T helper type 1‐related molecules as well as interleukin‐15 are hyperexpressed in the skin lesions of patients with pyoderma gangrenosum

Summary

Pyoderma gangrenosum (PG) is a rare, immune‐mediated skin disease classified into the group of neutrophilic dermatoses. Although a number of studies confirmed the central role of innate immunity, only few studies have investigated the possible contributing role of acquired immunity. In particular, no reports concerning T helper type 1 (Th1) and Th2 cells are available as yet. Therefore, 15 patients with PG, five with Sweet's syndrome (SS) and nine skin specimens from healthy controls (HC) were investigated, evaluating the expression of Th1‐related markers interleukin (IL)‐12, interferon (IFN)‐γ, C‐X‐C motif chemokine receptor 3 (CXCR3) and C‐C motif chemokine receptor 5 (CCR5), of the Th2‐related molecules IL‐4, IL‐5, IL‐13 and CCR3, of the co‐stimulatory axis CD40/CD40 ligand, of IL‐15 and the natural killer (NK) cell marker CD56 in skin lesions by immunohistochemistry. Patients with PG and SS showed a higher expression of Th1 markers than HC. Conversely, IL‐5‐ and CCR3‐expressing cells were less numerous in PG skin lesions compared to SS (P = 0·0157 and < 0·0001, respectively). Both CD40 and CD40L were expressed more in PG than in SS and HC (P < 0·0001 for both). Finally, the number of IL‐15⁺ and CD56⁺ cells was higher in the skin of patients with PG than in those of SS and HC (P < 0·0001 for both). Our results suggest that Th2 cells are down‐regulated in PG. At the same time, over‐expression of the co‐stimulatory axis CD40/CD40L amplifies the impairment of the Th1/Th2 balance. Both these findings might explain the most aggressive behaviour of PG in comparison to SS. Moreover, over‐expression of IL‐15⁺ and CD56⁺ cells may suggest a possible role of NK cells in the pathogenesis of the disease.

Introduction

Pyoderma gangrenosum (PG) is a rare cutaneous disorder belonging to the family of neutrophilic dermatoses, together with Sweet syndrome (SS) and subcorneal pustular dermatosis 1. These conditions are characterized by the presence of an inflammatory infiltrate of mature polymorphonuclear leucocytes and often occur in combination with systemic diseases, cancer or after using some drugs 2.

The main variant of PG, the classic ulcerative form, is characterized by a rapidly enlarging painful ulcer with erythematous and raised edges and undermined violaceous borders 3. Histology, despite non‐pathognomonic, shows dense neutrophilic dermal infiltrates during the active expanding lesions and marked tissue necrosis with surrounding mononuclear cell infiltrates during the fully developed ulcers 4.

Although the major role of innate immunity dysfunction has been demonstrated in the pathogenesis of PG, only few studies have evaluated the possible contributory role of acquired immunity. Indeed, during the recent past, several studies have focused upon the altered role of innate immunity, leading to classification of the disease in the autoinflammatory disorders group 5, 6. The latter are a group of immune‐mediated, organ‐specific or systemic diseases, marked by the absence of pathogenic autoantibodies or autoreactive T cells 7. Several studies support this hypothesis, including the association of PG with other autoinflammatory disorders such as Crohn's disease and Behçet's disease 8, 9 and the clinical efficacy of anti‐cytokine therapy, such as anti‐tumour necrosis factor (TNF) drugs 10. Moreover, Marzano et al. recently reported mutations in some autoinflammatory genes [Mediterranean fever (MEFV), NLR family pyrin domain containing 3 (NLRP3), NLRP12, nucleotide binding oligomerization domain containing 2 (NOD2), lipin 2 (LPIN2), proline‐serine‐threonine phosphatase interacting protein 1 (PSTPIP1)] in patients with PG and its syndromic form PASH (PG, acne, suppurative hydradenitis) 11.

However, few data are reported in the literature about the role of acquired immunity and more specifically about T cell populations and their functions in PG.

