Acute-phase ITIH4 levels distinguish multi-system from single-system Langerhans cell histiocytosis via plasma peptidomics

Background Langerhans cell histiocytosis (LCH) is a proliferative disorder in which abnormal Langerhans cell (LC)-like cells (LCH cells) intermingle with inflammatory cells. Whether LCH is reactive or neoplastic remains a controversial matter. We recently described Merkel cell polyomavirus (MCPyV) as a possible causative agent of LCH and proposed interleukin-1 loop model: LCH is a reactive disorder with an underlying oncogenic potential and we now propose to test this theory by looking for acute markers of inflammation. We detected MCPyV-DNA in the peripheral blood cells of patients with high-risk organ-type (LCH-risk organ (RO) (+)) but not those with non–high-risk organ-type LCH (LCH-RO (−)); this difference was significant. LCH-RO (−) is further classified by its involvement of either a single organ system (SS-LCH) or multiple organ systems (MS-LCH). In patients with LCH-RO (−), MCPyV-DNA sequences were present in LCH tissues, and significant differences were observed between LCH tissues and control tissues associated with conditions such as dermatopathic lymphadenopathy and reactive lymphoid hyperplasia. Although MCPyV causes subclinical infection in nearly all people and 22 % of healthy adults will harbor MCPyV in their buffy coats, circulating monocytes could serve as MCPyV reservoirs and cause disseminated skin lesions. Methods Plasma sample from 12 patients with LCH-RO (−) (5 MS-LCH and 7 SS-LCH) and 5 non-LCH patients were analyzed by peptidomics. Mass spectrometry (MS) spectra were acquired and peptides exhibiting quantitative differences between MS-LCH and SS-LCH patients were targeted. Results One new candidate biomarker, m/z 3145 was selected and identified after obtaining a MS/MS fragmentation pattern using liquid chromatography-MS/MS. This peak was identified as a proteolytic fragment derived from inter-alpha-trypsin inhibitor heavy chain 4 (ITIH4, [PDB: Q14624]). Conclusions Peptidomics of LCH have revealed that the level of acute-phase ITIH4 distinguishes MS-LCH-RO (−) from SS-LCH-RO (−). Acute-phase proteins serve non-specific, physiological immune functions within the innate immune system. LCH may be a reactive disorder with both underlying neoplastic potential of antigen presenting cells harboring BRAF mutations and hyper-immunity of other inflammatory cells against MCPyV infection. Among LCH-RO (−), MCPyV-DNA sequences were present in both MS-LCH tissues and SS-LCH tissues without significant differences. ITIH4 may show that LCH activity or LCH subtypes correlates with the systemic or localized reactions of MCPyV infection.


