- Email: [email protected]
Eosinophil extracellular DNA traps in skin diseases
Eosinophil extracellular DNA traps in skin diseases Dagmar Simon, MD,a Susanne Hoesli, MD,b Nina Roth,a Simon Staedler,a Shida Yousefi, PhD,b and Hans-Uwe Simon, MD, PhDb Bern, Switzerland Background: In the skin, eosinophils are found in a broad spectrum of diseases, including infectious diseases. Objective: In this study, we investigated whether eosinophil extracellular traps, structures containing DNA in association with eosinophil granule proteins able to bind and kill bacteria, are present in the skin under various pathologic conditions. Methods: Immunofluorescence staining was performed on sections of paraformaldehyde-fixed and paraffin-embedded skin biopsy tissues of 25 different eosinophilic skin diseases by using propidium iodide and an antibody to eosinophil cationic protein. Slides were evaluated by laser scanning microscopy. Results: Eosinophils releasing DNA together with eosinophil cationic protein were detected in infectious skin diseases such as ectoparasitosis and larva migrans. Further, we observed the extracellular DNA structures in allergic/reactive diseases (Wells syndrome, hypereosinophilic syndrome, positive reaction of atopy patch test, allergic contact dermatitis, drug hypersensitivity) and in autoimmune diseases (bullous pemphigoid, pemphigus foliaceus, dermatitis herpetiformis). The average number of eosinophils releasing DNA in the skin was usually below 10%, although in Wells syndrome the proportion was up to 30%. In areas with clusters of eosinophils, up to 50% of the eosinophils were seen to generate eosinophil extracellular traps. Conclusion: Eosinophil extracellular traps are seen in both infectious and noninfectious inflammatory skin diseases and are particularly common in Wells syndrome. (J Allergy Clin Immunol 2011;127:194-9.) Key words: ECP, eosinophils, eosinophil extracellular traps, DNA, skin disease
Eosinophils are found among infiltrating cells in a broad spectrum of skin diseases.1,2 The primary function of eosinophils has been related to the protection against helminth parasites.3 On the other hand, eosinophils are accused to cause tissue damage in helminth infections as well as in bronchial asthma by releasing granule cationic proteins.3,4 For instance, eosinophil cationic protein (ECP) and major basic protein damage target cell membranes and are thus highly cytotoxic.5-7 ECP and eosinophil-
From athe Department of Dermatology, Inselspital, Bern University Hospital, and bthe Institute of Pharmacology, University of Bern. Supported by grants from the Stanley Thomas Johnson Foundation (D.S.) and the Swiss National Foundation (S.Y., H.-U.S.) Disclosure of potential conflict of interest: The authors have declared that they have no conflict of interest. Received for publication August 19, 2010; revised October 31, 2010; accepted for publication November 4, 2010. Reprint requests: Dagmar Simon, MD, Department of Dermatology, Inselspital, CH-3010 Bern, Switzerland. E-mail: [email protected]. 0091-6749/$36.00 Ó 2010 American Academy of Allergy, Asthma & Immunology doi:10.1016/j.jaci.2010.11.002
194
Abbreviations used ACD: Allergic contact dermatitis AD: Atopic dermatitis APT: Atopy patch test BP: Bullous pemphigoid DH: Dermatitis herpetiformis DHS: Drug hypersensitivity ECP: Eosinophil cationic protein EET: Eosinophil extracellular trap HES: Hypereosinophilic syndromes PF: Pemphigus foliaceus PI: Propidium iodide WS: Wells syndrome
derived neurotoxin have also been identified as ribonucleases8 and possess antiviral activity.9 Recently, it has been suggested that eosinophils contribute to antibacterial defense mechanisms by releasing mitochondrial DNA in association with granule proteins.10 Similar extracellular structures can be formed by activated neutrophils and are called neutrophil extracellular traps.11,12 DNA-releasing eosinophils have been demonstrated in inflammatory diseases of the intestine only.10 Here we searched for these structures in different eosinophilic skin diseases. Because the extracellular DNA-containing structures generated by eosinophils are, like neutrophil extracellular traps, able to bind and kill bacteria,10,11 we used the term eosinophil extracellular traps (EETs) to describe this phenomenon in this study. In addition, we investigated the expression of IL-5, IFN-g, and eotaxins as potential triggers of EETs in EET-containing skin tissues as well as the location and type of cells undergoing apoptosis under these conditions.
METHODS Skin biopsies Skin specimens from biopsies taken for routine diagnostics were obtained from the archives of the histology laboratory of the Department of Dermatology, University of Bern. The study was approved by a hospital institutional board and the ethics committee of the Canton Bern. Representative specimens (numbers >1 are given) revealing tissue eosinophilia were selected from parasitic infestation (larva migrans, ectoparasitosis, n 5 3), atopic dermatitis (AD, n 5 5) including positive atopy patch test reaction (APT), allergic contact dermatitis (ACD), drug hypersensitivity (DHS, n53), urticaria (n 5 3), prurigo nodularis, eosinophilic pustular folliculitis, hypereosinophilic syndromes (HES, n 5 2), eosinophilic cellulitis/dermatitis (Wells syndrome, [WS], n 5 3), bullous pemphigoid (BP, n 5 5), pemphigus foliaceus (PF), dermatitis herpetiformis (DH, n 5 2), dermatomyositis, polyarteritis nodosa, eosinophilic fasciitis, Wegener granulomatosis (n 5 2), morphea (n 5 2), cutaneous T-cell lymphoma (n 5 4), B-cell lymphoma, pseudolymphoma (n 5 2), Langerhans cell histiocytosis (n 5 2), cutaneous mastocytosis, angiolymphoid hyperplasia with eosinophilia, and solid tumors (keratoacanthoma [n 5 2], sweat gland carcinoma, benign fibrous histiocytoma, melanoma metastasis).