Recently, interleukin (IL)‐17 and IL‐23, the two main cytokines related to the T helper type 17 (Th17) pathway, have been found over‐expressed in PG lesions, suggesting a potential role of Th17 cells in the pathogenesis of the disease 12. Furthermore, considering the increased levels of IL‐23, a therapy with ustekinumab, a monoclonal antibody that binds the shared p40 subunit of IL‐23 and IL‐12, has been proposed 13. Finally, in a previous study our group investigated the proportion of Th17 and regulatory T cells (T_regs) in PG and SS, demonstrating an imbalance between these two T cell types, with a significant reduction of T_regs in PG 14.

However, the contribution of adaptive immunity to the pathogenesis of PG seems not to be limited to the Th17 phenotype, as proved by the efficacy of several drugs such as cyclosporin and tacrolimus that act as unspecific inhibitors of the T cell receptor‐mediated signal transduction pathways 15. Accordingly, Th1 cells, that are able to stimulate neutrophil recruitment through the secretion of different cytokines and chemokines, may play an important role in the development of tissue damage in PG.

Therefore, in order to characterize further the role of T cells in the disease, in the present study we investigated the expression of cytokines and chemokine's receptors related to Th1 and Th2 pathways, as well as the co‐stimulatory molecules CD40 and CD40 ligand (CD40L), in the skin of patients with PG, including patients with SS and healthy individuals as controls (HC). Moreover, we also evaluated the expression of IL‐15, a cytokine associated with lymphocyte (particularly CD8⁺ T cell) survival and development and natural killer (NK) cell activation 16, and with an important pathogenic role in immune‐mediated gastrointestinal disorders such as coeliac disease or inflammatory bowel diseases 17, 18.

Accordingly, in order to investigate preliminarily the possible presence of NK cells in the inflammatory infiltrate of PG and SS, we evaluated the expression of CD56, a marker present among the different subsets of NK cells 19.

Materials and methods

Patients

Fifteen patients with ulcerative PG (seven males and eight females, age range = 37–69 years), five patients with SS (two males and three females, age range = 31–64 years) and nine HC (four males and five females, age range = 45–63 years) underwent excision of benign skin tumours in similar areas to those of the skin lesions of patients with PG and SS were included in the immunohistochemical study. All the PG and SS cases were idiopathic.

Inclusion criteria were the following: (i) patients at the first diagnosis of PG or SS; (ii) diagnosis established on the basis of clinical, histopathological and laboratory criteria 8; and (iii) the absence of previous topical or systemic treatment for at least 8 weeks at the time of the inclusion into the study protocol.

Exclusion criteria were the following: (i) any topical or systemic treatment in the 8 weeks before skin biopsy; and (ii) age lower than 18 years.

Upon informed consent, a 4‐mm punch biopsy from lesional skin was taken from all the patients. For patients with PG the biopsy was taken from the edge of the ulcer.

The trial was approved by the medical ethical committee of each hospital and was conducted according to the Declaration of Helsinki. All the patients and controls provided written informed consent to participate to the study.

Reagents

The monoclonal antibodies used for immunohistochemistry on cryostat cutaneous sections included: anti‐IL‐4 (1 : 20; Abcam, Cambridge, UK); anti‐IL‐5 (1 : 50; Abcam); anti‐IL‐12 (1 : 100; R&D Systems, Minneapolis, MN, USA); anti‐IL‐13 (1 : 100; Abcam); anti‐IL‐15 (1 : 300; Abcam); anti‐C‐X‐C motif chemokine receptor 3 (CXCR3) (1 : 300; R&D Systems); anti‐C‐C motif chemokine receptor 3 (CCR3) (1 : 50; R&D Systems); anti‐CCR5 (1 : 100; Abcam); anti‐interferon (IFN)‐γ (1 : 100; R&D Systems), anti‐CD40 (1 : 50; Abcam); anti‐CD40L (1 : 50; R&D Systems) and anti‐CD56 (1 : 100; ThermoFisher, San Diego, CA, USA).

Immunohistochemical analysis of the skin specimens

The expression of different Th1 (IL‐12, IFN‐g, CXCR3 and CCR5) and Th2 (IL‐4, IL‐5, IL‐13 and CCR3)‐related cytokines and chemokines, as well as that of the co‐stimulatory molecules CD40 and CD40L, of IL‐15 and of CD56, was analysed in serial sections of lesional skin biopsy specimens of 15 PG patients, five SS patients and nine HC using single immunohistochemical staining, as described previously 20.