Background
LCH is a proliferative disorder in which abnormal Langerhans cell (LC)-like cells (LCH cells) intermingle with inflammatory cells [1,2]. The appearance of LCH lesions under the microscope, which commonly contain only a small number of abnormal immune LCH cells surrounded by many normal immune cells initially led to questions about LCH as a cancer or immune disorder [3]. For decades, it was thought that the disease is a reactive disorder rather than a neoplastic process [4]. The discovery of the B-Rapidly Accelerated Fibrosarcoma gene (BRAF) [GenBank: G_007873] mutation in 2010 [5] gave new insights into LCH pathogenesis. Clonality [6,7] and BRAF mutation [5] suggest neoplasm [1], whereas granuloma formation [8] with spontaneous regression [1], and hypercytokinemia [1,9] indicate a reactive process. BRAF mutation does not distinguish LCH subclasses, which vary from self-healing to lethal [5]. Recently, there has been much discussion concerning whether or not LCH is considered a cancer. This discussion started when a recent change to patient information about LCH on the National Cancer Institute's website [10] stated, "LCH is a type of cancer that can damage tissue or cause lesions to form in one or more places in the body." On the contrary, the Histiocyte Society Scientific Committee [3] stated "All cancers are considered neoplasias, but not all neoplasias are cancerous." At present time no clear consensus seem to have been reached regarding whether LCH is reactive or neoplastic [1] and whether LCH is cancerous [10] or not [3].
Very recently, we [11] propose a new model for LCH pathogenesis in which the disease is a reactive disorder with underlying neoplastic potential. In other words, LCH is an inflammatory process that is prolonged by BRAF mutation (interleukin (IL)-1 loop model) [11].
The clinical course of LCH varies quite widely depending on the extent of organ involvement [2] and relationship between LCH subtypes and MCPyV and mutations are shown roughly in Table 1. Treatment of LCH should be planned according to the clinical presentation and the extent of organ involvement [2]. In LCH-RO (+), in which circulating precursor LCH cells as a reservoirs for MCPyV may relate to involving high risk organ ((extramedullary) hematopoietic organ) (Table 1), main aims of treatment are to increase survive and to reduce the incidence of late sequelae [2]. In LCH-RO (−), treatment differs widely from a wait-and-see approach for SS-LCH to systemic chemotherapy for MS-LCH [2]. However, the clinical significance of MCPyV infection and BRAF mutation has not been settled because they occur equally as frequently in MS-LCH-RO (−) and SS-LCH-RO (−) ( Table 1) [5].  Localized LCH may resolve spontaneously [2], which might be related to oncogene-induced senescence [45] relayed by interleukin-dependent inflammatory network [46]. We detected MCPyV-DNA in the peripheral blood mononuclear cells (PBMC) of patients with LCH-RO (+) but not those with LCH-RO (−); this difference was significant. In patients with LCH-RO (−), MCPyV-DNA sequences were present in LCH tissues There are no definite specific laboratory markers for LCH. In many cases, LCH presents with nonspecific inflammatory signs that arise from chronic inflammation [2], although comprehensive analysis of serum levels of cytokines/chemokines and growth factors in pediatric patients with LCH was done [19].
As mentioned above, we predict that acute phase markers are related to inflammatory activities of LCH subtypes (Table 1) and demonstrate, via plasma peptidomics, that acute-phase ITIH4 (inter-alpha-trypsin inhibitor heavy chain 4) levels can distinguish MS-LCH-RO (−) from SS-LCH-RO (−) in this paper.
All plasma sample measurements were repeated four times. The resulting 29 MS spectra per each measurement were combined using flexAnalysis version 2.4 to generate an integrated MS spectrum in the molecular mass range of m/z 1000-20,000. The mean relative peak intensities, normalized to a total ion current of m/z between 1000 and 6000, were expressed as arbitrary units (a.u.) [21].

Measurement of monoisotopic masses on BLOTCHIP®
Monoisotopic masses of statistically significant peaks were recorded in reflector mode on an Ultraflex II instrument (Bruker Daltonics).

Purification and MS/MS identification of peptides
Plasma samples were selected for peptide purification as follows: peptides exhibiting quantitative differences between MS-LCH and SS-LCH patients were targeted based on the statistical analysis of MS data. Blood samples containing the highest target peptide concentrations were used by Protosera as sources for further identification studies.
MS/MS spectra of the target peptides were obtained via liquid chromatography (LC)-MS/MS (Q Exactive, Thermo Fisher Scientific K. K., Yokohama, Japan); fragmentation data were applied to a "nonredundant" hu-

Statistical analysis
All statistical MS data analyses were conducted using ClinPro Tools version 2.2 (Bruker Daltonics) [22]. Peak heights exhibiting significant statistical differences between two groups (MS-LCH versus SS-LCH) were selected. MS data intensity comparisons between patients with MS-LCH and SS-LCH were performed using Student's t-test. Differences between values were considered statistically significant at P < 0.05.

