Urine metabolic profiling analysis revealed altered metabolic pathways in anaphylactic animals

Research paper

Urine metabolic profiling analysis revealed altered metabolic pathways in anaphylactic animals

Peng Liu ^1,2,3，Xincong Kang^1,2,3#，Xia Hu^2,3，Shanqing Pan⁴，Junge Liu³，Jian Kang⁵，   Hongqi Xie^1,2,3，Yunlin Wei⁶，Xiuling Ji⁶，Menghui Zhang^7*, Dongbo Liu ^1,2,3*

¹Hunan Provincial Key Laboratory of Crop Germplasm Innovation and Utilization, Hunan Agricultural University, Changsha, Hunan, China;

²State Key Laboratory of Subhealth Intervention Technology, Changsha, Hunan, China;

³Hunan Agricultural University, Changsha, Hunan, China;

⁴Hunan Research Center for Safety Evaluation of Drug, Changsha, Hunan, China;

⁵Department of Dermatology, The Third Xiangya Hospital of Central South University, Changsha, Hunan, China;

⁶Kunming University of Science and Technology, Kunming, China;

⁷State Key Laboratory of Microbial Metabolism, Shanghai Jiao Tong University, Shanghai, China

Running Head: Urine metabolic profiling analysis of anaphylactic animals

Chronic Disease Management and Subhealth Intervention 2014; Vol 1:1-13

* Corresponding authors:

Prof. Dongbo Liu, Hunan Agricultural University, Changsha, Hunan,   China

Tel: 86-731-84635293; Fax: 86-731-84635293; E-mail: chinasaga@163.com

Prof. Menghui Zhang, Shanghai Jiao Tong University, Shanghai, China

Tel: 86-21-34204878; Fax: 86-21-34204878; E-mail: mhzhang@sjtu.edu.cn

^#Liu and Kang contributed equally to the article.

Financial
support:Programs
of
international
science
and
technology
cooperation
(No:
2013DFA31790,
No:
S2013ZR0181).

This is an open-access article distributed under the terms of the International Standard Serial Number (2373-2806) and the International Union for Difficult-to-treat-Diseases (www/iudd.org).

Reproduction is permitted for personal, noncommercial use, provided that the article is in whole, unmodified, and properly cited.

Received: 09/25/2014; Accepted: 10/10/2014; Published: 10/15/2014.

Abstract

Routine diagnostic methods of anaphylaxis are usually based on history and physical exam with poor sensitivity and specificity, which challenge the identification and appropriate management of anaphylaxis. The application of metabolomics holds great potential for anaphylaxis diagnosis due to its high sensitivity and the ability to quantitatively measure the entire composition of metabolites. In this study, urine metabolic profiling was investigated to figure out altered metabolic pathways in anaphylactic animals by using Gas Chromatography-Mass Spectrometry (GC-MS). Principal Component Analysis (PCA) displayed obvious separation between the control group and two model groups (cattle albumin group and ovalbumin group). Two-group Partial Least Squares discriminant analysis (PLS-DA) supplemented with Martens’ uncertainty revealed twenty-two tentatively identified metabolites significantly different between control group and both model groups. The biological significance of these metabolites obtained from Kyoto Encyclopedia of Genes and Genomes (KEGG) revealed abnormal energy consumption of model groups in carbohydrate metabolism, lipid metabolism and amino acid metabolism. Moreover, the immune response modifications of cholesterol and benzene derivates were implied in the model groups. As a conclusion, urine metabolomic profiling would be a potential optimized tool to build metabolic probe library for anaphylaxis diagnosis and management.

Keywords:

Urine metabolic profiling; anaphylaxis; gas chromatography-mass

spectrometry; energy metabolism; immune response

                                         Introduction

Anaphylaxis is “a severe, life-threatening generalized or systemic hypersensitivity reaction” [1]. Its prevalence is as high as 2%, and appears to be rising, particularly in the younger age group [2-4]. Routine diagnostic methods of anaphylaxis are usually based on history and physical exam with little sensitivity and specificity, which challenge the identification and appropriate management of anaphylaxis [5]. Metabolomics provides an overview of the dynamic metabolic status of the organism and yield new biomarkers which can be a clinic tool for diagnosis of disease [6]. Studies regarding asthma [7,8] have established discriminate model subjects from healthy populations and uncovered some potential disease-relevant biomarkers. The application of metabolomics holds great potential for diagnosis of anaphylaxis due to its high sensitivity and the ability to quantitatively measure the entire composition of metabolites in animal models [9].