SIMON ET AL 195
J ALLERGY CLIN IMMUNOL VOLUME 127, NUMBER 1
FIG 1. Eosinophil extracellular traps (arrows). Accumulation of eosinophils releasing DNA (red) and ECP (green) in BP. DNA plus ECP, overlay (A), DNA (B), ECP (C).
Confocal laser scanning microscopy To detect EETs, immunofluorescence staining was performed on 6-mm paraformaldehyde-fixed and paraffin-embedded tissue sections by using propidium iodide (PI; Sigma, Buchs, Switzerland) to stain DNA together with a monoclonal mouse antihuman ECP antibody (Pharmacia Diagnostics AB, Uppsala, Sweden). The slides were analyzed by confocal laser-scanning microscopy (LSM510; Carl Zeiss MicroImaging, Jena, Germany). With an interval of 0.1 mm, slices were taken throughout the z-axis (z stacks) with 30 to 60 slices per stack. Image analysis was performed by using Imaris software (Bitplane, Zurich, Switzerland) to visualize the shapes and sizes of the DNA structures.10,12 Because EETs are difficult to detect in fixed tissues, stainings were performed in triplicate, and EETs were considered present if at least 2 of 3 sections were positively assessed. Moreover, immunofluorescence staining using antibodies directed to IL-5 (Santa Cruz Biotechnology, Santa Cruz, Calif) as well as IFN-g and eotaxin-1, eotaxin-2, and eotaxin-3 (all R&D Systems, Abington, United Kingdom) and appropriate Alexa Fluor 488–labelled secondary antibodies (Invitrogen Molecular Probes, Paisley, United Kingdom) was performed.13,14 To identify apoptotic cells, an antibody to cleaved caspase-3 (Cell Signaling Technology, Danvers, Mass) was used.15 Mouse monoclonal IgG1 and rabbit polyclonal IgG control antibodies (both from Dako Cytomation, Glostrup, Denmark) and normal goat IgG (R&D Systems) served as negative controls. Immunofluorescence staining was evaluated by at least 2 independent investigators (D.S., S.H., N.R., and S.S.). Using confocal microscopy, positive cells were counted in 10 consecutive fields (each 0.01 mm2; magnification 31000) of each specimen. Cell numbers are given as total or mean values (6SD in text or ranges in tables) if n >1.
RESULTS EETs are present in inflammatory skin derived from patients with different underlying diseases We first investigated whether EETs can be detected in skin tissues as a possible consequence of infectious and/or other triggers. Indeed, extracellular DNA in association with ECP was
detected in specimens of infectious skin diseases such as ectoparasitosis and larva migrans (Fig 1). Interestingly, we also found EETs in skin biopsies of WS and HES, as well as in the autoimmune bullous diseases BP, PF, and DH, and in allergic/reactive responses such as positive APT reaction, ACD, and DHS. In the remaining skin diseases investigated (AD, urticaria, prurigo nodularis, eosinophilic pustular folliculitis, dermatomyositis, polyarteritis nodosa, eosinophilic fasciitis, Wegner granulomatosis, morphea, and eosinophilic responses in association with different skin neoplasms), we identified no EETs. In all diseases except WS in which we identified EETs, the number of DNA plus ECP releasing eosinophils was below 10% of infiltrating eosinophils (Table I). In WS, the proportion was higher and ranged between 10% and 30%. In ectoparasitosis and APT, only single DNA-releasing cells among infiltrating eosinophils were found. Moreover, in most of the specimens, we observed eosinophil clusters with peak counts of 6 to 16 eosinophils per 0.01 mm2 field. In these areas, aggregates of EETs were detected with a higher proportion of eosinophils releasing DNA that could reach up to 50% in ACD and WS (Table I). Interestingly, the long and thin DNA structures often seemed to connect the releasing eosinophils with other cells, including other eosinophils (Fig 1). EETs could easily be distinguished from flame figures that are formed by broad collagen fibers covered with eosinophil granule proteins—for example, ECP (Fig 2, A and B). Flame figures were present in WS, HES, and the autoimmune-bullous diseases BP and PF. By immunofluorescence staining using an anti-ECP antibody, we detected extracellular eosinophil granule deposition in all specimens, which was usually not located adjacent to DNA-releasing eosinophils, except in WS. In WS, extracellular DNA structures and eosinophil granule deposits were observed in areas of eosinophil clusters (Fig 2, C).
196 SIMON ET AL
J ALLERGY CLIN IMMUNOL JANUARY 2011
TABLE I. Eosinophils generating EETs
Eosinophils, total or mean (range) cell numbers in 10 fields
Eosinophils releasing DNA with ECP 1, <10% 11, 10% to 30%
Proportion of DNAreleasing eosinophils in clusters 1, <10% 11, 10% to 30% 111, 31% to 50%
9 (3-14) 70 22 14 21 (7-29) 40 (24-53) 27 (25-28) 31 (19-53)
1 1 1 1 1 11 1 1
S 11 S 111 11 111 11 11
PF
30
1
11
DH
45 (8-16)
1
11
Disease
Ectoparasitosis Larva migrans APT ACD DHS WS HES BP
Localization of eosinophils releasing DNA
Dermis Deep dermis Upper dermis Upper dermis Dermis Dermis Dermis Upper dermis, lining at dermal-epidermal junction in prebullous stadium Upper dermis, intraepidermal Papillary dermis
S, Single eosinophils. In the indicated skin diseases, the absolute numbers of eosinophils, the proportion of eosinophils releasing EETs in the whole specimen and in eosinophil clusters, and localization of EET-releasing eosinophils are provided.