Results and discussion
Clinical profiles of patients in this study with LCH and non-LCH are summarized in Table 2. Plasma samples from LCH (n = 12) and non-LCH patients (n = 5) were subjected to BLOTCHIP® followed by MS analysis. Subsequent differential profiling analyses of the two sample sets were performed based on data obtained from the MS analysis. Quantitative differences in peptide peaks between the groups were noticeable primarily in the 2000-6500 m/z range. Statistical analysis indicated that 32 peptide peaks differed significantly between MS-LCH and SS-LCH patients at an average m/z of <6000 (Table 3 and Fig. 1a).
ITIH4 is an acute-phase-related protein [23] and potential new biomarker for distinguishing MS-LCH and SS-LCH (Table 3 and Fig. 1e). Acute-phase proteins serve non-specific, physiological immune functions within the innate immune system [24]. ITIH4 has been detected in animals during experimental bacterial and viral infections [25].
MCPyV causes subclinical infection in nearly all people [26]. Although 22 % of healthy adults will harbor   MCPyV in their buffy coats [27], circulating monocytes could serve as MCPyV reservoirs and cause disseminated skin lesions [28]. Martel-Jantin et al. [29] reported seroprevalence rate of MCPyV antibodies of children 12 months or younger (49/105) in Cameroon and pointed out the presence of maternal antibodies in very young children. Their data indicated that MCPyV infections mostly occurred during early childhood, after the disappearance of specific maternal antibodies [29]. On the contrary Tolstov et al. [30] reported seroprevalence rate of MCPyV antibodies of children of 1 year or younger (0/6) in patients with LCH. We [17] identified a relationship between LCH and MCPyV. MCPyV-DNA in PBMC correlated with LCH-RO (+) [17]. Among patients with LCH-RO (−) (MS-LCH and SS-LCH), MCPyV-DNA was restricted to lesional LCH cells [17], and we predict that primary MCPyV infection may influence the LCH subtype involving cells in an earlyactivated state [4]. Generally, no response is observed after secondary viral infection [25]. Primary respiratory syncytial virus infection at 6 months or younger often induces severe disease [31], although nearly all children are infected by 2-3 years of age [32]. Primary Epstein-Barr virus and cytomegalovirus infections in elderly individuals cause a severe condition called infectious mononucleosis; again, nearly all children are infected with these viruses [33]. Although no response is observed after MCPyV infection [26,27], Kumar et al. [34], however, found that MCPyV-specific T helper cells (in vitro model of a secondary infection) secrete several cytokines, including IL-10. IL-10 is an anti-inflammatory cytokine and is one of cytokines produced in LCH [2,35]. ITIH4 production is up-regulated by IL-6 [23], which is known produced in LCH [2], one of the mediators coordinating the interface between adaptive and innate immunity [36], and might be a target for the therapy [37]. Innate immune function between newborns and elderly is extremely different and large quantities of IL-6 after stimulation of receptors, such as Toll like receptor, by term newborns are indicated [38]. In LCH, MCPyV infection may induce hyperimmunity in both LCH cells [11] and other inflammatory cells [1,2]. Increased mRNA expression of BRAF, which was proven in MCPyV-positive non-small cell lung cancer [39], may also influence BRAF mutated precursor LCH cells.

Conclusions
LCH is a proliferative disorder in which LCH cells intermingle with inflammatory cells [1,2]. Very recently, we [11] proposed a new model for LCH pathogenesis in which the disease is a reactive disorder with both underlying neoplastic potential of LCH cells and other inflammatory cells such as T cells, macrophages, and eosinophils. In other words, LCH is an inflammatory process that is prolonged by mutations of antigen presenting cells (IL-1 loop model). The other inflammatory cells may be informed and instigated by the mutated antigen presenting cells. We tested this theory by examining the abundance of a well-established acute phase marker ITIH4. The presence of the acute phase markers observed is compatible with our proposed theory. Cytokine storm is one of characteristics of LCH and the fact that serum levels of some cytokines but not LCH histology, which is so uniform that pathologists cannot determine whether a given biopsy originates from a patient with LCH-RO (+) or LCH-RO (−) [40], deeply correlate to LCH activity and LCH subtypes [19,[41][42][43][44] explicitly reach this inflammatory reactive model in LCH.
Peptidomics suggests that the systemic or localized reaction against MCPyV may influence LCH activity or subtypes. Primary infection without maternal immunoglobulins against MCPyV [29,30] may influence LCH activity or subtypes. In the aspect of data validation, an orthogonal quantitation method has to be conducted,