Small-molecule metabolites play an important role in understanding disease phenotypes in metabolomics investigation [10]. Analysis of small-molecule metabolite profiling in body fluids, involving serum, urine, saliva, et al [11-13], is prevalently applied in diagnosing and monitoring the state of biological organisms [14,15]. Urine, because of its superiorities of less complex, ease of collection, metabolite-rich, higher thermodynamic stability and less intermolecular interaction comparing to other body fluids [16-18], is becoming increasingly popular in metabolomics experiments. Furthermore, metabolic profiling of urine provides a fingerprint of personalized endogenous metabolite markers that correlate to a number of factors such as gender, disease, diet, toxicity, medication and age [19]. Thus monitoring certain metabolite levels in urine has become an important way to detect early stages in disease, such as urological cancer [20], breast cancer [21,22] and especially in kidney disease [23].

However, there is no report of urine metabolomics analysis of anaphylaxis available. Therefore, the aim of this study was to figure out urine metabolic profiling to identify potential biomarkers of anaphylaxis for clinical diagnosis and appropriate monitoring of anaphylaxis.

Materials and Methods

Chemicals

All of the compounds and reagents were analytical grade and purchased from Sigma, except where stated otherwise. Urease (Cout: 5KU; Activity: 208U·mg^-1), N, O-bis (Trimethylsilyl) trifluoroacetamide (BSTFA) and 1 % trimethylchlorosilane (TMCS) were purchased from Tokyo Chemical Industry (Tokyo, Japan). Reference standards (L-arabinitol, cholesterol, glucose, myo-inositol, fumaric acid, octadecanoic acid，L-ascorbic acid), hydrochloride and pyridine (silylation-grade) were obtained from Aladdin Chemistry (Shanghai, China). Acetonitrile (HPLC-grade) was from Sinopharm Chemic Reagent (Shanghai, China). Ultrapure water was produced by a Milli-Q Reagent Water System (Millipore, USA).

Animal sensitization and sampling

All guinea pigs (300-350g) were randomly divided into control group (n = 12,    untreated, nonsensitized), ovalbumin group (n=12, guinea pigs sensitized to 0.1mg·mL⁻¹ intraperitoneal ovalbumin) and cattle albumin group (n=12, guinea pigs sensitized to 100mg·mL⁻¹ intraperitoneal cattle albumin). Animals in the model groups (ovalbumin and cattle albumin groups) were sensitized intraperitoneally with the corresponding allergen on the 1^st, 3^rd, and 5^th days. Then the sensitized guinea pigs were challenged by intravenous injection with ovalbumin (0.1mg·mL−1) or cattle albumin (100mg·mL−1), respectively, on the 14th day [9]. Urine samples were collected on the 21^th day after the first sensitization. Twenty-eight samples were collected with two (one in control and one in ovalbumin group) sampling failed. All samples were stored in aliquots at −80℃ until analysis.

Ethics Statement

This study was carried out in strict accordance with the recommendations in the guide for the Care and Use of Laboratory Animals of Hunan Provincial Institute for Drug Control (HPIDC). The protocol was approved by Laboratory
Animal
Management
Committee
(ethics), the permit number is SYXK (Hunan)-2010-0008, all efforts were made to minimize animal suffering.

Sample pretreatment and derivatization

All samples were thawed at 4℃ and vortex-mixed. Each 100μL aliquot urine sample was mixed with 35μL urease solution (10 mg·mL^-1), kept in a thermostatically controlled water bath (30 ℃) to hydrolyze urea. After adding 265μL icy acetonitrile to deprotein, samples were vortex mixed and centrifugated (12,000rpm, 15min, 4℃). Supernatant (100μL) was transferred into a GC vial and evaporated to dryness with nitrogen by using SE812 pressured gas-blowing concentrators (Shuaien technology Co., Ltd., Beijing, China). Oximation was conducted by adding methoxyamine hydrochloride (20mg·mL^-1, 25μL) dissolved in dry pyridine, vortex mixed for 5 min, incubated at 40℃ for 90min. Silylation were performed by mixed 25μL BSTFA with 1% TMCS and incubated at 70℃ for 60min. The standard solutions were derived following the protocol of samples [9].

