Langerhans cell histiocytosis: progress and controversies

Langerhans cell histiocytosis (LCH) is an inflammatory myeloid neoplasm of mixed cellularity, characterized by infiltrating pathological CD1a+/CD207+ dendritic cells (Egeler et al, 2010). Affecting both children and adults, LCH is clinically heterogeneous, ranging from self-resolving skin or single bone lesions to systemic forms involving bone marrow, liver and/or spleen. Disease outcome is highly variable, depending on the degree of involvement – multisystem involving risk-organs (MS-RO+), or not (RO-) or single system disease. Substantial progress has been made in treating LCH, known originally as histiocytosis X, the ‘X’ denoting unknown aetiology. Survival of children with high risk LCH has improved dramatically, to nearly 90%, establishing a standard of care (Gadner et al, 2013). Increased fundamental understanding squarely places LCH at an intersection between immune dysregulation, inflammation, and neoplasia (by discovery of MAP kinase pathway mutations). These substantial advances, together with the ‘mystery of the unresolved’ have incited extensive re-evaluation of all aspects of LCH including pathogenesis, stratification and treatment. The 5-year overall survival (OS) of RO+ patients (risk was defined as and is currently used as risk of mortality) in the latest randomized trial (LCH-III) of the Histiocyte Society (HS) was 84%, substantially higher than in the predecessor LCH-I (62%) and LCH-II (69%) trials (Gadner et al, 2013). Prolonged intense initial treatment, based on early response evaluation, earlier salvage therapy and better supportive care, may have contributed to the increased survival. Controversy over the explanation for the increase in survival may be a factor in scepticism about the value of current treatment approaches. Nevertheless, the highly improved outcomes of LCH-III (Gadner et al, 2013) define a standard of care. We suggest incorporating direct randomized comparison to this standard in trials testing new, experimental, therapeutic approaches. Established prognostic factors in LCH include disease extent at diagnosis, the presence of risk organ dysfunction, and early response to therapy (Lahey, 1981; Ladisch 1982). A requirement of histopathological verification of LCH and clinical stratification were introduced in the 1980s (Chu et al, 1987). Analysis of patients according to LCH risk group revealed marked differences in outcome not discernible without such stratification (Ladisch 1982). Currently, stratification is being further advanced by exploration of the mutation encoding BRAF V600E in LCH as a new molecular diagnostic marker in peripheral blood cell-free DNA (Héritier et al., 2017). More precise longitudinal monitoring of response to treatment and early signs of reactivation may thereby become possible, improving risk stratification and subsequent treatment decisions. Conversely, ignoring LCH risk stratification by both clinical and laboratory findings may lead to erroneous conclusions about the success, or failure, of new treatments. The importance of the systematic randomized LCH trials is underscored by the unexpected discovery of the true significance of reactivations, i.e., that they are usually less serious than the initial disease in extent and severity and are almost never associated with mortality (Gadner et al, 2013). Continued efforts to reduce reactivations remain important, but to reduce the risk of toxicity, the intensity of treatment of LCH reactivations should be tempered by this knowledge about their severity. This is in contrast to leukaemia (to which LCH has been compared), in which relapse usually has a poor prognosis, therefore warranting potentially more toxic treatments. Consideration of diabetes insipidus (DI) as always reflecting active LCH can also lead to overtreatment because DI is a permanent consequence of LCH, not active LCH itself (Abla et al, 2009). Progressive neurodegeneration (LCH-ND) develops in 5% of patients with LCH (McClain et al, 2018). Often severe and debilitating, the diagnosis is made by typical signal changes in the brain stem, basal ganglia, and cerebellum on neuroimaging, together with some neurological symptoms, such as ataxia, dysarthria, dysmetria, cognitive problems and behavioural abnormalities. The pathogenesis of LCH-ND is unknown. Earlier studies detected CD8+ T cells but not CD207+ histiocytes in biopsied lesions, suggesting an autoimmune phenomenon (Grois et al, 2005). More recently, migration of BRAF V600E-positive myeloid cells to particular regions of the brain via perivascular accumulation and parenchymal infiltration has also been found (McClain et al, 2018). Elucidating the pathogenesis and specific aetiology of LCH-ND will be critical to defining appropriate targets and trial design for experimental interventions. Characterization of LCH as a neoplastic disorder is supported by the finding that approximately 60% of cases harbour a somatic mutation that produces the oncogenic BRAF V600E variant (Badalian-Very et al, 2010). Notably, all LCH samples reported to date show evidence of activation of the MAP kinase pathway (Badalian-Very et al, 2010; Chakraborty et al, 2014). Most of the 40% of cases not expressing BRAF V600E have other genetic alterations that activate the pathway. These include other BRAF mutations, structural rearrangements of BRAF, and mutations in other components of the pathway, such as MAP2K1 (Brown et al, 2014; Chakraborty et al, 2014). The pathogenetic role played by these alterations in LCH is confirmed by the striking clinical responses to targeted inhibitors of BRAF or MEK1 (the product of the MAP2K1 gene) seen in LCH cases carrying activated mutations of those targets (Diamond et al, 2018). A small