Skin specimens were frozen immediately and stored at −80°C until use. Sequential cryostat sections (5 µm) were cut from each specimen, air‐dried and fixed in 100% acetone. The system EnVision⁺ horseradish peroxidase (HRP) system (Dako, Glostrup, Denmark) was used. According to the protocols, sections were washed in Tris‐buffered saline, coated with peroxidase block for 5 min, rinsed gently and then incubated with primary antibodies for 30 min. After washing, sections were incubated with peroxidase‐labelled polymer for 30 min, rinsed and then processed with substrate‐chromogen. Sections were then counterstained with Mayer's haematoxylin, cleared and mounted. Negative controls were subjected to the same treatment, but the passage with the primary antibody was omitted.

Two independent ‘blind’ observers evaluated the slides. For quantitative analysis, the stained cells were counted in three consecutive microscopic fields (×400), starting from the microscopic area immediately below the dermal–epidermal junction with the most intense inflammatory infiltrate. Only dermal cells were considered for the analysis.

Statistical analysis

The results are presented as medians (25th–75th percentiles). The Mann–Whitney U‐test was used to analyse the results of the immunohistochemical study. Results were considered significant with a P‐value < 0·05.

Results

Quantification of Th1 markers in PG compared to SS

To investigate the Th1 markers, we analysed the skin expression of IFN‐γ, IL‐12, CXCR3 and CCR5. The inflammatory infiltrate of both PG and SS was intensely stained by monoclonal antibodies against all markers (Fig. 1). Conversely, only sporadic positive cells were detected in HC specimens.

Figure 1.

Immunohistochemical staining for the T helper 1‐associated markers interleukin (IL)‐12, interferon (IFN)‐γ and the chemokine receptors C‐X‐C motif chemokine receptor 3 (CXCR3) and C‐C motif chemokine receptor 5 (CCR5) in skin biopsy specimens from patients with pyoderma gangrenosum (PG), Sweet syndrome (SS) and healthy controls (HC). Magnification: ×200.

Specifically, in both PG and SS, IFN‐γ, which represents the main effector molecule secreted by Th1 cells, was detected in both superficial and medium/deep dermis with some cells scattered in the epidermis (Figs 1 and 2). The number of IFN‐γ⁺ cells was similar in both groups of patients and significantly higher than in HC (P < 0·0001 for both PG and SS) (Fig. 3, Table 1).

Figure 2.

Higher magnification of representative fields showing interferon (IFN)‐γ⁺, C‐X‐C motif chemokine receptor 3 (CXCR3)⁺, interleukin (IL)‐4⁺, CD40 ligand⁺ and CD56⁺ cells in skin lesions of patients with pyoderma gangrenosum, as well as control isotype staining. Magnification: × 400.

Figure 3.

(a) Numbers of interleukin (IL)‐12⁺, interferon (IFN)‐γ⁺, C‐X‐C motif chemokine receptor 3 (CXCR3)⁺ and C‐C motif chemokine receptor 5 (CCR5)⁺ cells in the inflammatory infiltrate of skin biopsy specimens from patients with pyoderma gangrenosum (PG), Sweet syndrome (SS) and healthy controls (HC); all these T helper 1‐associated molecules were over‐expressed in comparison with HC. No significant differences were found between PG and SS. (b) Numbers of IL‐4⁺, IL‐5⁺, IL‐13⁺ and CCR3⁺ cells in the inflammatory infiltrate of skin biopsy specimens from patients with PG, SS and HC. T helper 2‐related cytokines and chemokines were over‐expressed compared to HC, while the number of IL‐5⁺ and CCR3⁺ cells was significantly lower in PG than in SS. (c) Numbers of CD40⁺ and CD40 ligand⁺ (CD40L) cells in the inflammatory infiltrate of skin biopsy specimens from patients with PG, SS and HC. Both PG and SS showed a significant over‐expression of CD40 and CD40L compared to HC. Moreover, both cell populations were significantly more numerous in PG than in SS. (d) Numbers of IL‐15⁺ and CD56⁺ cells in the inflammatory infiltrate of skin biopsy specimens from patients with PG, SS and HC; IL‐15⁺ and CD56⁺ cells were significantly more numerous in PG than in SS and HC. The numbers of positive cells reported are expressed as medians. *P < 0·05; y‐axis: analysed area expressed as cells per microscopic field, magnification ×400.