GC-MS method

Samples (1μL) were injected into an Agilent QP2010 GC system (Shimadzu, Japan) equipped with a DB5-MS column (30m × 0.25mm × 0.1μm film). Helium was used as carrier gas at a flow rate of 0.8 mL·min^-1. The split ratio is 20:1. The oven temperature increased from 60℃ to 85℃ at a rate of 2.5℃/min and from 85℃ to 280℃ at a rate of 9℃/min and then held at 280℃ for 4 min. The detector operated in the scan mode from 60m/z to 650m/z. The injector temperature, ion-source temperature and interface temperature were 280℃, 230℃ and 250℃, respectively [9].

Date extraction and statistical analysis

Pretreatment of raw data included filtering, peak detection, deconvolution, peak alignment and normalization. The metabolite peaks were identified from their ion ratios and retention times on the base of matching percentage by using National Institute of Standards and Technology (NIST) mass spectrum database or reference standards detected in the same conditions. To further interpret the biological significance of these metabolites associated with anaphylaxis, Kyoto Encyclopedia of Genes and Genomes (KEGG) (http://www.genome.jp/kegg/) database was applied to link them to metabolic pathways.

Principal Component Analysis (PCA), Partial Least Squares discriminant analysis (PLS-DA) and Martens’ uncertainty test were performed using built-in or in-house programs under the MATLAB® environment [24]. PERMANOVA used “adonis” function of package “vegan” in R (http://r-forge.r-project.org/projects/vegan/).

Results

Metabolomic analysis revealed twenty-two variables different between control and model groups

Anaphylaxis models expressed typical clinical symptoms of anaphylaxis（nose rubbing, coughing, wheezing, death et al）and high IgE level in serum after challenge with allergen [9], interpreting successful model establishment. Among 28 samples, a total of 712 aligned individuals peaks (variables) were tentatively identified from GC-MS ion chromatograms of urine based on NIST mass spectra library (Table S 1, not shown). PCA displayed obvious separation between the control group and two model groups, whereas the two model groups were partially overlapped (Fig.1). Application of PERMANOVA [25] with 9999 permutations on the data obtained a p value less than 0.001, indicating the significance of the group separation inside the data ( R package, adonis {vegan}). To identify the important variables contributing to the group separation, two-group PLS-DA supplemented with Martens’ uncertainty was applied [26,27] to the three groups respectively. The former was for discriminant modeling and the latter was for searching important discriminant variables. The result showed that 61 variables (Table S 2; not shown) were significantly different between control group and cattle albumin group while 99 variables (Table S 3; not shown) differed between control group and ovalbumin group. Twenty-two tentatively identified metabolites (Table 1) were found identical in these two variable sets. Among them, four were increased, namely glycerol, phosphoric acid, galactaric acid and 3-hydroxypregn-5-en-20-one hydrazone, while the others decreased.

Figure 1. Loading
plot
of
the
control
group
and
the
model
groups
(A)
from
PCA
(B)

from
PLS-DA (Green circles, control group; blue squares, cattle albumin group; red snowflake, ovalbumin group）

Table 1 Twenty-two metabolites different between control group and both model groups in PLS-DA analysis

No      Metabolite                Cattle-albumin/control    Ovalbumin/control                                     Pathway

表 1. PLS-DA分析获得的22种共同代谢标记物

序号代谢物             牛血清蛋白组/对照组      卵蛋白组/对照组            通路

Thirty-seven diverse variables were detected between cattle albumin group and ovalbumin group

As observed, 116 variables were significantly different between control group and either model group (61 variables between control group and cattle albumin group while 99 variables between control group and ovalbumin group, with 22 shared variables removed). Ninety-five variables differed between cattle albumin group and ovalbumin group by PLS-DA analysis. The 95 variables compared with the 116 variables, thirty-seven variables were found identical (Fig. 2, Table S 4; not shown) [28].

