Metabolomics Investigation of Cutaneous T Cell Lymphoma Based on UHPLC-QTOF / MS

Cutaneous T cell lymphoma (CTCL) is the most common primary malignant T cell lymphoma of the skin, presenting as erythema, plaques, tumors or erythroderma. Incidence is 1/10,000 people per year. Although, the actual incidence of the disease may be underestimated due to different diagnostic criteria (Chuang et al., 1990; Weinstock et al., 1999). The median onset age of CTCL has been reported as 57 years (Kim et al., 2003). But a report from China states the median onset age as being 47.5 years before 2000, and 34 years from 2001-2008, showing the onset of CTCL tends to be in the young (Li et al., 2008). Mycosis fungoides is the most frequent type of CTCL, which has an indolent course initiating as erythema, plaques for years and involving lymph nodes or visceral organs as the disease advances. The aggressive type of CTCL has a poor prognosis (Regina et al., 2002). The etiology of CTCL remains poorly understood, and occupational exposure, virus infection and genetic mutation have been proposed as etiological factors. The aberrant expression and function of the transcriptional factors and regulators of signal transduction have been reported in CTCL (Yin et al., 2013). It has been hypothesized that a dysfunctional regulation of small molecules plays a key role in the malignant formation. The occurrence and development of CTCL is a dynamic process regulated by multiple genes, ultimately leading


Introduction
Cutaneous T cell lymphoma (CTCL) is the most common primary malignant T cell lymphoma of the skin, presenting as erythema, plaques, tumors or erythroderma.Incidence is 1/10,000 people per year.Although, the actual incidence of the disease may be underestimated due to different diagnostic criteria (Chuang et al., 1990;Weinstock et al., 1999).The median onset age of CTCL has been reported as 57 years (Kim et al., 2003).But a report from China states the median onset age as being 47.5 years before 2000, and 34 years from 2001-2008, showing the onset of CTCL tends to be in the young (Li et al., 2008).Mycosis fungoides is the most frequent type of CTCL, which has an indolent course initiating as erythema, plaques for years and involving lymph nodes or visceral organs as the disease advances.The aggressive type of CTCL has a poor prognosis (Regina et al., 2002).The etiology of CTCL remains poorly understood, and occupational exposure, virus infection and genetic mutation have been proposed as etiological factors.The aberrant expression and function of the transcriptional factors and regulators of signal transduction have been reported in CTCL (Yin et al., 2013).It has been hypothesized that a dysfunctional regulation of small molecules plays a key role in the malignant formation.The occurrence and development of CTCL is a dynamic process regulated by multiple genes, ultimately leading

Metabolomics Investigation of Cutaneous T Cell Lymphoma Based on UHPLC-QTOF/MS
Qing-Yuan Zhou 1 , Yue-Lin Wang 1 , Xia Li 1 , Xiao-Yan Shen 2 , Ke-Jia Li 2 , Jie Zheng 2 , Yun-Qiu Yu 1 * to tumor metabolic changes.The use of metabolomics to study different blood samples from CTCL patients and healthy volunteers, by searching for small and diagnosis-related molecular biomarkers, may offer the opportunity for early diagnosis of CTCL.In this study, we aim to differentiate the expression of metabolic molecules in CTCL and map the potential biomarkers in CTCL plasma.
LC/MS has been widely applied in the analysis of biological sample according to Kunnathur et al. (Kunnathur et al.,2013).In metabolomics, tandem mass spectrometry was often used to detect as many metabolites as possible.In addition, TOF mass analyzer can provide a more accurate molecular weight of metabolites compared to the others.Ultrahigh performance liquid chromatography (UHPLC)-quadrupole time-of-flight mass spectrometry (QTOF/MS) was employed to profile the plasma of patients with CTCL.Differences in metabolomics data from the two groups were characterized using principal components analysis (PCA).Based on pattern recognition results, we aimed to establish a diagnosis model and explore the potential metabolic biomarkers for early diagnosis and staging in gastric cancer.
Sample pretreatment process is a critical factor for the success of metabolomics research studies.Therefore the major steps in sample pretreatment including plasma extraction method, stability of plasma samples at 25°C and anticoagulation in sample collection were investigated.

Patient recruitment
This prospective study was approved by the Ethics Committee of Ruijin Hospital, Shanghai Jiaotong University.Informed consent was obtained from all participants.CTCL plasma was obtained from 35 patients and matched plasma taken from healthy volunteers.The skin biopsy was examined by histopathological and immunohistochemical stain and T cell receptor gene rearrangement.The diagnosis was made by two experienced dermatopathologists.A computerized tomography (CT) scan and lymph node ultrasound examination were used to determine the tumor stage.(Xia et al., 2013) All patients were evaluated by the same dermatologist.

Sample collection
Venous blood was collected from 30 control subjects and 35 CTCL patients without medication.All the samples were then transferred into heparinized tubes and immediately centrifuged at 12,000 rpm for 10 minutes.Fresh plasma was frozen in liquid nitrogen during operation, then stored at -80°C until processing.To a 150 μL aliquot of plasma samples, 450 μL of methanol was added for protein precipitation.After centrifugation at 12,000 rpm for 10 minutes, an aliquot of 10 μL supernatant was injected for UHPLC/MS analysis.
Mass spectrometric detection was carried out on an Agilent 1260 Infinity series mass spectrometer (Agilent Corporation).The electrospray ionization source was set in positive and negative modes, respectively.Instrument parameters were set as follows: capillary voltage 3.5 kV, gas temperature 350°C and desolvation temperature 350°C.Nitrogen was used as desolvation and cone gas with a flow rate of 10 L/min, respectively.Full scan mode was employed in the mass range of 100-1000 amu.In the MS/MS experiments, argon was used as collision gas and collision energy was set according to the construction of metabolites.Data were collected in centroid mode.

