To cite this article: Niedoszytko M, Oude Elberink JNG, Bruinenberg M, Nedoszytko B, de Monchy JGR, te Meerman GJ, Weersma RK, Mulder AB, Jassem E, van Doormaal JJ. Gene expression profile, pathways, and transcriptional system regulation in indolent systemic mastocytosis. Allergy 2011; 66: 229–237.
Background: Mastocytosis is an uncommon disease resulting from proliferation of abnormal mast cells infiltrating skin, bone marrow, liver, and other tissues. The aim of this study was to find differences in gene expression in peripheral blood cells of patients with indolent systemic mastocytosis compared to healthy controls. The second aim was to define a specific gene expression profile in patients with mastocytosis.
Methods: Twenty-two patients with indolent systemic mastocytosis and 43 healthy controls were studied. Whole genome gene expression analysis was performed on RNA samples isolated from the peripheral blood. For amplification and labelling of the RNA, the Illumina TotalPrep 96 RNA Amplification Kit was used. Human HT-12_V3_expression arrays were processed. Data analysis was performed using GeneSpring, Genecodis, and Transcriptional System Regulators.
Results: Comparison of gene expression between patients and controls revealed a significant difference (P < 0.05 corrected for multiple testing) and the fold change difference >2 in gene expression in 2303 of the 48.794 analysed transcripts. Functional annotation indicated that the main pathways in which the differently expressed genes were involved are ubiquitin-mediated proteolysis, MAPK signalling pathway, pathways in cancer, and Jak-STAT signalling. The expression distributions for both groups did not overlap at all, indicating that many genes are highly differentially expressed in both groups.
Conclusion: We were able to find abnormalities in gene expression in peripheral blood cells of patients with indolent systemic mastocytosis and to construct a gene expression profile which may be useful in clinical practice to predict the presence of mastocytosis and in further research of novel drugs.
transcriptional system regulators
Naïve Bayes prediction
Mastocytosis is an uncommon disease resulting from proliferation of abnormal mast cells in different tissues including skin, bone marrow, liver, spleen, and lymph nodes (1). One of the key elements in the pathogenesis of the disease is the presence of a specific KIT mutation in mastocytes but also in other peripheral blood cells (1, 2). The clinical presentation of mastocytosis is heterogenous, varying from solely skin presentation found in urticaria pigmentosa and mastocytoma, to different forms of systemic disease including indolent systemic mastocytosis, smouldering systemic mastocytosis, aggressive systemic mastocytosis, and mast cell leukaemia (1). Of the adult patients with systemic mastocytosis, the large majority (ca. 90%) have the indolent form of the disease. Most symptoms (like anaphylaxis, hypotension, urticaria, and diarrhoea) are related to mast cell infiltration and degranulation (1). However, other symptoms such as osteoporosis, hypertension, pain syndromes, and neurological symptoms are only partially understood and may involve other mechanisms (1).
The mechanism(s) involved in the development of mastocytosis are mainly unknown (1). Activation of kinase pathways (i.e. D816V mutation of KIT), and IL-13 and rIL-4 polymorphisms have previously been shown to be relevant in this respect (3, 4). It seems quite likely that more than one pathway maybe located in other cells than mast cells alone are involved in the development of the disease. The studies by Garcia-Montero (2) showed that the KIT mutation is present not only in mast cells but also in myeloid and lymphoid cell lineages. Furthermore, the presence of the KIT mutation in various cell lineages was related to the prognosis of mastocytosis (2).
The aim of this study was to find differences in gene expression in peripheral blood cells of patients with indolent systemic mastocytosis compared to healthy controls. The second aim was to define a specific gene expression profile in patients with mastocytosis.