Fig. 2

Venn diagram of diverse variables distributed in three comparisons (http://bioinfogp.cnb.csic.es/tools/venny/index.html).

List 1 blue: different variables between control group and cattle albumin group

List 2 yellow: different variables between control group and ovalbumin group

List 3 green: different variables between cattle albumin group and ovalbumin group

Twenty-two variables were shared by “List1” and “List2”. Thirty-seven diverse variables were detected with 15 variables shared by “List 1” and “List 3” and 22 variables shared by “List 2” and “List 3”.

Tentatively identified metabolites were confirmed with the aid of reference standards

Four out of 22 tentatively identified metabolites and 3 out of 37 tentatively identified metabolites were further identified by comparing the retention time and ion ratios with available reference standards, namely L-arabinitol, glucose, myo-inositol, cholesterol, octadecanoic acid, fumaric acid, L-ascorbic acid (Table 2), consequently confirmed the results based on NIST mass spectra library.

Table 2 Metabolites identified with reference standards by

retention time and ion ratios

Retention time       Ion ratios             Sample/standards

23.30773/147/205/319Glucose

21.19073/103/147/205L-arabinitol

30.54243/121/145/353/368Cholesterol

25.36773/147/191/217/305Myo-inositol

26.86773/117/132/341Octadecanoic acid

23.21773/147/245Fumaric acid

23.02573/117/147/205/274L-ascorbic acid

Function analysis linked anaphylaxis to energy consumption

Information about biological function and involved pathways of the 22 selected metabolites which were obtained from the KEGG database implied to potential biomarkers of anaphylaxis. As shown in Table 1, glucose, 1,3-dyhydroxyl-benzene, myo-inositol, arabinitol, gluconic acid-6-phosphate, 5-hydroxyindole, benzeneacetic acid and cholesterol were decreased while glycerol, phosphoric acid, galactaric acid were increased in model groups. They were mainly related to carbohydrate metabolism, energy metabolism, lipid metabolism and amino acid metabolism with a ratio of 67% (Fig.3).

Fig. 3.    The function classification of 22 tentatively identified metabolites with differences between control group and both model groups in KEGG (red): carbohydrate metabolism; yellow: energy metabolism; green: lipid metabolism; blue: amino acid metabolism; brown: metabolism of cofactors and vitamins; pink: biosynthesis of other secondary metabolites; black: xenobiotics biodegradation and metabolism).

Discussion

In this study, anaphylaxis animal models were discriminated from the control in PCA analysis of GC-MS urinary metabolite spectra (Fig.1A and Fig.1B). The GC–MS method was validated by measuring repeatability (RSD<10 %) of seven tentatively identified metabolites, which were confirmed with the aid of reference standards (Table 2). Twenty-two potential biomarkers    were observed significantly different between control group and both model groups, and the disturbed metabolic pathway (Fig.4) revealed abnormal energy consumption of model groups in carbohydrate metabolism, lipid metabolism and amino acid metabolism.

Figure 4. Schematic diagram illustrating the most predominant disturbed metabolic pathways and the biochemical linkages among the biomarkers.

(Red block arrows) represent down-regulation in challenged animal compared to control group.

(Blue block arrows) represent up-regulation in challenged animal compared to control group.

Anaphylaxis is an acute multisystem syndrome with sudden energy expenditure with carbohydrates as the first source of energy. Five of the 22 identified metabolites, including glucose, L-arabinitol, myo-inositol, galactaric acid, gluconic acid-6-phosphate, were related with carbohydrate metabolism. Among them, glucose, L-arabinitol, myo-inositol and gluconic acid-6-phosphate were used as substrates in glycolysis or tricarboxylic acid cycle and found decreased during anaphylaxis [29]. On the other hand, as product of myo-inositol, galactaric acid was observed at higher level.