FIG 2. Release of ECP by eosinophils. A, A single eosinophil generating an EET consisting of DNA (red) and ECP (green) in larva migrans. B, Broad fibrous structures covered with ECP (green) forming a flame figure in WS. C, EETs (arrows) and eosinophil granule depositions with ECP (green; arrowheads) in WS. DNA was stained with PI (red). Magnification 31000.
IL-5 and eotaxins are expressed under conditions of EET formation In the EET-positive skin tissues, we investigated the numbers of cells expressing IL-5, IFN-g, and eotaxins 1 to 3 among the dermal infiltrating cells (Table II; Fig 3). The highest numbers of IL-5–expressing cells were observed in larva migrans and DHS. Among lymphocytes expressing IL-5, we usually observed a predominance of CD4-positive cells compared with CD8-positive cells. Eosinophils expressing IL-5 were found in APT, ACD, WS, HES, infectious diseases, and BP. In addition to IL-5, considerable expression of IFN-g was seen in both CD4-positive and CD8-positive cells in the same tissue specimen (Table II). Compared with IL-5 and IFN-g, the numbers of eotaxin-positive cells were less: the mean number was 17 6 16 per 10 fields. The highest numbers of eotaxin-expressing cells were seen in WS and HES. Eotaxins 2 and 3 were expressed in all tissues at mean levels of 28 6 22 and 16 6 12 cells, respectively. The mean cell count of eotaxin-1–positive cells was highest in WS, HES, and
infectious diseases. The numbers of eotaxin-3–expressing cells were highest in specimens in which eosinophil clusters with more than 30% of EET-releasing eosinophils occurred (P 5.046). No correlation was found between clustered EET-releasing eosinophils and numbers of IL-5, IFN-g, eotaxin-1, and eotaxin2–expressing cells.
Concurrent presence of EETs and apoptotic keratinocytes The pathogenic role of eosinophils has been associated with host defense, but also with tissue damage. To detect apoptotic cells in the vicinity of eosinophils, we stained specimen of skin diseases, in which we had detected EETs, with an antibody to cleaved caspase-3 and PI. Apoptotic cells in the dermis were seen in 14 of 23 specimens. In specimens of BP, DH, and DHS, we detected between 9 and 16 cleaved caspase-3–positive cells in the dermis, in the remaining diseases just 1 to 4 positive cells.
SIMON ET AL 197
J ALLERGY CLIN IMMUNOL VOLUME 127, NUMBER 1
TABLE II. Quantitative analysis of IL-5, IFN-g, and eotaxin expression in different skin diseases characterized by the presence of EETreleasing eosinophils* Disease
Ectoparasitosis Larva migrans APT ACD DHS WS HES BP PF DH
IL-5 (all IL-5 (CD4IL-5 (CD8IFN-g (all IFN-g (CD4IFN-g (CD8- Eotaxin-1 (all Eotaxin-2 (all Eotaxin-3 (all positive cells) positive cells) positive cells) positive cells) positive cells) positive cells) positive cells) positive cells) positive cells)
41 (27-55) 70 27 34 53 (32-75) 28 (17-43) 38 (34-41) 35 (19-49) 30 46 (45-46)
29 (7-12) 55 18 10 19 (7-27) 4 (3-6) 11 (11-11) 14 (9-21) 12 14 (5-16)
2 (1-3) 5 0 2 11 (1-26) 4 (0-11) 1 (2-5) 4 (0-11) 2 3 (1-5)
23 (19-30) 58 38 59 55 (33-75) 35 (31-40) 42 (37-47) 32 (26-44) 52 45 (44-46)
9 (8-11) 26 29 39 24 (14-37) 12 (5-24) 16 (11-20) 14 (7-19) 28 22 (16-28)
10 (9-10) 7 9 20 31 (19-38) 13 (3-25) 20 (14-26) 12 (1-17) 24 19 (15-22)
6 (2-12) 30 2 9 3 (0-9) 20 (12-26) 12 (10-13) 7 (0-16) 1 15 (7-22)
29 (21-36) 48 12 7 8 (3-17) 39 (25-57) 70 (45-94) 29 (13-60) 12 18 (15-10)
6 (4-10) 34 19 21 9 (1-16) 26 (18-38) 10 (5-14) 11 (7-19) 13 34 (13-55)
*Data are total or mean (ranges) cell numbers in 10 fields.
Apoptotic eosinophils were either not or very rarely detected. On the other hand, in the epidermis, which has just single positive cells in the basal layer under normal conditions,15 we observed cleaved caspase-3–positive keratinocytes scattered throughout the epidermis (DHS) or in groups (APT and ACD). In WS, HES, BP, PF, and DH, we found positive cells along the basal and suprabasal layers. In BP, eosinophils lined the dermalepidermal junction in prebullous stadiums (Fig 4, A). In these areas, eosinophils generating EETs have been found (Fig 4, B-D). Those parts of the epidermis covering blisters in BP and DH were intensively stained with the antibody to cleaved caspase-3.
DISCUSSION In this study, we demonstrate for the first time that EETs are present in several eosinophilic skin diseases. We observed these extracellular structures in infectious skin diseases such as larva migrans and ectoparasitoses. Interestingly, EETs were also detectable in other inflammatory skin diseases, such as allergic/ reactive and autoimmune diseases associated with eosinophilia. From a mechanistic point of view, EETs might be generated by IL-5/IFN-g and eotaxins, which were clearly expressed in all these diseases and sufficient to trigger the formation of these extracellular structures in vitro.10 In contrast, in other skin diseases, such as skin neoplasms with tissue-infiltrating eosinophils, EETs were not found. However, nothing is known about the survival and enzymatic degradation of EETs in tissues so far. It is possible that the time from taking the biopsy until fixation is crucial because of the presence of tissue nucleases. Therefore, the interpretation of our negative findings as well as the sometimes low numbers of EET-releasing eosinophils should be performed with caution. Clearly, additional research is required with respect to both numbers of investigated tissues and comparison of different methods of tissue preservation. In spite of these limitations, some conclusions might be possible, at least from those data in which we observed EETs. First, the diseases in which we found extracellular DNA structures were acute dermatoses, such as infectious diseases, ACD, DHS, or acute exacerbations of WS and HES. Interestingly, in a positive reaction of APT, which resembles an acute, 2-day-old exacerbation of AD induced by an allergen, we also detected EETs, whereas none were seen in 5 specimens of lesional AD skin.