Table 1.

Quantitative analysis on the numbers of positive cells for field (400X) in skin lesions of patients with pyoderma gangrenosum, Sweet's syndrome and healthy controls assessed by immunohistochemistry

	PG	SS	HC
IFN‐γ	30 (17·2–48·7)	33 (26–45)	1 (0–2)
IL‐12	19 (13–28)	22 (18–32)	1·5 (0·25–2)
CXCR3	42 (38·2–56·7)	40 (38–66)	4·5 (2–7)
CCR5	27 (21·5–30)	33 (14–36)	1 (0–1·7)
IL‐4	7·5 (5·7–10·2)	8 (4–11)	2 (0·2–3·7)
IL‐5	6 (3–15)	11 (8–42)	1 (1–2)
IL‐13	9 (6·5–14·5)	10 (9–42)	0·5 (0–1)
CCR3	33 (23–37)	54 (43–72)	1 (1–2·7)
CD40	48 (38–57·2)	25 (21–38)	10 (7·5–12)
CD40L	32 (23–48·5)	18 (15–27)	0 (0–0·7)
IL‐15	23 (19–27·5)	2 (1–3)	1·5 (1–2)
CD56	47 (37·5–55)	13 (11·2–15·5)	1 (0–1·7)

Positive‐cell counts are expressed as medians (25th–75th percentile). PG = pyoderma gangrenosum; SS = Sweet's syndrome; HC = healthy controls; CXCR3 = C‐X‐C motif chemokine receptor 3; IFN =interferon; IL = interleukin; CCR =  C‐C motif chemokine receptor; CD40L = CD40 ligand.

Moreover, the immunoreactivity for IL‐12, the leading cytokine associated with Th1 polarization, was similar between PG and SS and stronger compared with HC (P < 0·0001 for both) (Fig. 3, Table 1).

CXCR3 and CCR5 are chemokine receptors expressed sequentially on the Th1 cell surface after IL‐12 stimulation and are involved in the recruitment of macrophages, T cells and monocytes at inflammation sites 21. In both PG and SS, the reactivity for CXCR3 and CCR5 was significantly higher than in HC (P < 0·0001 for both); both molecules showed the same diffuse expression pattern as for IFN‐γ and IL‐12. CCR5⁺ cells were slightly less numerous in PG patients than in SS patients, without significant differences. Conversely, the cells expressing CXCR3 were more numerous in PG samples than in SS samples, although without statistical significance (Fig. 3, Table 1).

Quantification of Th2 markers in PG compared to SS

To analyse the Th2 pathway, we investigated the expression of IL‐4, IL‐5 and IL‐13, which represent the main Th2‐related cytokines, as well as that of CCR3, a chemokine receptor that is expressed predominantly on Th2 cell surface and eosinophils.

Few IL‐4⁺ cells were found in the superficial and medium/deep dermis of PG and SS lesional skin (Figs 2 and 4). The median number of IL‐4⁺ cells was higher in PG and SS than in HC; it was slightly lower in PG than in SS, although without significant differences (Fig. 3, Table 1).

Figure 4.

Immunohistochemical staining for the T helper 2‐associated markers interleukin (IL)‐4, IL‐5, IL‐13 and the C‐X‐C motif chemokine receptor 3 (CCR3) in skin biopsy specimens from patients with pyoderma gangrenosum (PG), Sweet syndrome (SS) and healthy controls (HC). Magnification: ×200.

A moderate immunostaining for IL‐5 and IL‐13 was detected in SS, while only slight reactivity was found in PG. Both cytokines showed the same distribution of IL‐4 (Fig. 4). Both mediators were expressed more in PG and SS than in HC (P < 0·0001 for both). The number of IL‐5⁺ cells was significantly lower in PG skin lesions than in SS (P = 0·0157; Fig. 3). By contrast, although IL‐13 was more expressed in SS than in PG, no statistical significance was reached. Finally, CCR3 expression was evident in the entire dermis of PG and SS skin specimens. The median number of CCR3 expressing cells was higher in both patient groups than in HC (P < 0·0001 for both). However, it was significantly lower in PG than in SS (P < 0·0001; Fig. 3, Table 1).

