You are in: eMedicine Specialties > Pediatrics: Hematology > Hematology HistiocytosisArticle Last Updated: Apr 12, 2007AUTHOR AND EDITOR INFORMATIONSection 1 of 11
  Author: Cameron K Tebbi, MD, Medical Director, Department of Pediatric Hematology-Oncology, Tampa Children's Hospital Cameron K Tebbi is a member of the following medical societies: American Association for Cancer Research, American Society of Clinical Oncology, and American Society of Pediatric Hematology/Oncology Coauthor(s): Robert J Arceci, MD, PhD, King Fahd Professor, Division of Pediatric Oncology, Johns Hopkins University School of Medicine; Thomas W Loew, MD, Director, Clinical Associate Professor of Pediatrics, Pediatric Hematology/Oncology Subspecialty Training Program, University of Iowa Hospitals and Clinics Editors: Gary R Jones, MD, Associate Medical Director, Clinical Development, Berlex Laboratories; Mary L Windle, PharmD, Adjunct Assistant Professor, University of Nebraska Medical Center College of Pharmacy, Pharmacy Editor, eMedicine.com, Inc; Gary D Crouch, MD, Program Director of Pediatric Hematology-Oncology Fellowship, Department of Pediatrics, Associate Professor, Uniformed Services University of the Health Sciences; Helen SL Chan, MBBS, FRCP(C), FAAP, Senior Scientist, Research Institute; Professor, Division of Hematology/Oncology, Department of Pediatrics, The Hospital for Sick Children, University of Toronto, Canada; Max J Coppes, MD, PhD, MBA, Executive Director, Center for Cancer and Blood Disorders, Children's National Medical Center Author and Editor Disclosure Synonyms and related keywords: histiocytosis, LCH, Langerhans cell histiocytosis, eosinophilic granuloma, Hand-Schüller-Christian disease, Letterer-Siwe disease, histiocytosis X, familial hemophagocytic lymphohistiocytosis, HLH, FHLH, sinus histiocytosis, SHML, juvenile xanthogranuloma, JXG, reticulohistiocytoma, acute monocyte leukemia, malignant histiocytosis, sinus hyperplasia, hepatosplenic T-cell lymphoma, histiocytic necrotizing lymphadenitis, dendritic lymphadenitis, follicular lymphadenitis, cutaneous LCH INTRODUCTIONSection 2 of 11
  BackgroundHistiocytosis encompasses a group of diverse disorders with the common primary event of the accumulation and infiltration of monocytes, macrophages, and dendritic cells in the affected tissues. Such a description excludes diseases in which infiltration of these cells occurs in response to a primary pathology. The clinical presentations vary greatly, ranging from mild to life threatening. Although nearly a century has passed since histiocytic disorders were recognized (Lampert, 1998), their pathophysiology remains an enigma, and treatment is nonspecific. These problems underscore the need for an improved understanding of the etiology and pathogenesis of histiocytosis. Over the past 50 years, the nomenclature used to describe histiocytic disorders has substantially changed to reflect the wide range of clinical manifestations and the variable clinical severities of some disorders that have the same pathologic findings. For example, the entity now referred to as Langerhans cell histiocytosis (LCH) was initially divided into eosinophilic granuloma, Hand-Schüller-Christian disease, and Letterer-Siwe disease, depending on the sites and severity. Later, these were found to be manifestations of a single entity and were unified under the term histiocytosis X (Lichtenstein, 1953). Most recently, this designation was changed to LCH to reflect the recognition of the primary cell involved and the pathophysiology of the disease (Chu, 1987; Nezelof, 1973). Although several histiocytic disorders are briefly discussed in this article (see History), the primary focus is on LCH.PathophysiologyImproved understanding of the pathology of histiocytic disorders requires knowledge of the origins, biology, and physiology of the cells involved. Normal histiocytes originate from pluripotent stem cells, which can be found in bone marrow (Katz, 1979). Under the influence of various cytokines (eg, granulocyte-macrophage colony-stimulating factor [GM-CSF], tumor necrosis factor-alpha [TNF-alpha], interleukin [IL]-3, IL-4), these precursor cells can become committed and differentiate to become a specific group of specialized cells. Committed stem cells can mature to become antigen-processing cells, with some possessing phagocytic capabilities. These cells include tissue macrophages, monocytes, dendritic cells, interdigitating reticulum cells, and Langerhans cells. Pluripotent stem cells can also be committed to produce dendritic cells. Each category of histiocytosis can be traced to reactive or neoplastic proliferation and disorder of cells in one of these groups (Gonzalez, 1990). The importance of dendritic cells in presenting antigens to T and B lymphocytes is increasingly recognized. Dendritic cells appear to develop in several pathways. Immature dendritic cells respond to GM-CSF (not to macrophage colony-stimulating factor [M-CSF]) and become committed to generating dendritic cells, which are “professional” antigen-presenting cells (APCs) (Steinman, 1997). These cells can capture antigen and migrate to lymphoid organs, where they present the antigens to naive T cells (Caux, 1995). Dendritic cells are also efficient stimulators of B-cell lymphocytes (Banchereau, 1998). Effective induction of antigen-specific T-cell responses requires interaction between the dendritic cells and T lymphocytes to prime the latter cells for expansion and appropriate function. The surface of the APC contains 2 peptide-binding proteins, ie, major histocompatability complex (MHC) classes I and II, which can stimulate cytotoxic T (TC) cells, regulatory T (Treg) cells, and helper T (TH) cells (Grewal, 1996). Although circulating T-cell lymphocytes can independently recognize antigens, their number is small. Dendritic cells display a large amount of MHC-peptide complexes at their surface and can increase the expression of costimulatory receptors and migrate to the lymph nodes, spleen, and other lymphoid tissues, where they activate specific T cells. The first signal may involve interaction between an MHC I– and/or MHC II–bound peptide on an APC with the T-cell receptor (TCRs) on the effector lymphocytes. TCRs can recognize fragments of antigen attached to MHC on the surface of an APC. Costimulatory interaction (ie, second signal) is between CD80(B7.1)/CD86(B7.2) on the dendritic cell, with CD28 on the T cells (Macatonia, 