When adenosine triphosphate (ATP) produced from carbohydrates is not enough, the organism will burn fat to produce energy. Three anaphylaxis-related metabolites, namely cholesterol, phosphoric acid and glycerol, were involved in lipid metabolism. Cholesterol decreased in urine due to its consumption during anaphylaxis. At the same time, as the catabolites of cholesterol and triglyceride, phosphoric acid and glycerol markedly enhanced. Moreover, in addition to serving as energy source, cholesterol processes immune-response modification in anaphylaxis as well. Mast cell-dependent proteolytic modification of HDL particles would reduce cholesterol efflux from macrophage foam cells ex vivo during anaphylactic shock in mouse [30]. It was reported that acute lowering of cholesterol enhanced mast cell degranulation [31] and cholesterol has interactions with inflammation in asthma [32].

Besides carbohydrate and lipid metabolisms, amino acid metabolism yields carbon skeleton for TCA cycle regarding energy producing. Lower level of benzeneacetic acid in urine of allergic animal might be the result of its consuming in amino acid metabolism. Furthermore, as benzene derivatives, 1, 3-dyhydroxyl-benzene, benzenepropanoic acid, together with benzeneacetic acid might influence biochemical and functional activities of immunecompetent cells and impair immune responses [33]. The inhibitory effects of benzene derivatives on immune responses were supposed to be mediated by interfering early transduction signals including PI3 kinase [34].

In addition to these twenty-two shared variables, thirty-seven diverse variables were observed between cattle albumin group and ovalbumin group, which showed partially overlapped in PCA scatter plot (Fig.1). This diversity may be due to the epitopes that can be recognized by T cell on ovalbumin were different from that on cattle albumin [35,36].

Conclusions

In this paper, metabolomics was first applied to investigate urine metabolic profiling in anaphylactic animals. Twenty-two potential biomarkers of anaphylaxis were observed in this investigation, which could be the metabolic probes of anaphylaxis. Among them, the disturbed metabolic pathway of metabolites (Fig.4) and the immune response modifications of cholesterol and benzene derivates were implied in the model groups. In summary, urine metabolomic profiling would be a potential optimized tool for anaphylaxis diagnosis and management.

Abbreviations

GC-MS: Gas Chromatography-Mass Spectrometry；PCA：Principal Component Analysis; PLS-DA：Partial Least Squares discriminant analysis；KEGG：Kyoto Encyclopedia of Genes and Genomes; BSTFA; O-bis (Trimethylsilyl) trifluoroacetamide; TMCS: trimethylchlorosilane; HPIDC：Hunan Provincial Institute for Drug Control；NIST：National Institute of Standards and Technology; ATP：adenosine triphosphate

Acknowledgements

The authors acknowledge financial support from the programs of international science and technology cooperation (No: 2013DFA31790, No: S2013ZR0181).

Conflict of Interest

The authors have declared that no conflict of interest exists.

References

1. Johansson SG, Bieber T, Dahl R, Friedmann PS, Lanier BQ, Lockey RF, et al. Revised nomenclature for allergy for global use: Report of the Nomenclature Review Committee of the World Allergy Organization,    October 2003.    The Journal of allergy and clinical immunology.

2004; 113: 832-6. doi:10.1016/j.jaci.2003.12.591.

2. Lieberman P, Camargo CA, Jr., Bohlke K, Jick H, Miller RL, Sheikh A, et al. Epidemiology of anaphylaxis: findings of the American College of Allergy, Asthma and Immunology Epidemiology of Anaphylaxis Working Group. Annals of allergy, asthma & immunology : official publication of the American College of Allergy, Asthma, & Immunology. 2006; 97: 596-602. doi:10.1016/S1081-1206(10)61086-1.

3. Lieberman P, Nicklas RA, Oppenheimer J, Kemp SF, Lang DM, Bernstein DI, et al. The diagnosis and management of anaphylaxis practice parameter: 2010 update. The Journal of allergy and clinical immunology. 2010; 126: 477-80 e1-42. doi:10.1016/j.jaci.2010.06.022.

4. Shen Y, Li L, Grant J, Rubio A, Zhao Z, Zhang X, et al. Anaphylactic deaths in Maryland (United States) and Shanghai (China): a review of forensic autopsy cases from 2004 to 2006. Forensic science international. 2009; 186: 1-5. doi:10.1016/j.forsciint.2008.12.007.