Plasma pretreatment
Plasma sample collection and handling procedures are critical for successful metabolomics research studies (Yin et al., 2013).Firstly, in the sample collection step, commonly used anticoagulants such as ethylenediaminetetraacetic acid (EDTA) and heparin were investigated.In order to avoid plasma matrix interference, globulin samples were added with EDTA and heparin, respectively, to analyze the matrix effect of anticoagulants.The results showed that in either positive or negative mode, an EDTA sample has an impurity peak at 1 min compared with heparin.Heparin was therefore chosen as anticoagulant in our plasma sample collection.Protein precipitation by methanol or acetonitrile was tested respectively.Data processing and identification LC/MSD ChemStation software (Agilent, Shanghai, China) was used for autoacquisition of LC total ion chromatograms (TICs) (Figure 1) and fragmentation patterns.The accurate mass number of every peak was available by XCMS-Online.
Compounds were mainly identified by searching the Human Metabolome Database (www.hmdb.ca)and by matching m/z (accurate mass number).Biomarkers were identified using biochemical databases such as KEGG and METLIN.

Statistical analysis
Qualitative analysis of MassHunter Acquisition Data was used to analyze the experimental data.Results are presented as means ±standard deviation (SD).Data that were not normally distributed were logarithmically transformed to obtain normal distribution before analysis.Continuous variables were analyzed by one-way ANOVA with Tukey's test.P>0.05 was considered to be statistically

Optimization of UHPLC/MS method
Different mobile phase systems were investigated, such as formic acid-methanol and formic acid-acetonitrile.Considering the separation and running time, the mobile phase adopted in our study consisted of solvent 0.1% formic acid and acetonitrile containing 0.05% formic acid with gradient elution program.

Precision and stability
The applied method was validated prior to the analysis of the experimental samples, including the precision of injection, the within-day stability and the repeatability of sample preparation.Extracted ion chromatographic peaks of four ions (m/z 137.0458, 1.356 min; m/z 242.2610, 11.190 min; m/z 496.3480, 14.332 min; m/z 524.3792, 15.502 min) distributed in different spectrum regions and retention time were selected for the method validation.Compared with the standards (retention time and m/z), four compounds were identified among the 36 potential biomarkers: Hypoxanthine, 1-Hexadecanol, LysoPC (18:0/0:0) and LysoPC (16:0/0:0).The relative standard deviations (RSDs) of peak intensities and retention time for the selected ions in pooled plasma sample were calculated.

UHPLC/MS metabolite profiles
The preparation of plasma samples for metabolic profiling analysis by UHPLC/MS involved a protein precipitation step to extract low-molecular weight compounds and remove the large amounts of proteins that would otherwise interfere with the UHPLC/MS analysis.More quantitative information was obtained from the positive ion mode than that collected under the negative ion mode, so that molecular ions (M+H)+ accounted for the majority of the mass spectrum.Figure 1 shows the representative positive base peak intensity.

Analysis of UHPLC/MS data
In order to find the metabolites with a significant change (i.e., potential biomarkers), partial least squaresdiscriminate analysis (PLS-DA) was constructed using the metabolite intensities as variables.As a classic unsupervised method (no prior knowledge concerning groups or tendencies within the data sets was necessary) for pattern recognition, PLS-DA was expected to pick out distinct variables as potential biomarkers through statistical protocols.The score plot for PLS-DA of control and CTCL patients is shown in Figure 2. Two groups were separated with a clear border.The results showed that the method could find potential biomarkers and distinguish CTCL patients from healthy people.

Biomarker identification
In the LC/MS TICs of samples from the study and control groups, the majority of the peaks were identified as endogenous metabolites based on the Human Metabolome Database, including amino acids, organic acids, inorganic acids, carbohydrates, fatty acids, aldehydes, amines, amides, polyols and pyrimidines (Wu et al., 2010).There were about 600 signals obtained.
SIMCA-P was used to find metabolites whose contents were significantly different between CTCL and control samples (Figure 3).According to the analysis results, we chose absolute values with P>0.05 as potential biomarkers (Table 2).

Study of stability and precision
Extracts of plasma samples were stored at 25°C and injected at 0, 4, 8, 12 and 24 hours, respectively.The RSDs of peak intensities and retention time for the selected ions in pooled plasma sample were calculated (Table 3) (Yang et al., 2013).
Precision of injection was carried out by the continuous detection of five injections of the same standard sample which was considered a potential biomarker.RSDs ranged from 0.09-1.6%for retention time and from 3.5-6.0%for peak intensity.The results are shown in Table 3.

Study of matrix effect
Globulin samples were added with EDTA and heparin respectively to analyze the matrix effect of anticoagulants (Figures 4) (Yang et al., 2013).Either in positive or negative mode, EDTA sample has an impurity peak at 1 min compared with heparin.Therefore, we chose heparin as the anticoagulant based on its better pre-treatment results.