FIG 3. IL-5 and eotaxin–expressing cells in skin tissues containing EETs. A, IL-5–expressing (green) and CD4-positive (red) cells in APT. Double-positive cells (yellow) are indicated by arrows, single IL-5–expressing cells (green) by arrowheads. Single CD4-positive cells (red) are also visible. B, Eotaxin2–positive cells in BP. Eosinophils (arrows) and noneosinophilic cells (arrowhead) express eotaxin-2.
Second, most of the diseases with EETs, except WS and HES, were related to an exogenous (parasite, allergen) or endogenous (autoantigen, drug) pathogenic trigger. Taken together, these data point to the possibility that EETs are present in acute phases or exacerbations of eosinophilic skin diseases. Therefore, EETs might be generated by eosinophils that are newly recruited to the skin. The eosinophils generating EETs revealed the same size and shape compared with other eosinophils in the tissues and looked morphologically intact. They were most likely nonapoptotic because they were caspase-3–negative and had normal nuclear morphology. It should be noted that extracellular depositions was also found in the absence of DNA release. For instance, we detected ECP-positive extracellular granules in all eosinophilic skin biopsies. Moreover, we observed ECP in association with flame figures in WS and HES as well as BP and PF. This suggests that ECP release is not dependent on and can occur independently of DNA release. Alternatively, it is possible that ECP is less subject to degradation compared with DNA and therefore can be accumulated more easily. Why eosinophils release and form EETs is not clear. It can be hypothesized that these DNA structures guarantee that toxic granule proteins are effectively brought to a pathogen/antigen, which might be trapped by the DNA nets. On the other hand, it can be speculated that with the association of toxic granule proteins to
198 SIMON ET AL
J ALLERGY CLIN IMMUNOL JANUARY 2011
FIG 4. Cleaved caspase-3 expression and EETs in BP. A, Caspase-3–positive cells throughout the epidermis, beneath eosinophils (arrows) in the dermis. Nuclei were counterstained with PI. B-D, Eosinophil at the dermal-epidermal junction (dashed line) generating an EET (arrow). DNA plus ECP, overlay (B), DNA (C), ECP (D). Magnification 3400 (A), 31000 (B-D).
the DNA structures, collateral tissue damage might be limited. Indeed, the relation of eosinophil peak infiltration including the formation of EETs to apoptotic cells in the dermis was not obvious, supporting this hypothesis. On the other hand, we observed large numbers of eosinophils and EETs in proximity to apoptotic keratinocytes of patients with BP, implying that eosinophils with or without the help of EETs are involved in keratinocyte damage including blister formation. Future studies including clinical trials using specific antieosinophil therapies may bring more clarity to the pathogenic role of eosinophils in this process. We thank E. Kozlowski for technical assistance.
Key messages d
Eosinophils generating EETs are present in inflammatory skin diseases.
d
Eosinophil extracellular traps containing DNA and eosinophil granule proteins are distinct from flame figures and eosinophil granule depositions.
REFERENCES 1. Simon D, Simon HU. Eosinophilic disorders. J Allergy Clin Immunol 2007;119: 1291-300.
J ALLERGY CLIN IMMUNOL VOLUME 127, NUMBER 1
2. Simon D, Wardlaw A, Rothenberg ME. Organ-specific eosinophilic disorders of the skin, lung, and gastrointestinal tract. J Allergy Clin Immunol 2010;126: 3-13. 3. Klion AD, Nutman TB. The role of eosinophils in host defense against helminth parasites. J Allergy Clin Immunol 2004;113:30-7. 4. Frigas E, Motojima S, Gleich GJ. The eosinophilic injury to the mucosa of the airways in the pathogenesis of bronchial asthma. Eur Respir J 1991;13: S123-35. 5. Gleich GJ, Frigas E, Loegering DA, Wassom DL, Steinmuller D. Cytotoxic properties of the eosinophil major basic protein. J Immunol 1979;123:2925-7. 6. Kroegel C, Costabel U, Matthys H. Mechanism of membrane damage mediated by eosinophil major basic protein. Lancet 1987;1:1380-1. 7. Young JD, Peterson CG, Venge P, Cohn ZA. Mechanism of membrane damage mediated by human eosinophil cationic protein. Nature 1986;321:613-6. 8. Gleich GJ, Loegering DA, Bell MP, Checkel JL, Ackerman SJ, McKean DJ. Biochemical and functional similarities between human eosinophil-derived neurotoxin and eosinophil cationic protein: homology with ribonucleases. Proc Natl Acad Sci U S A 1986;83:3146-50.
