Reflections on Eosinophils and Flame Figures: Title and subTitle BreakWhere There's Smoke There's Not Necessarily Wells Syndrome

Eosinophils were first described in 1879 by Paul Ehrlich, who recognized their ability to stain with acid dyes, particularly eosin. The distinctive tinctorial properties of eosinophil granules give these cells a prominence in stained sections that contrasts with their lack of diagnostic power and still undefined role in pathogenesis. Eosinophils may be seen in skin biopsy specimens from patients with various inflammatory and neoplastic disorders, but they are among the diagnostic criteria in a limited number of diseases, including Wells syndrome, angiolymphoid hyperplasia with eosinophilia, and eosinophilic pustulosis.^{1 - 2}

Eosinophils are associated with pathophysiologic mechanisms that strongly implicate them in the pathogenesis of many cutaneous diseases. Eosinophil involvement has been well documented in cutaneous disorders characterized by swelling,³ particularly chronic urticaria,⁴ pressure urticaria,⁵ and episodic angioedema.⁶ Cutaneous edema was the common clinical thread in the first 4 cases reported by Wells.⁷ Eosinophils also have a role in eczematoid dermatoses such as atopic dermatitis, prurigo nodularis, and onchocercal dermatitis⁸ ; blistering diseases such as pemphigoid and incontinentia pigmenti; vasculitis,¹ with evidence that eosinophils are a marker for drug-induced disease⁹ ; and fibrosing conditions such as scleroderma.¹⁰ The classification of eosinophil-associated diseases has been hampered by properties of eosinophils that are not directly comparable with those of other inflammatory cells. Eosinophils are attracted into tissues and are activated by cytokines and other stimuli; interleukin 5 is a principal factor inducing eosinophil proliferation, differentiation, and activation.¹¹ Once activated, eosinophils degranulate via 3 mechanisms: piecemeal, vesicular transport, and cytolysis.¹¹ The eosinophils' toxic granule proteins and other phlogistic factors are then released into tissues. Eosinophil-associated lesions exhibit 1 or more of the following 3 patterns of eosinophil involvement: intact eosinophils in varying numbers, intact eosinophils with extracellular granule protein deposition proportionate to the numbers of intact cells, and prominent extracellular eosinophil granule protein deposition in the presence of few or no intact cells. In many eosinophil-associated diseases, the third pattern is found and is considered evidence that eosinophils are attracted into tissues and that their granule contents are released through cytolysis. Thus, the eosinophils are no longer morphologically identifiable as cells, but their distinctive granules can be recognized using electron microscopy and their granule proteins can be detected in tissues by special stains (Figure).

The phlogistic potential of activated eosinophils includes generation of toxic oxygen metabolites; production of lipid-derived mediators including leukotrienes, prostaglandins, and platelet-activating factor; release of enzymes including collagenase and matrix metalloproteinase–9; and secretion of many cytokines.¹¹ These properties are in addition to the deposition of granule proteins that are toxic to cells and tissues and can persist in tissues for long periods (Table).^{11 ,13} Eosinophils also have numerous membrane components that further involve them in inflammatory processes,^{11 - 12} including receptors that enable them to become activated by their own secreted products. Membrane receptors on eosinophils also facilitate their activation by mediators from other cells, including mast cells. Secreted eosinophil products affect physiologic processes. For example, eosinophil major basic protein, which is extensively deposited in atopic dermatitis,¹⁴ blocks the feedback control of acetylcholine on type 2 muscarinic (M2) receptors,¹⁵ causing uninhibited acetylcholine release. In animal models, this blockage is associated with hyperactive airways¹⁶ ; extrapolating this phenomenon to the skin may explain pharmacophysiologic abnormalities such as white dermographism. In addition, eosinophil granule proteins and leukotriene C₄, along with platelet-activating factor produced by eosinophils, directly provoke vasopermeability. Eosinophils also likely work in concert with other cells. Prostaglandin E₂ from eosinophils, injected with histamine from mast cells, produces a larger wheal, with pruritus of longer duration, than does either prostaglandin E₂ or histamine alone. Substance P, synthesized and released from eosinophils, and eosinophil granule proteins, alone or with oxygen metabolites, induce histamine release from mast cells. It is, therefore, reasonable to conclude that the myriad products and activities of eosinophils make these cells important factors in a variety of diseases.¹¹ In particular, eosinophil degranulation probably contributes to flame figure formation in Wells syndrome.^{17 - 18} Moreover, a correlation between disease severity and levels of both interleukin 5 and eosinophil cationic protein has been found in Wells syndrome¹⁹ and interleukin 5 is implicated in hereditary eosinophilic cellulitis.²⁰