proportion of LCH cases have no detectable genetic abnormalities and the cause of their MAP kinase pathway activation remains to be determined. LCH tissues should always be tested for activating BRAF or MAP2K1 mutations because their presence may guide the choice of second line therapies in patients with resistant or multiply-relapsed disease. A cardinal feature of LCH lesions is a robust inflammatory infiltrate surrounding the pathological histiocytes. The precise signals that attract specific types of inflammatory cells have not been defined, but investigators have documented the expression of many different cytokines and chemokines that could play a role. The effect of these cells on LCH pathogenesis also remains unclear. They may contribute to the pathobiology of disease or only to the bulk of disease in involved organs but not be drivers of disease, perhaps analogous to the situation in Hodgkin lymphoma. Also similar to Hodgkin lymphoma is expression of PD-L1 (also termed CD274) by LCH histiocytes (Gatalica et al, 2015). Whether immune therapy, such as PD-1 (PDCD1)/PDL-1 blockade, has a therapeutic role in LCH, or whether dysregulation of cytokines, such as macrophage colony-stimulating factor may contribute to pathogenesis, remain to be seen. The HS randomized trials have shown that clinical questions in LCH can be successfully addressed through rigorous prospective international collaboration. The rarity of LCH (incidence of 4·6/million/year) is not a contraindication to rigor; witness the successful completion of the randomized HS LCH-I to LCH-III trials. Consequently, circumspection is warranted in designing pilot studies in LCH, particularly avoiding a drift to relaxation of strict diagnostic and stratification criteria and avoiding the urge to replace highly effective standard therapy with something ‘novel’. Combining patients with or without RO+ disease in a non-stratified analysis, a recent small retrospective study of LCH treatment with cytarabine monotherapy reported a 93% 1-year progression-free survival (PFS) of newly diagnosed patients and a 3-year PFS of 41% in relapsed patients (Simko et al, 2015). However, long-term survival in the severe form (risk-organ positive) of LCH alone is already nearly 90% (Gadner et al, 2013). Thus, before considering replacing vinblastine/prednisone therapy as primary therapy based on small retrospective studies with short follow-up and small differences, we recommend conducting randomized trials, considering the unknown effects of such therapy on the risks of reactivations and late sequelae, and resisting temptation to combine different LCH stratification groups (with their potentially very different outcomes) into a single statistical analysis because of too low patient enrolment. The discovery of BRAF and MAP2K1 mutations in LCH has led to targeted therapies acting upon the RAS/RAF/MEK/ERK pathway. Small series and anecdotal case reports of refractory and relapsed LCH have shown responses to the BRAF inhibitors, vemurafenib and dabrafenib (Diamond et al, 2018; Bhatia et al, 2018). Myelotoxicity of BRAF inhibition seems to be less than that of nucleoside analogues (cladribine, clofarabine) or haematopoietic stem cell transplantation. There are caveats, however. In adults with melanoma, BRAF inhibitors have multiple toxicities, including severe cutaneous toxicity, cardiac toxicity, squamous cell carcinoma and, more rarely, secondary pancreatic cancer (Sloot et al, 2014). Lack of paediatric long-term safety data may raise questions about using these drugs in children with LCH. Resistance to BRAF inhibition has been reported commonly in malignancies in adults (Lito et al, 2013) but only very rarely in the histiocytoses. The approach of combining a RAF inhibitor with a MEK inhibitor has merit in adult malignancies and may find a place in LCH. An ongoing international trial is addressing this by testing the combination of dabrafenib and a MEK inhibitor, trametinib, in adults and children with refractory or relapsed LCH (NCI MEK116540; NCT02124772). Importantly, cancers with RAS mutations may actually be stimulated by BRAF inhibition, potentially causing progression; fortunately, such mutations are rare in LCH and can easily be detected (Nelson et al, 2014). Finally, the optimal design of therapy is unknown, given that 75% of adults with histiocytoses recur after stopping BRAF inhibitors (Cohen Aubart et al, 2017), underscoring the risk that it is not curative. Consequently, in children with LCH, a ‘stopping’ study of BRAF or MEK inhibitors, rigorously controlled and carefully noting potential acute and late toxicities (unknown), optimal duration (unknown) and cost (very high) is needed. To develop effective new approaches to LCH, it will be important to consider past advances and issues not yet resolved, as well as new challenges. The recent suggestion that LCH is a neoplastic disorder has steered thought processes to regarding patients with severe LCH as potential candidates for novel targeted therapies. Therapies with RAF and MEK1 inhibitors and clofarabine seem promising, but it is too early to conclude whether they can cure patients with refractory or multiply relapsed high-risk LCH, and it should be noted that the current standard of care is successful for most patients. The treatment of patients with neurodegeneration or with relapsed/refractory risk-organ positive LCH remains challenging. Finally, having the goal of optimizing outcomes, is it possible to simplify complex treatments for LCH to make them accessible globally where financial and other resources might be limited (Narula et al, 2018)? OA and SL developed the concepts; OA, BR, and SL wrote and edited the manuscript.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.