SIMON ET AL 199
9. Rosenberg HF, Domachowske JB. Eosinophils, eosinophil ribonucleases, and their role in host defence against respiratory virus pathogens. J Leukoc Biol 2001;70:691-8. 10. Yousefi S, Gold JA, Andina N, Lee JJ, Kelly AM, Kozlowski E, et al. Catapult-like release of mitochondrial DNA by eosinophils contributes to antibacterial defense. Nat Med 2008;14:949-53. 11. Brinkmann V, Reichard U, Goosmann C, Fauler B, Uhlemann Y, Weiss DS, et al. Neutrophil extracellular traps kill bacteria. Science 2004;303:1532-5. 12. Yousefi S, Mihalache C, Kozlowski E, Schmid I, Simon HU. Viable neutrophils release mitochondrial DNA to form neutrophil extracellular traps. Cell Death Differ 2009;16:1438-44. 13. Simon D, H€osli S, Kostylina G, Yawalkar N, Simon HU. Anti-CD20 (rituximab) treatment improves atopic eczema. J Allergy Clin Immunol 2008;121:122-8. 14. Straumann A, Conus S, Grzonka P, Kita H, Kephart G, Bussmann C, et al. Anti-interleukin-5 antibody treatment (mepolizumab) in active eosinophilic oesophagitis: a randomised, placebo-controlled, double-blind trial. Gut 2010;59:21-30. 15. Simon D, Lindberg RLP, Kozlowski E, Braathen LR, Simon HU. Epidermal caspase-3 cleavage associated with interferon-gamma expressing lymphocytes in acute atopic dermatitis lesions. Exp Dermatol 2006;15:441-6.
Eosinophils are found among infiltrating cells in a broad spectrum of skin diseases.1,2 The primary function of eosinophils has been related to the protection against helminth parasites.3 On the other hand, eosinophils are accused to cause tissue damage in helminth infections as well as in bronchial asthma by releasing granule cationic proteins.3,4 For instance, eosinophil cationic protein (ECP) and major basic protein damage target cell membranes and are thus highly cytotoxic.5-7 ECP and eosinophil-
From athe Department of Dermatology, Inselspital, Bern University Hospital, and bthe Institute of Pharmacology, University of Bern. Supported by grants from the Stanley Thomas Johnson Foundation (D.S.) and the Swiss National Foundation (S.Y., H.-U.S.) Disclosure of potential conflict of interest: The authors have declared that they have no conflict of interest. Received for publication August 19, 2010; revised October 31, 2010; accepted for publication November 4, 2010. Reprint requests: Dagmar Simon, MD, Department of Dermatology, Inselspital, CH-3010 Bern, Switzerland. E-mail: [email protected]. 0091-6749/$36.00 Ó 2010 American Academy of Allergy, Asthma & Immunology doi:10.1016/j.jaci.2010.11.002
194
Abbreviations used ACD: Allergic contact dermatitis AD: Atopic dermatitis APT: Atopy patch test BP: Bullous pemphigoid DH: Dermatitis herpetiformis DHS: Drug hypersensitivity ECP: Eosinophil cationic protein EET: Eosinophil extracellular trap HES: Hypereosinophilic syndromes PF: Pemphigus foliaceus PI: Propidium iodide WS: Wells syndrome
derived neurotoxin have also been identified as ribonucleases8 and possess antiviral activity.9 Recently, it has been suggested that eosinophils contribute to antibacterial defense mechanisms by releasing mitochondrial DNA in association with granule proteins.10 Similar extracellular structures can be formed by activated neutrophils and are called neutrophil extracellular traps.11,12 DNA-releasing eosinophils have been demonstrated in inflammatory diseases of the intestine only.10 Here we searched for these structures in different eosinophilic skin diseases. Because the extracellular DNA-containing structures generated by eosinophils are, like neutrophil extracellular traps, able to bind and kill bacteria,10,11 we used the term eosinophil extracellular traps (EETs) to describe this phenomenon in this study. In addition, we investigated the expression of IL-5, IFN-g, and eotaxins as potential triggers of EETs in EET-containing skin tissues as well as the location and type of cells undergoing apoptosis under these conditions.
METHODS Skin biopsies Skin specimens from biopsies taken for routine diagnostics were obtained from the archives of the histology laboratory of the Department of Dermatology, University of Bern. The study was approved by a hospital institutional board and the ethics committee of the Canton Bern. Representative specimens (numbers >1 are given) revealing tissue eosinophilia were selected from parasitic infestation (larva migrans, ectoparasitosis, n 5 3), atopic dermatitis (AD, n 5 5) including positive atopy patch test reaction (APT), allergic contact dermatitis (ACD), drug hypersensitivity (DHS, n53), urticaria (n 5 3), prurigo nodularis, eosinophilic pustular folliculitis, hypereosinophilic syndromes (HES, n 5 2), eosinophilic cellulitis/dermatitis (Wells syndrome, [WS], n 5 3), bullous pemphigoid (BP, n 5 5), pemphigus foliaceus (PF), dermatitis herpetiformis (DH, n 5 2), dermatomyositis, polyarteritis nodosa, eosinophilic fasciitis, Wegener granulomatosis (n 5 2), morphea (n 5 2), cutaneous T-cell lymphoma (n 5 4), B-cell lymphoma, pseudolymphoma (n 5 2), Langerhans cell histiocytosis (n 5 2), cutaneous mastocytosis, angiolymphoid hyperplasia with eosinophilia, and solid tumors (keratoacanthoma [n 5 2], sweat gland carcinoma, benign fibrous histiocytoma, melanoma metastasis).
SIMON ET AL 195
J ALLERGY CLIN IMMUNOL VOLUME 127, NUMBER 1
FIG 1. Eosinophil extracellular traps (arrows). Accumulation of eosinophils releasing DNA (red) and ECP (green) in BP. DNA plus ECP, overlay (A), DNA (B), ECP (C).