1995; Scheicher, 1995; Sprent, 1997). A combination of the 2 signals activates the T cell, resulting in upregulation of the expression of CD40L, which, in turn, can interact with the dendritic cell–expressed CD40 receptor (Grewal, 1996). In perforin-deficient mice, abnormally heightened cytokine production by T cells is due to overstimulation by APCs after a viral infection (Lykens, 2006). This cell-to-cell interaction between dendritic cells and T cells generates an antigen-specific T-cell response. The effective function of antigen presentation by dendritic cells is presumed to reflect that these cells, in addition to MHC molecules, express a high density of other costimulatory factors. Dendritic cells can produce several cytokines, including IL-12, which is critical for the development of TH1 cells from naive CD4+ T cells (Macatonia, 1995; Scheicher, 1995). Ligation of CD40 on dendritic cells triggers the production of large amounts of IL-12, which enhances T-cell stimulatory capacity. This observation suggests that feedback to dendritic cells results in signals that are critical for induction of immune responses. The nature of the latter interaction and requirement for optimal dendritic cell activation is not fully understood. Dendritic cells in culture derived from human blood monocytes exposed to GM-CSF and IL-4 followed by maturation in a monocyte-conditioned medium have heightened antigen-presenting activity (Romani, 1996; Reddy, 1997). Monocyte-conditioned media contain critical maturation factors to catalyze this process. Dendritic cells are present in tissues in a resting state and cannot stimulate T cells. Their role is to capture and phagocytize antigens, which, in turn, induce their maturation and mobilization (Inaba, 1993; Moll, 1993; Reis e Sousa, 1993; Svensson, 1997). Immature dendritic cells reside in blood, lungs, spleen, heart, kidneys, and tonsils, among other tissues. Their function is to capture antigen and migrate to the draining lymphoid organs to prime CD4+ and CD8+ T cells. In the process of their function, these cells mature and increase their capacity to express costimulatory receptors and decrease their capacity to process antigen. These cells can phagocytize, forming pinocytic vesicles for sampling and concentrating their surrounding medium, which is called macropinocytosis (Sallusto, 1995). Immature dendritic cells express receptors that mediate endocytosis, including C-type lectin receptors, such as the macrophage mannose receptor and DEC205, FC-gamma, and FC-epsilon receptors (Steinman, 1997). Microbial components, as well as IL-1, GM-CSF, and TNF-alpha, have an important role in cellular response (Dinarello, 1991 and 1993; Strieter, 1994) and can stimulate maturation of dendritic cells, whereas IL-10 opposes it (Buelens, 1997). Mature dendritic cells possess numerous fine processes (veils, dendrites) and have considerable mobility. These cells, rich in MHC classes I and II, have abundant molecules for T-cell binding and co-stimulation, which involves CD40, CD54, CD58, CD80/B7-1, and CD86/B7-1. Mature dendritic cells express high levels of IL-12. High levels of CD83 (a member of the immunoglobulin [Ig] superfamily), and p55 or fascin (an actin-bundling protein) are present in these cells, as opposed to the low levels that are present in the immature cells (Steinman, 1997). IL-1 enhances dendritic cell function. This effect appears to be indirect and due to activation of TNF receptor–associated factors (TRAFs). Mature dendritic cells also express high levels of the NF-kappaB family of transcriptional control proteins. These proteins regulate the expression of several genes encoding inflammatory and immune proteins. Signaling by means of the TNF-receptor family (eg, TNF-R, CD40, TNF-related activation-induced cytokine [TRANCE], receptor activator of NF-kappaB [RANK]) activates NF-kappaB. Immunologic response of dendritic cells to a given antigen partly involves the triggering of signal-transduction pathways involving the TNF-R family and TRAFs. Information regarding the fate of dendritic cells after these events is sparse. Dendritic cells disappear from the lymph nodes 1-2 days after antigen presentation, possibly because of apoptosis (Ingulli, 1997). CD95 (Fas) is suggested to have a role in the death of the dendritic cell (Bjorck, 1997; Koppi, 1997). Although dendritic cells express CD95, CD95 ligation does not induce apoptosis (Willems, 2000). Experiments indicate that immature dendritic cells are partially susceptible to death receptor–mediated apoptosis (Leverkus, 2000). TNF-related apoptosis-inducing ligand (TRAIL) may bind to 5 separate receptors. Functional cytoplasmic death domains characterize TRAIL-R1 receptors, TRAIL-R2 receptors, and CD95 receptors. In contrast, TRAIL-R3 is a membrane-anchored truncated receptor, and TRAIL-R4 does not have a functional death domain. Dendritic cells express CD95, TRAIL-R2, and TRAIL-R3 in comparative levels. Similar to the role of CD95L, that of TRAIL-mediated apoptosis of mature dendritic cells has been controversial. Data regarding in vitro TRAIL-mediated apoptosis in these cells has been reported (Wang, 1999) and disputed (Leverkus, 2000). Mature dendritic cells are resistant to TRAIL- and CD95L-mediated apoptosis (Leverkus, 2000). C-FLIP, which is the caspase-8 inhibitory protein capable of inhibiting death receptor–mediated apoptosis, is highly expressed in mature dendritic cells, whereas only low levels are found in immature cells (Leverkus, 2000). Overexpression of C-FLIP inhibits signals of death receptor (Irmler, 1997). C-FLIP expression on dendritic cells is upregulated during maturation (Willems, 2000). Note that engagement of CD95 on immature dendritic cells by CD95L induces phenotypic and functional maturation of these cells. In addition, a CD95-activated dendritic cell upregulates the expression of MHC class II and costimulatory receptors, which is essential for the function of these cells. Furthermore, such engagement upregulates the expression of dendritic-cell lysosome-associated membrane protein (DC-LAMP) and causes the secretion of proinflammatory cytokines, including IL-1 beta and TNF-alpha (Rescigno, 2000). The function of normal Langerhans cells is cutaneous immunosurveillance. These cells can migrate to the regional lymph nodes and potentially present antigen to paracortical T cells and cause their transformation to interdigitating dendritic cells. Some cancer cells