5. Bjornsson HM, Graffeo CS. Improving diagnostic accuracy of anaphylaxis in the acute care setting. The western journal of emergency medicine. 2010; 11: 456-61.

6. Watkins SM, German JB. Metabolomics and biochemical profiling in drug discovery and development. Current opinion in molecular therapeutics. 2002; 4: 224-8.

7. Carraro S, Giordano G, Reniero F, Carpi D, Stocchero M, Sterk PJ, et al. Asthma severity in childhood and metabolomic profiling of breath condensate. Allergy. 2013; 68: 110-7. doi:10.1111/all.12063.

8. Ho WE, Xu YJ, Xu F, Cheng C, Peh HY, Tannenbaum SR, et al. Metabolomics reveals altered metabolic pathways in experimental asthma. American journal of respiratory cell and molecular biology. 2013; 48: 204-11. doi:10.1165/rcmb.2012-0246OC.

9. Hu X, Wu GP, Zhang MH, Pan SQ, Wang RR, Ouyang JH, et al. GC-MS-based metabolic profiling reveals metabolic changes in anaphylaxis animal models. Analytical and bioanalytical chemistry. 2012; 404: 887-93. doi:10.1007/s00216-012-6129-x.

10. de la Luz-Hernandez KR, Rojas-del Calvo L, Rabasa-Legon Y, Lage-Castellanos A, Castillo-Vitlloch A, Diaz J, et al. Metabolic and proteomic study of NS0 myeloma cell line following the adaptation to protein-free medium. Journal of proteomics. 2008; 71: 133-47. doi:10.1016/j.jprot.2008.02.003.

11. Ellis DI, Dunn WB, Griffin JL, Allwood JW, Goodacre R. Metabolic fingerprinting as a diagnostic tool. Pharmacogenomics. 2007; 8: 1243-66. doi:10.2217/14622416.8.9.1243.

12. Griffin JL, Mann CJ, Scott J, Shoulders CC, Nicholson JK. Choline containing metabolites during cell transfection: an insight into magnetic resonance spectroscopy detectable changes. FEBS letters. 2001; 509: 263-6.

13. Seger C, Sturm S. Analytical aspects of plant metabolite profiling platforms: current standings and future aims. Journal of proteome research. 2007; 6: 480-97. doi:10.1021/pr0604716.

14. Santini MT, Rainaldi G, Romano R, Ferrante A, Clemente S, Motta A, et al. MG-63 human osteosarcoma cells grown in monolayer and as three-dimensional tumor spheroids present a different metabolic profile: a (1)H NMR study. FEBS letters. 2004; 557: 148-54.

15. Atherton HJ, Jones OA, Malik S, Miska EA, Griffin JL. A comparative metabolomic study of NHR-49 in Caenorhabditis elegans and PPAR-alpha in the mouse. FEBS letters. 2008; 582: 1661-6. doi:10.1016/j.febslet.2008.04.020.

16. Bujak R, Daghir E, Rybka J, Koslinski P, Markuszewski MJ. Metabolomics in urogenital cancer. Bioanalysis. 2011; 3: 913-23. doi:10.4155/bio.11.43.

17. Zhang A, Sun H, Wu X, Wang X. Urine metabolomics. Clinica chimica acta; international journal of clinical chemistry. 2012; 414: 65-9. doi:10.1016/j.cca.2012.08.016.

18. Lamb E, Dawnay A. Effects of sample preparation and storage on urine albumin deter19. Gu H, Pan Z, Xi B, Hainline BE, Shanaiah N, Asiago V, et al. 1H NMR metabolomics study of age profiling in children. NMR in biomedicine. 2009; 22: 826-33. doi:10.1002/nbm.1395.

20. Urquidi V, Rosser CJ, Goodison S. Molecular diagnostic trends in urological cancer: biomarkers for non-invasive diagnosis. Current medicinal chemistry. 2012; 19: 3653-63.

21. Nam H, Chung BC, Kim Y, Lee K, Lee D. Combining tissue transcriptomics and urine metabolomics for breast cancer biomarker identification. Bioinformatics. 2009; 25: 3151-7. doi:10.1093/bioinformatics/btp558.