Confocal laser scanning microscopy To detect EETs, immunofluorescence staining was performed on 6-mm paraformaldehyde-fixed and paraffin-embedded tissue sections by using propidium iodide (PI; Sigma, Buchs, Switzerland) to stain DNA together with a monoclonal mouse antihuman ECP antibody (Pharmacia Diagnostics AB, Uppsala, Sweden). The slides were analyzed by confocal laser-scanning microscopy (LSM510; Carl Zeiss MicroImaging, Jena, Germany). With an interval of 0.1 mm, slices were taken throughout the z-axis (z stacks) with 30 to 60 slices per stack. Image analysis was performed by using Imaris software (Bitplane, Zurich, Switzerland) to visualize the shapes and sizes of the DNA structures.10,12 Because EETs are difficult to detect in fixed tissues, stainings were performed in triplicate, and EETs were considered present if at least 2 of 3 sections were positively assessed. Moreover, immunofluorescence staining using antibodies directed to IL-5 (Santa Cruz Biotechnology, Santa Cruz, Calif) as well as IFN-g and eotaxin-1, eotaxin-2, and eotaxin-3 (all R&D Systems, Abington, United Kingdom) and appropriate Alexa Fluor 488–labelled secondary antibodies (Invitrogen Molecular Probes, Paisley, United Kingdom) was performed.13,14 To identify apoptotic cells, an antibody to cleaved caspase-3 (Cell Signaling Technology, Danvers, Mass) was used.15 Mouse monoclonal IgG1 and rabbit polyclonal IgG control antibodies (both from Dako Cytomation, Glostrup, Denmark) and normal goat IgG (R&D Systems) served as negative controls. Immunofluorescence staining was evaluated by at least 2 independent investigators (D.S., S.H., N.R., and S.S.). Using confocal microscopy, positive cells were counted in 10 consecutive fields (each 0.01 mm2; magnification 31000) of each specimen. Cell numbers are given as total or mean values (6SD in text or ranges in tables) if n >1.
RESULTS EETs are present in inflammatory skin derived from patients with different underlying diseases We first investigated whether EETs can be detected in skin tissues as a possible consequence of infectious and/or other triggers. Indeed, extracellular DNA in association with ECP was
detected in specimens of infectious skin diseases such as ectoparasitosis and larva migrans (Fig 1). Interestingly, we also found EETs in skin biopsies of WS and HES, as well as in the autoimmune bullous diseases BP, PF, and DH, and in allergic/reactive responses such as positive APT reaction, ACD, and DHS. In the remaining skin diseases investigated (AD, urticaria, prurigo nodularis, eosinophilic pustular folliculitis, dermatomyositis, polyarteritis nodosa, eosinophilic fasciitis, Wegner granulomatosis, morphea, and eosinophilic responses in association with different skin neoplasms), we identified no EETs. In all diseases except WS in which we identified EETs, the number of DNA plus ECP releasing eosinophils was below 10% of infiltrating eosinophils (Table I). In WS, the proportion was higher and ranged between 10% and 30%. In ectoparasitosis and APT, only single DNA-releasing cells among infiltrating eosinophils were found. Moreover, in most of the specimens, we observed eosinophil clusters with peak counts of 6 to 16 eosinophils per 0.01 mm2 field. In these areas, aggregates of EETs were detected with a higher proportion of eosinophils releasing DNA that could reach up to 50% in ACD and WS (Table I). Interestingly, the long and thin DNA structures often seemed to connect the releasing eosinophils with other cells, including other eosinophils (Fig 1). EETs could easily be distinguished from flame figures that are formed by broad collagen fibers covered with eosinophil granule proteins—for example, ECP (Fig 2, A and B). Flame figures were present in WS, HES, and the autoimmune-bullous diseases BP and PF. By immunofluorescence staining using an anti-ECP antibody, we detected extracellular eosinophil granule deposition in all specimens, which was usually not located adjacent to DNA-releasing eosinophils, except in WS. In WS, extracellular DNA structures and eosinophil granule deposits were observed in areas of eosinophil clusters (Fig 2, C).
196 SIMON ET AL
J ALLERGY CLIN IMMUNOL JANUARY 2011
TABLE I. Eosinophils generating EETs
Eosinophils, total or mean (range) cell numbers in 10 fields
Eosinophils releasing DNA with ECP 1, <10% 11, 10% to 30%
Proportion of DNAreleasing eosinophils in clusters 1, <10% 11, 10% to 30% 111, 31% to 50%
9 (3-14) 70 22 14 21 (7-29) 40 (24-53) 27 (25-28) 31 (19-53)
1 1 1 1 1 11 1 1
S 11 S 111 11 111 11 11
PF
30
1
11
DH
45 (8-16)
1
11
Disease
Ectoparasitosis Larva migrans APT ACD DHS WS HES BP
Localization of eosinophils releasing DNA
Dermis Deep dermis Upper dermis Upper dermis Dermis Dermis Dermis Upper dermis, lining at dermal-epidermal junction in prebullous stadium Upper dermis, intraepidermal Papillary dermis
S, Single eosinophils. In the indicated skin diseases, the absolute numbers of eosinophils, the proportion of eosinophils releasing EETs in the whole specimen and in eosinophil clusters, and localization of EET-releasing eosinophils are provided.
FIG 2. Release of ECP by eosinophils. A, A single eosinophil generating an EET consisting of DNA (red) and ECP (green) in larva migrans. B, Broad fibrous structures covered with ECP (green) forming a flame figure in WS. C, EETs (arrows) and eosinophil granule depositions with ECP (green; arrowheads) in WS. DNA was stained with PI (red). Magnification 31000.