Co‐stimulatory molecules CD40 and CD40L in the skin of patients with PG compared to SS

In PG and SS patients, the cells expressing the co‐stimulatory molecules CD40 and its ligand CD40L were distributed in the superficial and medium dermis of lesional skin (Figs 2 and 5). Both molecules were over‐expressed significantly in PG and SS compared to HC (P < 0·0001 and P = 0·009, respectively). Interestingly, their numbers were significantly higher in patients with PG than in SS patients (P = 0·0087 for CD40 and P = 0·0073 for CD40L, respectively; Fig. 3, Table 1)

Figure 5.

Immunohistochemical staining for the co‐stimulatory molecules CD40 and CD40 ligand (CD40L) and for interleukin (IL)‐15 and CD56 in skin biopsy specimens from patients with pyoderma gangrenosum (PG), Sweet syndrome (SS) and healthy controls (HC). Magnification: ×200.

IL‐15 and CD56 in the skin of patients with PG compared to SS

IL‐15 is a proinflammatory cytokine which promotes lymphocyte and NK cell proliferation and activity. In our study, IL‐15⁺ cells were located in perivascular areas within the superficial and medium dermis of PG lesional specimens (Fig. 5). Interestingly, their number was found to be significantly higher in PG compared to either SS (P < 0·0001) or HC (P < 0·0001), where a very low number of positive cells was detected (Fig. 3, Table 1).

CD56 is a neural cell adhesion molecule expressed on the surface of NK cells and is considered as a prototypic marker of these cells. In our study, CD56⁺ cells were distributed mainly in the superficial and medium dermis, mainly in perivascular areas, with few elements scattered through the epidermis both in PG and SS (Figs 2 and 5), while in HC skin they were almost absent. In agreement with IL‐15 expression, CD56⁺ cells were significantly more numerous in PG compared to SS (P < 0·0001).

Discussion

Recent studies have demonstrated clearly that neutrophilic dermatoses and, in particular, PG and SS can be included among the group of autoinflammatory disorders that are correlated with hyperactivation of the innate immune system 5. It is of note that even acquired immunity could play a contributory role in the effector phases of both diseases. Accordingly, CD3⁺ T lymphocytes were found in the inflammatory infiltrate of both PG (especially at the wound's edge), and SS 22; moreover, in patients with PG, a T cell clonal expansion was described previously, supporting the possibility of an aberrant T cell response driving its development 23. Nevertheless, the role of acquired immunity in the pathogenesis of these dermatoses still remains poorly investigated. As mentioned above, the proportion of T_reg and Th17 cells has been investigated previously by our group 14; conversely, no data are present in the literature regarding Th1 and Th2 cells and their differential expression between PG and SS. For this reason, the main objective of the present study was to analyse the Th1‐ and Th2‐related cytokines and chemokine receptors in these two neutrophilic disorders.

As expected, Th1 mediators were found hyperexpressed in both PG and SS samples, without significant differences. All these mediators were expressed in superficial and deep dermis, suggesting that Th1 cells may have a synergistic role both in leucocyte recruitment and local tissue damage. Th1 cells are activated mainly by tissue macrophages and other antigen‐presenting cells, such as dendritic cells 24. The role of the latter in neutrophilic dermatoses has yet to be determined, but it could be hypothesized that they are able to process an as‐yet‐identified autoantigen, driving Th1 polarization by the release of IL‐12.