disrupt dendritic-cell function, blocking the development of tumor-specific immune responses and allowing tumors to evade recognition (Kiertscher, 2000; Pirtskhalaishvili, 2000). To counteract this effect, dendritic cells may produce the antiapoptotic protein Bcl-xL (Pirtskhalaishvili, 2000). Stimulation of dendritic cells by CD154, IL-12, or IL-15 increases expression of this protein. The information gained from normal physiology of dendritic cells may potentially lead to treatment modalities for histiocytic disorders. FrequencyInternationalThe incidence of LCH is 4.0-5.4 per million population. However, because many bone and skin lesions may not be diagnosed as LCH, this rate may be an underestimate (Carstensen, 1993). The estimated incidence of neonatal LCH, determined by using the population-based German Childhood Cancer Registry, is 1-2 per million neonates (Minkov, 2005). Mortality/MorbiditySee History. SexThe overall male-to-female ratio is 1.5:1. The male-to-female ratio in individuals who have a single organ system involvement is 1.3:1, and the male-to-female ratio in individuals with multisystem disease 1.9:1 (Salotti, 2006). AgeThe disease can occur in individuals of any age. LCH can be congenital (Stiakaki, 1997) or may occur in adults (Malpas, 1998). The disease is seen in all age groups, ranging from neonates to adults (Arico, 2003; Gotz, 2004; Subramanian, 2004; Minkov, 2005). The incidence peaks in children aged 1-3 years (Nicholson, 1998). In one study, the age at diagnosis was 0.09-15.1 years. Patients with single system involvement were older (0.1-15.1 y) than those with multisystem involvement (0.09-14.8 y) (Salotti, 2006). CLINICALSection 3 of 11
HistoryClinical presentationLangerhans cell histiocytosis (LCH) can be local and asymptomatic, as in isolated bone lesions, or can involve multiple organs and systems, with clinically significant symptoms and consequences. The clinical manifestations depend on the site of the lesions and on the organs and systems involved and their functions (see Physical). ClassificationClassification of diseases involving histiocytic and dendritic cells is difficult, and classification systems must include a broad range of diseases. Therefore, most systems have been incomplete and arbitrary. The location of lesions and the extent of the disease substantially affect the course of the disease and the patient's prognosis. Decisions regarding treatment are based on the extent of the disease. Classification of the World Health Organization Table 1 shows the classification of histiocytic and dendritic cell disorders the World Health Organization (WHO) proposed (Harris, 1999). Table 1. Classification of Histiocytosis Syndromes in Children
The WHO classification of neoplastic disorders of histiocytes and dendritic cells is as follows:
Classification of the Histiocyte Society Table 2 shows the working classification of histiocytosis syndromes from the Histiocyte Society. Table 2. Histiocyte Society Classification of Histiocytosis Syndromes
The following, adapted from the Writing Group of the Histiocyte Society (Chu, 1987), describes confidence levels for the diagnosis of class I LCH:
Previous and other classifications LCH formerly was divided into 3 disease categories: eosinophilic granuloma, Hand-Schüller-Christian disease, and Letterer-Siwe disease, depending on the severity and extent of involvement. This classification and its related risk groups no longer are used. Systems based on these categories were meant to reflect the extent of involvement and its relationship to the patient's prognosis (Lahey, 1962; Lucaya, 1971). A clinical-grouping system for LCH based on age, extent of the disease, and organ dysfunction, as once constructed (Arceci, 1995), can provide a means to compare patient data and prognoses. Various categories, such as restricted and extensive multiorgan involvement, have also been proposed. Data regarding treatment results are needed to validate any classification system. The Histiocyte Society developed a classification based on risk groups that arose from the first and second international (LCH I and II, respectively) trials of chemotherapy (Broadbent 1998; Gadner, 2001). At-risk organs and systems identified in those trials included the liver, lung, spleen, and hematopoietic system. This risk classification is currently being used in the treatment protocol of the third international study for LCH (LCH III). Patients are stratified into 3 groups: (1) at-risk patients, or those with multisystemic involvement including 1 or more at-risk organs; (2) low-risk patients, or those with multisystem involvement not including at-risk organs; and (3) other patients, or those with single-system multifocal bone disease or localized involvement of special sites (intraspinal extension or involvement of the paranasal, parameningeal, periorbital, or mastoid region) that can lead to persistent soft-tissue swelling. In the trial, at-risk patients are randomly assigned to 1 of 2 treatment arms. Low-risk patients receive standard therapy for 6-12 months, and those with multifocal bone or special-site involvement receive the standard therapy for 6 months. Other histiocytosesAlthough this article focuses mainly on LCH, other histiocytoses are as follows:
PhysicalWhen the disease is focal, establishing the diagnosis of LCH depends on a high level of suspicion. When advanced multisystem involvement is observed, diagnosis is often easy. Adequate workup to determine the extent of the disease and possible complications is essential. Biopsy and pathologic evaluation are needed to establish the diagnosis.
CausesThe causes of most histiocytoses are not known. Factors implicated in the etiology and pathophysiology of these disorders include infections, especially viral infections (Leahy, 1993); cellular and immune dysfunction (Hage, 1993; Willman, 1994; Yu, 1994), including dysfunction of lymphocytes and cytokines (Kannourakis, 1994; Egeler, 1999 and 1999); neoplastic mechanisms; genetic factors (Ornvold, 1990; Katz, 1991; Yu, 1995; Corbeel, 2004; Chen, 1994; Kuwabara, 1990; Halton, 1992; Mader, 1996); cellular adhesion molecules (Ruco, 1993; de Graaf, 1994 and 1995); and their combinations. Although human herpes virus 6 has been found in lesions of LCH, its etiologic significance has been questioned (Ruco, 1993; McClain, 1994; Slacmeuler, 1999). Extensive searches for evidence of viral infection have been unrevealing (Mierau, 1994).