22. Huang J, Sun J, Chen Y, Song Y, Dong L, Zhan Q, et al. Analysis of multiplex endogenous estrogen metabolites in human urine using ultra-fast liquid chromatography-tandem mass spectrometry: a case study for breast cancer. Analytica chimica acta. 2012; 711: 60-8. doi:10.1016/j.aca.2011.10.058.

23. Abbiss H, Maker GL, Gummer J, Sharman MJ, Phillips JK, Boyce M, et al. Development of a non-targeted metabolomics method to investigate urine in a rat model of polycystic kidney disease. Nephrology. 2012; 17: 104-10. doi:10.1111/j.1440-1797.2011.01532.x.

24. Centner V, Massart DL, de Noord OE, de Jong S, Vandeginste BM, Sterna C. Elimination of uninformative variables for multivariate calibration. Anal Chem. 1996; 68: 3851-8. doi:10.1021/ac960321m.

25. Anderson MJ. A new method for non-parametric multivariate analysis of variance. Austral Ecology 2001; 26: 32-46.

26. N.M. Faber MJM, P. Geladi, M. Sjöström, L.M.C. Buydens and G. Kateman. Random error bias in principal component analysis. Part I. Derivation of theoretical predictions. Anal Chim Acta. 1995; 304: 257-71.

27. N.M. Faber MJM, P. Geladi, M. Sjöström, L.M.C. Buydens and G. Kateman. Random error bias in principal component analysis. Part II, Application of theoretical predictions to multivariate problems. Anal Chim Acta. 1995; 304 273-83.

28. Oliveros JC. An interactive tool for comparing lists with Venn Diagrams.   VENNY; 2007.

29. Sicherer SH, Leung DY. Advances in allergic skin disease, anaphylaxis, and hypersensitivity reactions to foods, drugs, and insects in 2010. J Allergy Clin Immunol. 2011; 127: 326-35. doi:10.1016/j.jaci.2010.11.024.

30. Judstrom I, Jukkola H, Metso J, Jauhiainen M, Kovanen PT, Lee-Rueckert M. Mast cell-dependent proteolytic modification of HDL particles during anaphylactic shock in the mouse reduces their ability to induce cholesterol efflux from macrophage foam cells ex vivo. Atherosclerosis. 2010; 208: 148-54. doi:10.1016/j.atherosclerosis.2009.07.027.

31. Surviladze Z, Draberova L, Kovarova M, Boubelik M, Draber P. Differential sensitivity to acute cholesterol lowering of activation mediated via the high-affinity IgE receptor and Thy-1 glycoprotein. Eur J Immunol. 2001; 31: 1-10. doi: 10.1002/1521-4141(200101)31: 1&#60; 1::AID-IMMU1&#62;3.0.CO;2-W.

32. Yeh YF, Huang SL. Dietary cholesterol enhances pulmonary eosinophilic inflammation in a murine model of asthma. International archives of allergy and immunology. 2001; 125: 329-34. doi:53834.

33. Borrelli I, Loffredo S, Staiano RI, Frattini A, Marone G, Triggiani M. Benzene metabolites inhibit the release of

mination. Annals of clinical biochemistry. 1989; 26 ( Pt 6): 560-2.

proinflammatory mediators and cytokines from human basophils. Int J Immunopathol Pharmacol. 2010; 23: 737-44.

34. Triggiani M, Loffredo S, Granata F, Staiano RI, Marone G. Modulation of mast cell and basophil functions by benzene metabolites. Curr Pharm Des. 2011; 17: 3830-5.

35. Majorek KA, Porebski PJ, Dayal A, Zimmerman MD, Jablonska K, Stewart AJ, et al. Structural and immunologic characterization of bovine, horse, and rabbit serum albumins. Molecular immunology. 2012; 52: 174-82. doi:10.1016/j.molimm.2012.05.011.

36. Nakajima-Adachi H, Koike E, Totsuka M, Hiraide E, Wakatsuki Y, Kiyono H, et al. Two distinct epitopes on the ovalbumin 323-339 peptide differentiating CD4(+)T cells into the Th2 or Th1 phenotype. Bioscience, biotechnology, and biochemistry. 2012; 76: 1979-81.