IL-5 and eotaxins are expressed under conditions of EET formation In the EET-positive skin tissues, we investigated the numbers of cells expressing IL-5, IFN-g, and eotaxins 1 to 3 among the dermal infiltrating cells (Table II; Fig 3). The highest numbers of IL-5–expressing cells were observed in larva migrans and DHS. Among lymphocytes expressing IL-5, we usually observed a predominance of CD4-positive cells compared with CD8-positive cells. Eosinophils expressing IL-5 were found in APT, ACD, WS, HES, infectious diseases, and BP. In addition to IL-5, considerable expression of IFN-g was seen in both CD4-positive and CD8-positive cells in the same tissue specimen (Table II). Compared with IL-5 and IFN-g, the numbers of eotaxin-positive cells were less: the mean number was 17 6 16 per 10 fields. The highest numbers of eotaxin-expressing cells were seen in WS and HES. Eotaxins 2 and 3 were expressed in all tissues at mean levels of 28 6 22 and 16 6 12 cells, respectively. The mean cell count of eotaxin-1–positive cells was highest in WS, HES, and
infectious diseases. The numbers of eotaxin-3–expressing cells were highest in specimens in which eosinophil clusters with more than 30% of EET-releasing eosinophils occurred (P 5.046). No correlation was found between clustered EET-releasing eosinophils and numbers of IL-5, IFN-g, eotaxin-1, and eotaxin2–expressing cells.
Concurrent presence of EETs and apoptotic keratinocytes The pathogenic role of eosinophils has been associated with host defense, but also with tissue damage. To detect apoptotic cells in the vicinity of eosinophils, we stained specimen of skin diseases, in which we had detected EETs, with an antibody to cleaved caspase-3 and PI. Apoptotic cells in the dermis were seen in 14 of 23 specimens. In specimens of BP, DH, and DHS, we detected between 9 and 16 cleaved caspase-3–positive cells in the dermis, in the remaining diseases just 1 to 4 positive cells.
SIMON ET AL 197
J ALLERGY CLIN IMMUNOL VOLUME 127, NUMBER 1
TABLE II. Quantitative analysis of IL-5, IFN-g, and eotaxin expression in different skin diseases characterized by the presence of EETreleasing eosinophils* Disease
Ectoparasitosis Larva migrans APT ACD DHS WS HES BP PF DH
IL-5 (all IL-5 (CD4IL-5 (CD8IFN-g (all IFN-g (CD4IFN-g (CD8- Eotaxin-1 (all Eotaxin-2 (all Eotaxin-3 (all positive cells) positive cells) positive cells) positive cells) positive cells) positive cells) positive cells) positive cells) positive cells)
41 (27-55) 70 27 34 53 (32-75) 28 (17-43) 38 (34-41) 35 (19-49) 30 46 (45-46)
29 (7-12) 55 18 10 19 (7-27) 4 (3-6) 11 (11-11) 14 (9-21) 12 14 (5-16)
2 (1-3) 5 0 2 11 (1-26) 4 (0-11) 1 (2-5) 4 (0-11) 2 3 (1-5)
23 (19-30) 58 38 59 55 (33-75) 35 (31-40) 42 (37-47) 32 (26-44) 52 45 (44-46)
9 (8-11) 26 29 39 24 (14-37) 12 (5-24) 16 (11-20) 14 (7-19) 28 22 (16-28)
10 (9-10) 7 9 20 31 (19-38) 13 (3-25) 20 (14-26) 12 (1-17) 24 19 (15-22)
6 (2-12) 30 2 9 3 (0-9) 20 (12-26) 12 (10-13) 7 (0-16) 1 15 (7-22)
29 (21-36) 48 12 7 8 (3-17) 39 (25-57) 70 (45-94) 29 (13-60) 12 18 (15-10)
6 (4-10) 34 19 21 9 (1-16) 26 (18-38) 10 (5-14) 11 (7-19) 13 34 (13-55)
*Data are total or mean (ranges) cell numbers in 10 fields.
Apoptotic eosinophils were either not or very rarely detected. On the other hand, in the epidermis, which has just single positive cells in the basal layer under normal conditions,15 we observed cleaved caspase-3–positive keratinocytes scattered throughout the epidermis (DHS) or in groups (APT and ACD). In WS, HES, BP, PF, and DH, we found positive cells along the basal and suprabasal layers. In BP, eosinophils lined the dermalepidermal junction in prebullous stadiums (Fig 4, A). In these areas, eosinophils generating EETs have been found (Fig 4, B-D). Those parts of the epidermis covering blisters in BP and DH were intensively stained with the antibody to cleaved caspase-3.
DISCUSSION In this study, we demonstrate for the first time that EETs are present in several eosinophilic skin diseases. We observed these extracellular structures in infectious skin diseases such as larva migrans and ectoparasitoses. Interestingly, EETs were also detectable in other inflammatory skin diseases, such as allergic/ reactive and autoimmune diseases associated with eosinophilia. From a mechanistic point of view, EETs might be generated by IL-5/IFN-g and eotaxins, which were clearly expressed in all these diseases and sufficient to trigger the formation of these extracellular structures in vitro.10 In contrast, in other skin diseases, such as skin neoplasms with tissue-infiltrating eosinophils, EETs were not found. However, nothing is known about the survival and enzymatic degradation of EETs in tissues so far. It is possible that the time from taking the biopsy until fixation is crucial because of the presence of tissue nucleases. Therefore, the interpretation of our negative findings as well as the sometimes low numbers of EET-releasing eosinophils should be performed with caution. Clearly, additional research is required with respect to both numbers of investigated tissues and comparison of different methods of tissue preservation. In spite of these limitations, some conclusions might be possible, at least from those data in which we observed EETs. First, the diseases in which we found extracellular DNA structures were acute dermatoses, such as infectious diseases, ACD, DHS, or acute exacerbations of WS and HES. Interestingly, in a positive reaction of APT, which resembles an acute, 2-day-old exacerbation of AD induced by an allergen, we also detected EETs, whereas none were seen in 5 specimens of lesional AD skin.
FIG 3. IL-5 and eotaxin–expressing cells in skin tissues containing EETs. A, IL-5–expressing (green) and CD4-positive (red) cells in APT. Double-positive cells (yellow) are indicated by arrows, single IL-5–expressing cells (green) by arrowheads. Single CD4-positive cells (red) are also visible. B, Eotaxin2–positive cells in BP. Eosinophils (arrows) and noneosinophilic cells (arrowhead) express eotaxin-2.