Conversely, in both PG and SS we found reduced expression of IL‐4, IL‐5 and IL‐13 with respect to Th1‐related molecules, suggesting a possible impairment of the Th2‐type response and a consequent predominant Th1‐related inflammation, as happens in several autoinflammatory diseases. In fact, patients with CD, who have a higher risk of developing PG and SS than normal individuals, display an augmented number of Th1 cells and related transcription factors and cytokines both in peripheral blood and mucosa 25. Moreover, in Behçet's disease, several reports have documented the over‐expression of Th1 cytokines 26 and, in particular, skin pathergy reaction, a non‐specific tissue hyper‐reactivity to minor trauma involving epithelial disruption that represents a common feature of Behçet's disease and PG, was related to an exaggerated Th1‐type response 27. Finally, even in psoriasis, where neutrophils play a major role in the development of skin lesions, some studies documented increased levels of IFN‐γ and IL‐12 28, 29, together with down‐regulation of IL‐4, IL‐5 and IL‐10 28.

Interestingly, in our study the number of IL‐5 and CCR‐3 expressing cells was significantly lower in PG compared to SS. Accordingly, Th2 cell depression was also reported in the peripheral blood of PG patients 30.

Th2 cells are involved in wound repair 31; therefore, the marked reduction of Th2‐mediators may justify the tendency of PG lesions to evolve rapidly into an enlarging ulcer, as well as their refractivity in the absence of immunosuppressive treatments.

We have also found a higher expression of the CD40/CD40L axis in PG samples. The latter represents a co‐stimulatory axis of the TNF/TNF receptor superfamily with a central role in Th1 polarization by inducing IL‐12 secretion 32. Moreover, the CD40/CD40L system promotes inflammation via other mechanisms at the bridge between innate and adaptive immunity pathways, such as the up‐regulation of adhesion molecules and the production of various cytokines and chemokines such as IL‐1, TNF‐α, IL‐8 and regulated on activation, normal T cell expressed and secreted (RANTES), and enzymes such as matrix metalloproteinase 33. Our findings provide other evidence for the more aggressive behaviour of the inflammatory background in PG, and suggest a stronger impairment between the Th1 and Th2 activity compared to SS. The data from the present study confirmed those of a previous work of our group, where similar results were found by using protein microarray technique 34.

The other aim of our study was to evaluate the expression of IL‐15 in neutrophilic dermatoses. IL‐15 plays a wide range of functions: it is able to protect neutrophils from apoptosis, modulate phagocytosis and stimulate the secretion of IL‐8, TNF‐α, IL‐6 and IL‐1 35; in addition, as described previously, IL‐15 plays a central role in NK cell biology.

IL‐15 has long been investigated in inflammatory bowel diseases and coeliac disease 8, 18. In particular, patients with Crohn's disease or ulcerative colitis showed an increased expression of IL‐15 in mucosal tissues 17, while decreasing levels of this cytokine has been documented in patients with Crohn's disease responding to infliximab therapy 36.

Interestingly, in our study, IL‐15⁺ cells were significantly more numerous in PG skin lesions than in SS and HC, where only few IL‐15 expressing cells could be detected. Its expression was particularly concentrated near perivascular spaces, suggesting that it could be involved in leucocyte recruitment. Moreover, none of our PG patients was affected by inflammatory bowel disorders, suggesting that IL‐15 may have an independent role in the pathogenesis of PG, even in its sporadic variant. As for IL‐15, we detected over‐expression of the NK cell marker CD56 in the inflammatory infiltrate of PG with respect to SS. This result might suggest the involvement of IL‐15 in the NK cell recruitment and activation in PG.

The limitation of our study concerns the small study group; however, it is in line with other experimental studies on PG and SS and mirrors the low incidence of these diseases among the general population.

In conclusion, our results confirmed that Th1 mediators are hyperexpressed in the lesional skin of PG and SS. The over‐expression of IL‐12 suggests a possible role of antigen‐presenting cells in the activation of Th1 cells in an antigen‐dependent manner.

Compared to SS, PG shows a significant depression of the Th2 markers together with an increased expression of the CD40/CD40L system. These data are in agreement with the more aggressive clinical course of the disease.

Finally, for the first time to our knowledge, we demonstrated a high expression of IL‐15 and the NK cell marker CD56 in PG. Despite the small number of patients, our results might suggest a potential and even unexplored role of NK cells in the pathogenesis of PG. Accordingly, although further studies are needed to confirm our preliminary results, IL‐15 could represent a potential therapeutic target to affect NK cell recruitment in PG lesions.

Disclosure

The authors declare that they have no conflicts of interest.