DIFFERENTIALSSection 4 of 11
Acute Lymphoblastic Leukemia Acute Myelocytic Leukemia Anemia, Chronic Atopic Dermatitis Craniopharyngioma Diabetes Insipidus Diaper Dermatitis Lymphadenopathy Lymphohistiocytosis Lymphoproliferative Disorders Mastoiditis Non-Hodgkin Lymphoma Osteomyelitis Otitis Media Splenomegaly
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Type of Study | Study | Involvement | With Monostotic Lesion | |
| Laboratory | Hemoglobin and/or hematocrit | Monthly | Every 6 mo | None |
| Leukocyte count and differential cell count | Monthly | Every 6 mo | None | |
| Liver function tests* | Monthly | Every 6 mo | None | |
| Coagulation studies | Monthly | Every 6 mo | None | |
| Urine osmolality test after overnight water fast | Every 6 mo | Every 6 mo | None | |
| Radiography | Chest, posteroanterior and lateral | Monthly | Every 6 mo | None |
| Skeletal survey | Every 6 mo | None | Once at 6 mo | |
*Measurements of alanine transaminase (ALT), aspartate transaminase (AST), and alkaline phosphatase.
Table 4. Indication for Laboratory Evaluations Based on Findings in LCH
| Evaluation | Indication | Follow-Up Interval |
| Bone-marrow aspiration biopsy | Anemia, leukopenia, or thrombocytopenia | 6 mo |
| Pulmonary function tests | Abnormal chest radiographic findings, tachypnea, intercostal retractions | 6 mo |
| Lung biopsy after bronchoalveolar lavage, if available* | Abnormal findings on pretreatment chest radiography to rule out infection | None |
| Small-bowel series and biopsy | Unexplained chronic diarrhea, failure to thrive, malabsorption | None |
| Hepatic ERCP, angiography, or biopsy | High liver enzyme levels and hypoproteinemia not caused by protein-losing enteropathy to rule out active LCH vs liver cirrhosis | When all evidence of disease resolves but hepatic dysfunction persists |
| IV gadolinium-enhanced MRI of brain and hypothalamic-pituitary | Visual, neurologic, hormonal abnormalities | 6 mo |
| Panoramic radiography of the teeth, mandible, and maxilla; consultation with an oral surgeon | Oral involvement | 6 mo |
| Endocrine investigation | Growth failure, diabetes insipidus, hypothalamic syndromes, galactorrhea, precocious or delayed puberty; hypothalamic and/or pituitary abnormality on CT or MRI | None |
| Consultation with an audiologist and an otolaryngologist | Aural discharge, impaired hearing | 6 mo |
*Diagnostic findings on bronchoalveolar lavage obviate lung biopsy.
Regardless of the clinical severity, the histopathology of LCH is generally uniform. To some extent, the location and age of the lesion may influence the histopathology of the disease (Schmitz, 1998). Early in the course of the disease, lesions tend to be cellular and contain aggregates of pathologic Langerhans cells (PLCs), intermediate cells, interdigitating cells, macrophages, T cells, and giant histiocytes. Mitotic figures number 0-23 per 10 high-power fields (Risdall, 1983).
Multinucleated giant cells are common, and some may exhibit phagocytosis. Lesions may also include eosinophils, necrotic cells, and LCH cells. With time, the cellularity and number of LCH cells are reduced, and macrophages and fibrosis become eminent. The infiltrates tend to destroy epithelial cells. Table 5 shows the phenotypes and cell-marker characteristics of LCH.
Table 5. Cell Markers and Phenotypes of Histiocytic and Related Disorders
| Cell Marker | LCH | SHML | Follicular Dendritic Tumor | Histiocytic Sarcoma | Acute Monocytic Leukemia | Anaplastic Large-Cell Lymphoma |
|---|---|---|---|---|---|---|
| CD1a | + | - | - | - | - | - |
| CD4 | + | + | - | + | + | + |
| CD21 | - | +/- | + | - | - | - |
| CD25 | - | + | - | + | + | ++ |
| CD30 | - | - | - | - | - | ++ |
| CD35 | - | + | + | - | - | - |
| CD45 | - | + | - | +/- | + | +/- |
| CD68 | - | + | - | + | + | +/- |
| ALK-1 | - | - | - | - | - | + |
| S-100 | + | + | - | +/1 | - | - |
| Lysozyme | - | + | - | + | + | - |
Table 6 shows specialized stains for diagnosing these disorders, and Table 7 shows labeling pattern of histiocytes and dendritic cells.
Table 6. Stains for Diagnosing Histiocytosis
| Type of Test | Stain | Mononuclear Phagocytic System | Langerhans Cells | Interdigitating Dendritic Cells | Dendritic Reticulum Cells |
| Frozen-section enzyme histochemistry | Nonspecific esterase | - | - | - | - |
| Acid phosphatase | + | - | - | - | |
| ATPase | - | + | + | - | |
| Lambda-mannosidase | - | + | - | - | |
| 5' nucleotidase | - | - | - | + | |
| Immunohistochemistry | CD14 (Leu M3/MY4) | + | + | + | + |
| CD11 C (Leu M5) | + | + | + | + | |
| CD68 (EBM 11) | + | - | - | - | |
| CD1a | - | + | + | - | |
| Paraffin-section immunohistochemistry | HLA-DR | + | + | + | + |
| CD68 | + | - | - | - | |
| Mac 387 | + | - | - | - | |
| Lysozyme | + | - | - | - | |
| Alpha-antitrypsin | + | - | - | - | |
| S-100 | - | + | + | - | |
| Peanut agglutinin | Diffuse | Halo and dot | Halo and dot | - |
Note.—ATPase = adenosine triphosphatase; HLA = human leukocyte antigen.