A total of 22 Caucasian patients with indolent systemic mastocytosis from the Department of Allergology, University Medical Center Groningen (UMCG) were studied [median age 50 range 35–73 years; 7 (31%) men and 15 (68%) women]. All patients underwent standard diagnostic procedures based on WHO guidelines for the workup of systemic mastocytosis including bone marrow histopathological, cytological, and flow cytometric (CD2, CD25) examinations of bone marrow. To detect the KIT D816V mutation, we used two different techniques in time. Initially, RNA was isolated from EDTA anti-coagulated bone marrow cells with the help of the QIAamp®RNA Blood MINI Kit (Qiagen, Westburg, Leusden, the Netherlands). The Promega Reverse Transcriptase kit (Promega Benelux, Leiden, the Netherlands) was used to synthesize c-DNA from approximately 500 ng RNA. The resulting c-DNA was amplified using previously described primers with the following PCR conditions: 30 cycles of denaturation (1 min at 95°C), annealing (1 min at 61°C), and extension (2 min at 72°C), followed by 7 min at 72°C and subsequent cooling (5). The resulting 346 -bp PCR product was digested with the help of Hae III and Hinf I (BioLabs, Westburg, Leusden, the Netherlands), resulting in restriction fragments of 171, 127, and 48 (not detected) base pairs to detect the wild type and 157, 127, 48 (not detected), and 14 (not detected) base pairs to detect the Asp 816Val mutation. The restriction fragments were separated on a 6% agarose Multi purpose (Roche, Almere, the Netherlands) gel and visualized using ethidium bromide (patient no. 1, 3, 5, 7, 9, 14, 15, 16, 17, 20, and 22). From December 2007, detection of the KIT D816V mutation was performed with a real-time qPCR using previously published (6) primers 5′-TTGTGATTTTGGTCTAGCCAGACT-3′ and 5′-GTGC-CATCCACTTCACAGGTAG-3′ (patient no. 2, 4, 8, 11, 12, 18, and 19). Urinary histamine metabolites and serum tryptase measurements were also taken (Table 1) (1). A group of 43 healthy Caucasian subjects [median age 50 range 19–73 years, 22 (51%) men and 21 (49%) women] were used as controls. They were nonrelated partners from patients with inflammatory bowel disease visiting the outpatient department of the inflammatory bowel disease unit of the UMCG.
|Patient no.||Gender||Age at diagnosis (years)||UP||Serum tryptase (μg/l)||Urine MH (μmol/mol creat)||Urine MIMA (mmol/mol creat)||MCs in bone marrow aspirate (%)||CD2 immuno-phenotype||CD25 immuno-phenotype||D816V KIT mutation in bone marrow cells||≥2 aggregates of ≥15 MCs in bone marrow||Abnormal morphology of ≥25% of MCs in bone marrow||Histological bone marrow cellularity|
The study was approved by the Medical Ethical Committee of the UMCG (METc 2008/340).
Collection of blood samples
PAXgene blood RNA tubes (Qiagen, Valencia, CA, USA) were used for RNA sampling. All tubes were immediately frozen and stored in −20°C till RNA isolation (maximal period 2 months). RNA was isolated using PAXgene blood RNA Kit CE (Qiagen, Venlo, the Netherlands). All RNA samples were stored in −80°C till labelling and hybridization.
The quality and concentration of RNA were determined using 2100 Bioanalyzer (Agilent, Amstelveen, the Netherlands) and the Agilent RNA 6000 Nano Kit. Samples with RNA integrity number >7.5 were used for further analysis on expression arrays.
For amplification and labelling of the RNA with the Illumina TotalPrep 96 RNA Amplification Kit (Applied Biosystems, Nieuwerkerk ad IJssel, the Netherlands), 200 ng of RNA from each sample was used. The human HT-12_V3_expression arrays (Illumina, San Diego, CA, USA) were processed according to the manufacturer’s protocol. Slides were scanned immediately using Illumina BeadStation iScan (Illumina).
Image and data analysis
First line check, background correction and quantile normalization of the data were carried out with Genomestudio Gene Expression Analysis module v 1.0.6 Statistics (San Diego, CA, USA). Entities of which at least 75% of the samples had a signal intensity value above 20th percentile in 100% of the samples of at least two groups were included for further analysis.
Data analysis was performed using GeneSpring package version 8.0.0 (Agilent Technologies Santa Clara. CA, USA). Genes of which expression was significantly different between the compared groups were chosen based on a log2fold change >2 in gene expression, t-test P-value <0.05 and corrected for multiple testing by the Benjamin–Hochberg method. The naïve Bayes prediction model was used to build a prediction model which might be used in diagnosis of mastocytosis (7). Naïve Bayesian classifier assumes that the impact of single gene expression is unrelated to other genes in the prediction model. The method does not take into account the interactions of the genes composing the model or gene environmental interactions.