As eosinophil biology continues to be understood, we will be able to better classify eosinophil-associated skin diseases. Recent advances in our understanding of the hypereosinophilic syndromes (HES) provide an emerging paradigm. In the past, the HES were classified by broad-based inclusion criteria, including the following: peripheral blood eosinophilia of 1500 eosinophils per microliter or greater for at least 6 months or for less than 6 months with evidence of organ damage, signs and symptoms of multiple organ involvement, and no evidence of parasitic or allergic disease or of other known causes of peripheral blood eosinophilia. Eosinophils are thought to cause much of the end-organ damage in all HES variants through elaboration of their products, and clinical improvement usually parallels a decrease in the eosinophil count. Two specific types of HES, myeloproliferative and lymphocytic HES, have recently been characterized, along with clinical presentations of HES that are not subclassified (unclassified HES).²¹ Myeloproliferative HES has many features of eosinophilic leukemia, but the lack of eosinophil immaturity and absence of clonal expansion preclude classification as leukemia. With the finding that patients with HES, including those with mucosal ulcers and an aggressive disease course, dramatically respond to imatinib mesylate (Gleevec; Novartis Pharmaceuticals Corp, East Hanover, NJ),²² the target of imatinib, a tyrosine kinase inhibitor, was quickly identified.²³ The novel kinase expressed in HES results from an 800-kb deletion on chromosome 4, which produces a fusion gene. This fusion gene is composed of the kinase domain of platelet-derived growth factor receptor-α linked to a previously uncharacterized gene resembling Fip1, an essential component of the Saccharomyces cerevisiae polyadenylation mechanism. The resultant product, FIP1L1–platelet-derived growth factor receptor-α fusion tyrosine kinase, is a constitutively activated tyrosine kinase that transforms hematopoietic cells in vitro and in vivo and is exquisitely sensitive to inhibitory effects of imatinib. Imatinib mesylate produces complete hematologic remission in patients with HES who express this mutation and regression of their associated skin lesions. With imatinib treatment, the mutation often becomes undetectable. Cutaneous disorders in patients with myeloproliferative HES include severe mucosal ulcers and sequelae from thromboembolysis. Patients with lymphocytic HES exhibit T-cell clonality.²⁴ The clonal lymphocyte populations in such patients frequently have unusual surface phenotypes including CD3⁺CD4⁻CD8⁻ and CD3⁻CD4⁺. These T cells display activation markers such as CD25, and they secrete T-helper 2 cell cytokines including high levels of interleukin 5. The lymphocytic HES variant usually follows a benign course, and T-cell clones remain stable for years; however, CD3⁻CD4⁺ T cells and other clonal T cells may undergo progressive transformation and evolve into lymphoma. Throughout their disease course, patients with lymphocytic HES often have severe pruritus, eczema, erythroderma, urticaria, and angioedema. Patients with unclassified HES, such as those with episodic angioedema and eosinophilia (Gleich syndrome)⁶ and with the constellation of nodules, eosinophilia, rheumatism, and dermatitis (NERDS),²⁵ also have developed T-cell clones. To summarize, we know that there are separate eosinophil-associated diseases within the HES spectrum that are defined pathogenetically and that carry implications for treatment and understanding of the disease course.