Table 7. Labeling Pattern of Histiocytes and Dendritic Cells
| Marker | Histiocytes | Langerhans Cells | Interdigitating Cells | Follicular Dendritic Cells |
|---|---|---|---|---|
| CD1a | 0 | S | 0 | 0 |
| S-100 protein | 0 | S | S | W |
| HLA-DR | S | S | S | W |
| CD45 | S | W | W | 0 |
| Alpha-naphthyl acetate esterase | S | W | W | W |
| ATPase | W | S | S | S |
| Fascin | 0 | 0 | S | S |
| R4/23 | 0 | 0 | 0 | S |
Note.—0 = no staining; S = strong and constant; W = weak or inconstant.
Langerhans cells express CD1a antigen, HLA-DR, and a subunit S-100 protein (see Image 5). Upon morphologic evaluation, cells are 12 µm in diameter with moderately abundant cytoplasm and a medium-sized folded, indented, or lobulated nucleus that has vesicular chromatin with 1-3 nucleoli and an elongated central groove producing a coffee-bean appearance (see Image 6 and Image 8) (Favara, 1987; Dehner, 1991).
The histopathology of the LCH does not appear to be prognostic of the outcome of the disease (Risdall, 1983). The aggregation of Langerhans cells is observed in various disorders, such as lymphomas (eg, Hodgkin disease), lung cancer, and thyroid cancer. However, these disorders are secondary and resolve with control of the primary disorder (Neumann, 1986; Egeler, 1993).
In LCH, the cytoplasm and, rarely, the nucleus contain the characteristic structures termed Birbeck granules (see Image 7).These trilaminar organelles are 190-360 nm long and approximately 33 nm wide, with a central striated line. These are derived from cytoplasmic membrane and are involved in receptor (T6)-mediated and non–receptor-mediated endocytosis. An electron microscopic finding of racquet-shaped granules in the cells can be helpful in confirming the pathologic diagnosis. Relatively nonspecific findings include cytoplasmic, trilaminar, membranous loops and laminated substructures of lysosomes (Mierau, 1982).
Birbeck granules are the products of internalization of complexes originating from cell-membrane antigens and corresponding antibodies. CD1a antigen can be detected in paraffin-embedded tissues to provide for reliable diagnostic testing (Krenacs, 1993; Emile, 1995). ATPase peanut lectin and alpha-D-mannoside can be positive in the dendritic reticulum. An electron microscopic finding of racquet-shaped granules in the cells can be helpful to confirm the pathologic diagnosis (Writing Group of the Histiocyte Society, 1987). Enzyme histochemistry and immunocytochemistry can also aid in the diagnosis of histiocytosis (Malone, 1991).
The organs and tissues most commonly involved are the bones, skin, lymph nodes, bone marrow, lungs, hypothalamic-pituitary axis, spleen, liver, GI tract, and orbits. Multisystemic involvement is common (Surico, 2000). Bones can be involved in isolation or as a part of multisystemic disease. The skull or large bones are often involved. Bone lesions may contain an accumulation of eosinophils, multinucleated giant histiocytes, necrosis, and hemorrhage. The term eosinophilic granuloma was previously used to describe single bone lesions of LCH.
Cutaneous involvement can also be singular or can be a part of generalized involvement (Lookingbill, 1984; Hashimoto, 1991; Johno, 1994; Munn, 1998). A spontaneously regressing nodular form of cutaneous LCH is reported in infants; it involves deep dermis with a nodular aggregate of histiocyte and is called congenital self-healing reticulohistiocytosis (Kapila, 1985; Hashimoto, 1973 and 1999). In general, skin lesions have a pattern of diffuse or multifocal nodular aggregation of PLCs deep in the papillary dermis; destruction of epidermal-dermal interface; and infiltration of histiocytes, T-cell lymphocytes, and eosinophils. The lymph nodes and thymus can be involved as a primary site or as a part of multiorgan and systemic involvement. The most common sites are the cervical, inguinal, axillary mediastinal, and retroperitoneal areas (Motoi, 1980; Seyrig, 2005).
Five histologic motifs have been recognized. These include sinusoidal, limited sinusoidal, epithelioid granulomatous, partial effacement, and total effacement. However, the prognostic significance of these appearances is not proven. The cellular composition includes Langerhans cells, macrophages, multinucleated giant histiocytes, T lymphocytes, and eosinophils (Favara, 1997). Histologic involvement may have different appearances in lesions from separate sites. In some instances, lymph nodes are massive and cause airway obstructions.
Suppuration resembling infection has been reported (Sacks, 1986). The bone marrow may be normal or heavily involved. Bone marrow lesions may be focal with pathologic infiltration of Langerhans cells or may contain neutrophils, eosinophils, lymphocytes, multinucleated cells, fibrosis, and (in rare cases) eosinophilic accumulation. Pulmonary involvement is more common in adults than in children (especially adults with a history of smoking), but it occurs in 20-40% of all patients (Ha, 1992; Vassallo, 2000). Small cysts can coalesce and rupture into the pleural cavity, leading to pneumothorax (Donnelly, 2000).
CNS involvement, including pituitary involvement, is often part of systemic disease (Smets, 1997; Yule, 1997; Histiocyte Society, 1997; Barthez, 2000). The CNS is rarely a primary site of LCH involvement (Sivalingam, 1977; Greenwood, 1982; Moscinski, 1985; Penar, 1987; Itoh, 1992; Hund, 1999; Bergman, 1997). The most common site of CNS involvement in persons with LCH is the hypothalamic-pituitary axis, which results in diabetes insipidus in 10-50% of patients (Dunger, 1989). Pathologic changes in the cerebellum, basal ganglia, and pons have been reported (Grois, 2004). Local involvement of the temporal lobe has also been observed and represents a neurodegenerative disorder that is thought to be similar to a paraneoplastic syndrome (Cagli, 2004).