Second, most of the diseases with EETs, except WS and HES, were related to an exogenous (parasite, allergen) or endogenous (autoantigen, drug) pathogenic trigger. Taken together, these data point to the possibility that EETs are present in acute phases or exacerbations of eosinophilic skin diseases. Therefore, EETs might be generated by eosinophils that are newly recruited to the skin. The eosinophils generating EETs revealed the same size and shape compared with other eosinophils in the tissues and looked morphologically intact. They were most likely nonapoptotic because they were caspase-3–negative and had normal nuclear morphology. It should be noted that extracellular depositions was also found in the absence of DNA release. For instance, we detected ECP-positive extracellular granules in all eosinophilic skin biopsies. Moreover, we observed ECP in association with flame figures in WS and HES as well as BP and PF. This suggests that ECP release is not dependent on and can occur independently of DNA release. Alternatively, it is possible that ECP is less subject to degradation compared with DNA and therefore can be accumulated more easily. Why eosinophils release and form EETs is not clear. It can be hypothesized that these DNA structures guarantee that toxic granule proteins are effectively brought to a pathogen/antigen, which might be trapped by the DNA nets. On the other hand, it can be speculated that with the association of toxic granule proteins to
198 SIMON ET AL
J ALLERGY CLIN IMMUNOL JANUARY 2011
FIG 4. Cleaved caspase-3 expression and EETs in BP. A, Caspase-3–positive cells throughout the epidermis, beneath eosinophils (arrows) in the dermis. Nuclei were counterstained with PI. B-D, Eosinophil at the dermal-epidermal junction (dashed line) generating an EET (arrow). DNA plus ECP, overlay (B), DNA (C), ECP (D). Magnification 3400 (A), 31000 (B-D).
the DNA structures, collateral tissue damage might be limited. Indeed, the relation of eosinophil peak infiltration including the formation of EETs to apoptotic cells in the dermis was not obvious, supporting this hypothesis. On the other hand, we observed large numbers of eosinophils and EETs in proximity to apoptotic keratinocytes of patients with BP, implying that eosinophils with or without the help of EETs are involved in keratinocyte damage including blister formation. Future studies including clinical trials using specific antieosinophil therapies may bring more clarity to the pathogenic role of eosinophils in this process. We thank E. Kozlowski for technical assistance.
Key messages d
Eosinophils generating EETs are present in inflammatory skin diseases.
d
Eosinophil extracellular traps containing DNA and eosinophil granule proteins are distinct from flame figures and eosinophil granule depositions.
REFERENCES 1. Simon D, Simon HU. Eosinophilic disorders. J Allergy Clin Immunol 2007;119: 1291-300.
J ALLERGY CLIN IMMUNOL VOLUME 127, NUMBER 1
2. Simon D, Wardlaw A, Rothenberg ME. Organ-specific eosinophilic disorders of the skin, lung, and gastrointestinal tract. J Allergy Clin Immunol 2010;126: 3-13. 3. Klion AD, Nutman TB. The role of eosinophils in host defense against helminth parasites. J Allergy Clin Immunol 2004;113:30-7. 4. Frigas E, Motojima S, Gleich GJ. The eosinophilic injury to the mucosa of the airways in the pathogenesis of bronchial asthma. Eur Respir J 1991;13: S123-35. 5. Gleich GJ, Frigas E, Loegering DA, Wassom DL, Steinmuller D. Cytotoxic properties of the eosinophil major basic protein. J Immunol 1979;123:2925-7. 6. Kroegel C, Costabel U, Matthys H. Mechanism of membrane damage mediated by eosinophil major basic protein. Lancet 1987;1:1380-1. 7. Young JD, Peterson CG, Venge P, Cohn ZA. Mechanism of membrane damage mediated by human eosinophil cationic protein. Nature 1986;321:613-6. 8. Gleich GJ, Loegering DA, Bell MP, Checkel JL, Ackerman SJ, McKean DJ. Biochemical and functional similarities between human eosinophil-derived neurotoxin and eosinophil cationic protein: homology with ribonucleases. Proc Natl Acad Sci U S A 1986;83:3146-50.
SIMON ET AL 199
9. Rosenberg HF, Domachowske JB. Eosinophils, eosinophil ribonucleases, and their role in host defence against respiratory virus pathogens. J Leukoc Biol 2001;70:691-8. 10. Yousefi S, Gold JA, Andina N, Lee JJ, Kelly AM, Kozlowski E, et al. Catapult-like release of mitochondrial DNA by eosinophils contributes to antibacterial defense. Nat Med 2008;14:949-53. 11. Brinkmann V, Reichard U, Goosmann C, Fauler B, Uhlemann Y, Weiss DS, et al. Neutrophil extracellular traps kill bacteria. Science 2004;303:1532-5. 12. Yousefi S, Mihalache C, Kozlowski E, Schmid I, Simon HU. Viable neutrophils release mitochondrial DNA to form neutrophil extracellular traps. Cell Death Differ 2009;16:1438-44. 13. Simon D, H€osli S, Kostylina G, Yawalkar N, Simon HU. Anti-CD20 (rituximab) treatment improves atopic eczema. J Allergy Clin Immunol 2008;121:122-8. 14. Straumann A, Conus S, Grzonka P, Kita H, Kephart G, Bussmann C, et al. Anti-interleukin-5 antibody treatment (mepolizumab) in active eosinophilic oesophagitis: a randomised, placebo-controlled, double-blind trial. Gut 2010;59:21-30. 15. Simon D, Lindberg RLP, Kozlowski E, Braathen LR, Simon HU. Epidermal caspase-3 cleavage associated with interferon-gamma expressing lymphocytes in acute atopic dermatitis lesions. Exp Dermatol 2006;15:441-6.