Thirty-five years ago, Wells⁷ described 4 patients who had recurrent edematous infiltrative plaques resembling cellulitis at onset and later morphea. He termed the condition recurrent granulomatous dermatitis with eosinophilia. Histologically, the lesions were characterized by diffuse infiltration with eosinophils, histiocytes, and foci of amorphous or granular material associated with connective tissue fibers of the dermis, which he named flame figures. In 1979, Wells and Smith²⁶ introduced the term eosinophilic cellulitis in their description of 8 additional patients with the syndrome. They also discussed 9 patients with the histologic features of eosinophilic cellulitis associated with a variety of other clinical diagnoses, including pemphigoid, eczema, and tinea.²⁶ This and subsequent reports of flame figures in lesions from patients with arthropod bites, parasitic infections, drug reactions, and other inflammatory conditions^{2 ,27} indicate that the flame figure per se is not diagnostic of one clinical disorder. However, the converse, namely, that a diagnosis of Wells syndrome can be made in the absence of flame figures, should be met with skepticism.

In this issue of the ARCHIVES, Caputo et al²⁸ describe 19 patients with a broad spectrum of clinical features. Approximately half had some degree of peripheral blood eosinophilia (percentages of eosinophils were provided, but not total leukocyte or eosinophil counts) and all had eosinophils in skin lesions. Some form of granulomatous infiltrate was found in 14 of 19 patients, and edema in 10 patients. The histologic pattern consisted of eosinophils plus microgranulomas or granulomatous changes (with or without edema) in 5 patients; only edema and eosinophil infiltration were found in another 5 patients. Flame figures were noted in biopsy specimens from 9 patients (fewer than half). The authors comment that the “concept of Wells syndrome has been somewhat blurred in the recent past because of its clinical polymorphism and suggestive but not distinctive histopathologic characteristics.”^{28 (p1160)} They “defend the view on Wells syndrome as a distinct disease”^{28 (p1160)} and state that they have “categorized 7 clinical variants of the disease.”^{28 (p1161)} As noted, eosinophil infiltration and degranulation is a well-documented aspect of many clinical syndromes characterized by cutaneous swelling, some of which must be considered in the clinical differential diagnosis with Wells syndrome. The histologic combination of granulomas and eosinophils prompts another set of differential diagnoses. Therefore, in light of what is known about the role of eosinophils in cutaneous diseases, what criteria should be used in evaluating this series of patients? Based on the article, one could postulate that the spectrum of Wells syndrome is wider than previously appreciated. However, one might also consider that the authors are including multiple eosinophil-associated diseases within Wells syndrome, perhaps and understandably because there is no clear way to otherwise categorize these dermatoses.

How many features are necessary to make up a syndrome, or, rather, how few? Because so many lesions are included under the umbrella of edema and eosinophilia, a minimum requirement for Wells syndrome should be the histologic finding of flame figures, with the recognition that this abnormality is not diagnostic or alone sufficient for diagnosis. A syndrome may exhibit some but not unlimited heterogeneity of criteria. Establishing the boundaries that define common diseases is easier than establishing the range of features that characterize syndromes that have been described in a sprinkling of small series or individual cases reported in the literature over decades. With respect to Wells syndrome in particular and eosinophil-associated skin disease in general, before attempting to expand the criteria for inclusion, it may be preferable to wait for our understanding of eosinophil biology to catch up. Clarifying the full effect of eosinophil involvement, including the role of eosinophil products and the interaction of eosinophils with and relative importance of other infiltrating cells, especially T cells and histiocytes, may offer a molecular means to more accurately categorize these cutaneous syndromes.

Correspondence: Dr Leiferman, Department of Dermatology, University of Utah, Room 4B454, School of Medicine, 30 N 1900 E, Salt Lake City, UT 84132-2409 (kristin.leiferman@hsc.utah.edu).

Financial Disclosure: Dr Leiferman has served on the Dermatology Advisory Panel, Beiersdorf AG; been a consultant for and received a research grant from GlaxoSmithKline Inc (co-investigator for MHE100185 Phase III Multicenter Treatment Study of Hypereosinophilic Syndromes With Mepolizumab, 2004-2006); received an honorarium from Novartis Pharmaceuticals Corp; and been a consultant for 3M. Her husband has been a consultant for GlaxoSmithKline Inc; received honoraria and research grants from GlaxoSmithKline Inc (principal investigator for MHE100185 Phase III Multicenter Treatment Study of Hypereosinophilic Syndromes With Mepolizumab, 2004-2006), Novartis Pharmaceuticals Corp, and Genentech Inc; and held stock in Ception and Eosynos Inc.