Involvement of the anterior pituitary is relatively uncommon. However, it can result in growth-hormone deficiency or, in rare cases, panhypopituitarism (Kaltsas, 2000). Cerebellar dysfunction with incoordination and white matter changes has been reported (Rosenfield, 1990). LCH may affect the spleen and liver. Primary involvement of the liver is uncommon (Kaplan, 1999; Steiner, 2005). Involvement of the liver is often part of multiorgan disease. Even when PLCs are not present, sclerosing cholangitis can be observed (Kaplan, 1999). Liver infiltration may result in tissue damage and increased enzyme levels, jaundice, coagulation disorders, and hypoalbuminemia.
Involvement of the GI tract is probably more common than is clinically recognized (Nanduri, 1999). Lesions in the stomach, small bowel, colon, and rectum have been reported (Nihei, 1983; Idlibi, 1984; Nanduri, 1999; Egeler, 1990). The usual pathology of GI involvement with LCH includes infiltration of lamina propria and submucosa with glandular, mucosal, and, possibly, villus atrophy. Diarrhea and GI bleeding can be the presenting features of the disease. Involvement of the pancreas is rare (Yu, 1993 and 1994).
LCH rarely involves the intraocular structures. Isolated eye disease has been reported (Sikic, 2005). Lytic lesions of the orbit and resulting soft-tissue extension may cause proptosis. Ptosis and optic atrophy rarely occur. Patients with ear involvement often present with chronic otorrhea, lesions of the external auditory meatus, middle-ear involvement, and mastoid involvement. Although rare, involvement of the genital tract has been reported (Batey, 2002; Montero, 2003; Dietrich, 2004).
In patients with LCH and hematopoietic involvement, Langerhans cell infiltration is often not evident; however, other abnormalities are common. These include abnormal M:E ratio; hyperplasia and dysplasia of megakaryocytes, including mononucleated and bilobed micromegakaryocytes and paratrabecular and grouped megakaryocytes; existence of neutrophil remnants in megakaryocytes (emperipolesis); increased numbers of macrophages; hemophagocytosis; and myelofibrosis (Galluzzo, 2006).
Optimal treatment of Langerhans cell histiocytosis (LCH) has not been established. In ideal cases, the differences between normal cells and PLCs should be used to guide treatment of the disease. However, a lack of sufficient information has hampered specific therapy.
Substantial variation of the disease and the fact that 10-20% of patients achieve spontaneous regression complicate comparisons of current nonspecific therapies (Muscolo, 2003; Yamaguchi, 2004; Rajendram, 2005). Several agents, including drugs for cancer chemotherapy, have been effective in the treatment.
Some suggest that treatment of LCH should be conservative and limited to individuals with constitutional symptoms, such as pain, fever, failure to thrive, and vital organ disorder (Mclelland, 1990). Treatment of these disorders must often be tailored to the patient’s prognostic factors, such as the patient’s age, extent of the disease, sites of involvement, and complications. For example, the general agreement is that simple biopsy and curettage are adequate treatment for a solitary bone lesion (Ladisch, 1994; Arceci, 1998).
See the Medication section for a discussion of agents used in the treatment of LCH.
Multidisciplinary care is essential for all patients. Consultation with an oral surgeon and an otolaryngologist, among others, may be required.
The aim of therapy in histiocytosis is to relieve clinical symptoms and prevent complications of the disease. For single-system disease (eg, of the skin or bone), no therapy or only local therapy may be necessary, although further treatment may be needed in certain circumstances (Broadbent, 1998).
Topical therapy
Localized skin lesions, especially in infants, can spontaneously regress. If treatment is required, topical corticosteroids may be tried. Use of extemporaneously prepared topical 0.02% nitrogen mustard has also been advocated (Mermann, 1955; Sheenan, 1991; Hoeger, 2000). This agent, initially used systemically, appears to provide rapid response within 10 days (Sheenan, 1991) with minimal adverse effects, such as contact allergy. Scarring at the site of the lesion is thought to be due to the disease and not therapy (Hoeger, 2000). In one study, skin lesions promptly healed in 14 of 22 children, and 2 had partial responses (Sheenan, 1991). Low-dose radiation therapy to the local lesions is often effective but is rarely needed. For unresponsive skin lesions, low-dose mild systemic therapy can be used.
Chemotherapy for multisystemic disease with local or constitutional symptoms is used (Arceci, 1999). Single agents or adjuvant use of several chemotherapeutic agents and/or biologic-response modifiers may be effective. Published therapies include corticosteroids, vinca alkaloids, antimetabolites-nucleoside analogs, immune modulators such as cyclosporine (Mahmoud, 1991), antithymocyte globulin (Stephan, 1993), biologic-response modifiers such as IL-2 and INFs (Sato, 1990), cellular treatment, and exchange transfusion (Ladisch, 1982). Most reports of treatment modalities lack controls, with most authors citing the rarity of the disease as justification for the lack (Komp, 1990; D’Angio, 1994).
Single-agent therapy
Purine analogs with activity for treatment of Langerhans cell histiocytosis (LCH) include 2-chlorodeoxyadenosine (2CdA; cladribine [Leustatin]) and 2-deoxycoformycin (2CDF; pentostatin [Nipent]) (Weitzman, 1999; Saven, 1993 and 1994; McCowage, 1996; Dimopoulos, 1997; Stine, 1997); 2CdA has been found to be particularly toxic to monocytes (Carrera, 1990). Justification for the use of 2CdA is that some histiocytes are derived from monocytes (Saven, 1993). In a review of 15 patients with multiorgan involvement receiving 2CdA and 2 receiving 2CDF, 6 had complete responses, 3 had partial responses, 5 had no response, and 1 died early. Fourteen had previously received significant treatments (Weitzman, 1999).
As a single agent, cyclosporine has been used in pretreated patients with advanced LCH. Cyclosporine, a cyclic endecapeptide immunosuppressant of fungal origin, inhibits immune responses. The proposed mechanism of action is blockage of the transmission and synthesis of lymphokines, such as IL-2 and INF (ie, INF-alpha inhibition of the accessory cell function of Langerhans cells and reduced capacity of dendritic cells to enhance mitogenic stimulation of lymphocytes). Cyclosporine is postulated to disrupt abnormal cytokine-dependent activation of lymphocytes and histiocytes in the liver, spleen, lymph nodes, and bone marrow. The activation of lymphocytes is presumed to be secondary to uncontrolled proliferation of Langerhans cells. Furthermore, cyclosporine can inhibit cytokine-mediated cellular activation that potentially contributes to phagocytosis and disease progression (Mahmoud, 1991).
