL15) and showed increased IL-23 production after ex vivo stimulation. In resolved lesions, expression of IL23 and IL15 was still higher than in healthy skin. Upon ex vivo stimulation, these cells were still capable of producing IL-23 [83]. In a comprehensive review on the role of Langerhans cells in psoriasis, Eidsmo and Martini [81] hypothesized that Langerhans cells are not only found in close proximity to T cells, but are also able to cross-talk with both T cells and keratinocytes. This, together with the ability of Langerhans cells to produce IL-23, implies that Langerhans cells could play a role in renewed activation of T cells and disease recurrence.

Last but not least, Johnson-Huang and colleagues [46] showed that NB-UVB leads to clearance of psoriasis by suppressing the IL-17/IL-23 axis. They compared responders vs. non-responders after 6 weeks of NB-UVB therapy and looked at myeloid DCs and T cells and their inflammatory products using immunohistochemistry and qPCR. Inflammatory myeloid DC levels were reduced and expression of DC-related cytokines (e.g., INOS, IL20, IL23p19, IL12/IL23p40) was decreased in responsive plaques. Furthermore, CD3 T cell counts were reduced in responders, and cell counts strongly correlated with the defined “response score”. In addition, expression of IFNG, IL17, IL22, and IL-17/IL-22 downstream genes β-defensin 4 (BD4) and myxovirus-resistance 1 (MX1) was significantly suppressed. All of these changes were not observed in non-responsive plaques. Interestingly, the “response score” correlated with IL17 and IL22 expression, but not with IFNG expression. To determine whether these results show a direct effect of NB-UVB or merely changes that occur when psoriasis clears up in general, Johnson-Huang’s group performed an in vitro experiment where PBMCs were isolated, treated with NB-UVB, and measured for cytokine expression after 4 h. Intracellular staining showed that production of IFNγ was suppressed by 85% and that IL-17 and IL-22 produced by CD3+ T cells were decreased by 45% and 89%, respectively [46]. It is fascinating to learn that a nature-derived therapy such as UVB phototherapy seems to interfere with the very same pathway targeted by today’s most modern drug treatments, i.e., the IL-17/23 axis.

Numerous studies have shown that the molecular imprints of psoriasis do not fully resolve in macroscopically cleared skin even after the most efficacious targeted treatment available nowadays, including anti-IL-17 and anti-IL-23 antibody administration. Resolved lesions have a molecular scar composed of about 250 gene products that are not fully normalized after treatment. These transcripts can be grouped into inflammation-associated genes such as IL17, IL22 and others and skin structure-related genes [62]. Moreover, pathogenic memory T cells seem to persist at sites of clinically resolved psoriatic lesions. Even after long-term therapy, these cells do not lose their ability to produce IL-17A [76, 78], which can signal to keratinocytes and stimulate their proliferation [80]. Although the role of innate cells has not been studied extensively, Langerhans cells isolated from resolved psoriatic lesions can produce IL-23 after stimulation, which makes them another potential player in the restart of activation of “sleeping” TRMs [83] and possible reappearance of psoriatic lesions [76] (Fig. 3). The question now is this: what initially triggers (or retriggers) the process of psoriatic recurrence after a treatment has been stopped? Could IL-17 be stored and hidden in certain cells of the skin, perhaps mast cells, waiting to be released after endogenous and/or exogenous stimulation to activate the inflammatory loop again? What about the neurogenic inflammation of the skin [84] that plays an essential role at least in the itch of psoriasis? Might a nerve scar left behind after successful treatment contribute to the reappearance of psoriatic lesions, be it during or after continuous treatment? Last but not the least, may the interplay of the microbiome and AMPs [85, 86] help trigger psoriatic recurrence? This may be in particular important for pustular psoriasis, in which IL-1/IL-36 plays an important role in the pathogenesis [87‐89]. A vicious loop between AMPs such as cathelicidin (LL-37) and IL-36 signaling may drive psoriatic disease [90‐93]. In fact, IL-36 receptor blockade revealed promising results in pustular psoriasis [94] and may also be a therapeutic option in plaque-type psoriasis [95]. Future work will have to address all of these open questions in order to improve and advance overall treatment strategies for psoriasis and allow long-term, continuous disease control.