Partial and complete responses have been recorded in a small number of patients. Patients with partial response had achieved a complete response with prednisone and vinblastine chemotherapy. Cyclosporine A has also been used in familial erythrophagocytic lymphohistiocytosis (FEL). In one report of 2 children whose disease was resistant to steroids and etoposide, durable remission was obtained with this agent (Abella, 1997).
INF-alpha had some effect in anecdotal cases of LCH (Jakobson, 1987; Sato, 1990).
Treatment of multifocal relapsing and resistant bone lesions in LCH is challenging. Langerhans cells are capable of releasing cytokines, which are potent activators of osteoclasts and can result in the lytic lesions seen in the disease. Pamidronate, a bisphosphonate agent, has been reported to induce response or result in disease stability in a very small group of patients (Ramanujachar, 2006).
Multiagent therapy
Most chemotherapy agents for the treatment of LCH are used in combination. The length of therapy is arbitrarily chosen. In some studies, patients were stratified by risk factor (Gadner, 1994). Use of a combination of cytarabine arabinoside (Ara-C), vincristine, and prednisolone to treat disseminated LCH with organ dysfunction has been reported.
In a study of 18 pediatric patients with LCH and multiorgan involvement, 8 had additional organ dysfunction; 8 of 10 patients with organ involvement achieved complete remission (Egeler, 1993). Five of 8 patients with additional dysfunction achieved complete remission. Four (22%) of 18 patients developed diabetes insipidus. Two with organ dysfunction died at the time of the report. The regimen was described as being mildly toxic and relatively well tolerated. In this regimen, cytarabine (100 mg/m2/d for 4 consecutive days), vincristine (1.5 mg/m2 on day 1), and prednisone (40 mg/m2/d for 4 wk followed by 20 mg/m2 for 20 d) were administered. The combination of vincristine and cytarabine was repeated every other week for 4 weeks. Thereafter, the interval was extended by 1 week until this combination was administered every 6 weeks, until complete remission was achieved (4-16 wk).
In a multicenter study in 1983-1988, Italian investigators assigned 70 patients with biopsy-proven LCH into good- or poor-prognosis groups, depending on their organ dysfunction (Ceci, 1993). Sixteen patients with limited disease were treated with surgery alone, 5 received immunotherapy with thymus extract then chemotherapy, and 49 patients received chemotherapy with vinblastine (5.5 mg/m2/wk for 3 mo). Poor responders in this group were then treated with doxorubicin (20 mg/m2 intravenously for 2 d every 3 wk for 3 mo). Patients who did not improve with this regimen were administered etoposide (200 mg/m2 intravenously) for 3 consecutive days every 3 weeks for at least 3 months or until their disease progressed. The poor-prognosis group (11 patients) received doxorubicin (20 mg/m2 on days 1 and 2), prednisone (40 mg/m2 by mouth on days 1-29), vincristine (1.5 mg/m2 intravenously once a week for 4 wk starting on day 8), andcyclophosphamide(400mg/m2 on days 15 and 29 for 9 courses).
Only 1 of 10 patients with good prognosis had a favorable response during therapy with thymus extract. Of 54 patients receiving chemotherapy (49 as first-line treatment), 34 achieved complete remission with vinblastine, and 8 had a recurrence after 4-22 months. Of 15 patients achieving remission with etoposide, 1 had a relapse 10 months after therapy. In 11 patients with poor prognoses, 7 had progressive disease, and 6 died within 9 months of diagnosis. Organ dysfunction appeared to significantly affect survival, with only 46% of patients surviving for 12 months. The main complication was diabetes insipidus, which occurred in 20% of patients. The overall incidence of disease-related disabilities was 48%.
In the Austrian and German DAL-HX 83/90 study, patients were stratified into 3 groups: those with multifocal bone disease (group A), those with soft-tissue involvement but without organ dysfunction (group B), and those with organ dysfunction (group C) (Gadner, 1994). Induction therapy consisted of etoposide (60 mg/m2/d for 5 d on days 1-5, followed by weekly dosing of 150 mg/m2), prednisone (40 mg/m2 on days 1-28), and vinblastine (6 mg/m2 starting at week 3 of therapy). Maintenance therapy was risk related and consisted of vinblastine, 6-mercaptopurine, and prednisone in all patients, with etoposide added in group B and methotrexate and etoposide added in group C. Mortality rates for groups A, B, and C were 8%, 9%, and 38%, respectively.
An organized international approach to LCH has been successful (Arceci, 1999). Using the Histiocyte Society’s LCH I protocol (Ladisch, 1994), investigators prospectively and randomly assigned patients with multisystemic LCH who met criteria based on standard diagnostic evaluation (Broadbent, 1989). Patients received vinblastine (6 mg/m2 intravenously weekly for 24 wk) or etoposide (150 mg/m2 intravenously on 3 consecutive days every 3 wk for 24 wk). All patients received methylprednisolone (30 mg/kg intravenously for 3 consecutive days [maximum daily dose of 1 g]). Of the 447 patients who were registered from various countries, 192 had multisystemic disease, and 136 were randomly assigned (72 to the vinblastine arm and 64 to the etoposide arm).
Patients were evaluated at predetermined intervals. Responses at 6 weeks appeared to differentiate responders from nonresponders, who had poor outcomes. Neither the patients’ ages nor the number, type, or dysfunction of the organs differentiated the groups. At 6 weeks, 51 (50%) of 103 patients achieved a complete response or substantial disease regression, whereas 32 (31%) had stable disease or partial or mixed responses. Disease progression was reported in 19 patients. At 26 months, the mortality rate was 18%. Among the patients who died, 4 had an initial response, 5 had intermediate responses, and 9 had initial nonresponses.
The protocol allowed nonresponders to switch to another treatment arm. Only 34% of patients who had switched had favorable results. Disease recurrence was observed in 11 patients who received vinblastine and in 8 who received etoposide. The 2 arms were statistically similar in terms of initial responses, recurrences, and mortality rates. The overall probability of diabetes insipidus was 42%.
The randomized LCH II study of the Histiocyte Society was performed to compare the effects of oral prednisone with vinblastine (with or without etoposide) in patients with multisystemic disease. Patients were divided into low- or high-risk groups. All patients received prednisone (40 mg/m2/d for 28 d with weekly reduction afterward) and vinblastine (6 mg/m2 intravenously weekly for 6 wk). The low-risk group received continuation therapy with vinblastine (6 mg/m2 during weeks 9, 12, 15, 18, 21, and 24), as well as 5-day pulses of prednisone during the same weeks. Patients in the low-risk group were excluded from randomization.
Patients in the high-risk group were randomly assigned to treatment A or B. Treatment consisted of an initial 6 weeks of therapy with prednisolone and weekly vinblastine and continuation therapy, pulses of vinblastine and/or oral prednisone as in the low-risk group, and daily doses of 6-mercaptopurine (50 mg/m2 during weeks 6-24). Treatment B was the same as treatment A, with the addition of etoposide (150 mg/m2 administered on day 1 of weeks 9, 12, 15, 18, 21, and 24). Results of this protocol have not yet been published.
The LCH III of the Histiocyte Society study is designed to determine if methotrexate administered during the initial 2 courses of treatment affects outcomes. In addition, investigators will determine whether maintenance therapy reduces the risk of recurrent disease and improves overall outcomes.
Radiation therapy
Radiation therapy is effective in LCH. Doses ranging from 750-1500 cGy are usually administered, resulting in good local control of single lesions or metastasis, which can occur in critical areas or cause permanent damage. Fractionated doses of radiotherapy have also been used (Jahraus, 2004).
Treatment for recurrent or refractory disease
The severity of the recurrent disease often dictates the type of therapy that is most likely to be helpful. For example, recurrence of an isolated bone lesion can often be treated with nonsteroidal anti-inflammatory drugs (NSAIDs) or intralesional steroid injections. When bone lesions are multiple and cause clinically significant morbidity, systemic therapy can be helpful. In such circumstances, patients often respond to the same drugs that they previously received, such as vinblastine and/or corticosteroids. Extensive recurrence of skin disease, including refractory perianal or vulvar involvement, often requires systemic chemotherapy.
When patients do not have an early (ie, by 6 wk of therapy) response to vinblastine, corticosteroids, methotrexate, 6-mercaptopurine, or even etoposide, alternate therapies should be administered. Although several immunomodulatory agents, such as cyclosporine, have been used in patients with refractory disease, the results have been inconsistent. Cytotoxic chemotherapy often needs to be administered as well.
Several studies, including an international phase II trial, demonstrated notable activity of 2CdA. This agent was originally used to treat patients with refractory hairy-cell leukemia and chronic lymphocytic leukemia. Response rates were more than 50%. 2CdA is both lympholytic and monolytic, making it a potentially ideal drug to use in LCH, which is characterized by reactive lymphocytic and dendritic and macrophage components. Response rates to 2CdA have been particularly good in patients with extensive skin and bone disease, and in some patients with pulmonary involvement. Overall response rates have been about 30-40% in children. In a study with a small number of adults, the response rate was less than 70%. In 2 reports, a combination of 2CdA and Ara-C seemed to have major effects in a small group of children with refractory disease, but clinically significant grade 4 toxicities and a sepsis-related death were reported (Bernard, 2005; Maschan, 2005).
For some patients whose disease does not respond to 2CdA alone, the combination of 2CdA and high-dose cytarabine has been effective. A similar regimen has also been effective in patients with relapses of acute myelogenous leukemia. Until additional information is obtained with this drug combination, the true response rate and the duration of response are difficult to determine.
Other approaches to the treatment of patients with refractory LCH that are being tested or developed and include agents such as thalidomide, which is used to inhibit TNF-alpha and INF-gamma production (Wu, 2005) In some studies, only patients with low-risk disease were likely to respond to thalidomide, whereas high-risk patients with organ involvement were not (Mauro, 2005; McClain, 2005). Further recognition of NF-kappaB pathway may improve the success of targeted therapy for LCH (Brown, 2005).
Targeting humanized antibodies against lineage-specific antigens, such as CD1a antigens on LCH cells, is another treatment being developed. The application of inhibitors of activated cytokine receptors and their downstream signal-transduction pathways is also an important area of future therapeutic trials. Although hematopoietic stem-cell transplantation has been successful in some patients with refractory LCH, identifying patients who might benefit from such high-risk therapy is difficult, and this treatment is associated with significant acute and chronic complications.
Specific therapies, including monoclonal antibodies against the CD1a or CD52 epitopes found on Langerhans cells, are emerging (McClain, 2005).
Local therapy with various agents has been reported. Intralesional infiltration of corticosteroids for treatment of localized LCH has been advocated (Putters, 2005).
Myeloablative therapy followed by bone-marrow or stem-cell transplantation in disease refractory to the conventional therapy has been reported (Akkari, 2003; Steiner, 2005). However, reporting of positive results are likely to bias such reports.
Intravenous immunoglobulin has been used to treat neurodegenerative LCH (Kaplinsky, 2005). However, to the authors’ knowledge, no formal study has been done to conclusively affirm the benefit of such a treatment.
The need to develop effective treatments and, ultimately, strategies to prevention progressive fibrosis of the lung, sclerosing cholangitis, and fibrosis of the liver, and the neurodegenerative pattern of CNS involvement is immense. Additional clinical trials are needed to determine whether agents such as 2CdA or specific inhibitors of fibrosis can improve the outcomes of patients with these complications.