Differences in gene expression between patients with mastocytosis and healthy controls were also analysed using transcriptional system regulators (TSR) and factor analysis (FA) described by Fehrmann et al. (10). This method uses principal components derived from the correlation between expressed genes in 15.000 Affymetrix expression arrays.
The gene specific weights were applied to normalized log transformed Illumina transcript data, using the average for every gene, if more than one transcript was available. This procedure was performed for the first 50 principal components identified by Fehrmann et al. (10), resulting in a new set of 50 data points for each person. Subsequently, FA was performed on the component scores, further reducing the set of data points per person to eight explaining 75% of the variance of the 50 original principal component scores. The correlation between the factor scores is caused by the much lower heterogeneity of the data in comparison with the 15.000 arrays used by Fehrmann et al. The aim of this method is to use the correlation structure between genes to find scores that have a higher reproducibility because the signal of many genes is added. The biological interpretation of the factors is derived from those genes that have the strongest contribution to the compound score. Gene related specificity is lost, but problems with overfitting and low reliability of individual gene signals are strongly reduced. Factor analysis was performed with Systat 12.0 (San Jose, CA, USA) and component scores were computed with a computer program written in Delphi 5.0(Austin, TX, USA) available on demand from GTM.
Power calculation to find differences in expression is difficult to compute a priori, as we have no knowledge of the impact of systemic mastocytosis on expression. Considering that the phenotype has a substantial impact on health, we assumed that even with a small number of individuals compared, significant differences between cases and controls would be present even after correction for multiple testing. The analysis based on metagenes is more sensitive than single gene analysis as signals from many genes are combined and errors cancelled out.
Clinical data were analysed with Statistica 8.0 (StatSoft, Tulsa, OK, USA).
Whole genome gene expression analysis was performed on RNA samples isolated from all blood cells in whole blood. From all 48.804 probes present in the array, 48.794 transcripts had sufficient data for further analysis.
Comparison of gene expression profiles between patients and controls revealed a significant difference in 5086 of the analysed transcripts. A fold change difference >2 in gene expression was found in 2330 of those transcripts among which 1951 (84%) were upregulated and 379 (16%) downregulated. Functional annotation indicated that the main pathways in which the differently expressed genes were involved are ubiquitin-mediated proteolysis, MAPK signalling pathway, pathways in cancer, Jak-STAT signalling, and p53 signalling pathway (Table 2). The most important processes influenced by mastocytosis are transcription, cell cycle, protein transport, and signal transduction (Table 3).
|27 genes||129 (37435)||27 (1769)||5.93719e−11||1.82272e−08||(KEGG) 04120 :Ubiquitin-mediated proteolysis|
|35 genes||262 (37435)||35 (1769)||3.3285e−08||5.10924e−06||(KEGG) 04010 :MAPK signalling pathway|
|37 genes||320 (37435)||37 (1769)||5.6426e−07||5.77426e−05||(KEGG) 05200 :Pathways in cancer|
|22 genes||150 (37435)||22 (1769)||2.38024e−06||0.000182683||(KEGG) 04630 :Jak-STAT signalling pathway|
|14 genes||67 (37435)||14 (1769)||2.43652e−06||0.000149602||(KEGG) 04115 :p53 signalling pathway|
|18 genes||108 (37435)||18 (1769)||3.05782e−06||0.000156458||(KEGG) 04110 :Cell cycle|
|15 genes||87 (37435)||15 (1769)||1.3072e−05||0.000573302||(KEGG) 04210 :Apoptosis|
|20 genes||145 (37435)||20 (1769)||1.70886e−05||0.000655773||(KEGG) 00230 :Purine metabolism|
|20 genes||150 (37435)||20 (1769)||2.8221e−05||0.000962649||(KEGG) 04310 :Wnt signalling pathway|
|190 genes||1516 (37 435)||190 (1769)||3.22857e−35||1.51743e−33||GO:0006350 :transcription (BP)|
|61 genes||376 (37 435)||61 (1769)||3.51927e−17||8.27027e−16||GO:0007049 :cell cycle (BP)|
|53 genes||376 (37 435)||53 (1769)||1.34649e−12||2.10951e−11||GO:0015031 :protein transport (BP)|
|134 genes||1700 (37 435)||134 (1769)||4.64627e−09||5.45937e−08||GO:0007165 :signal transduction (BP)|
|51 genes||471 (37 435)||51 (1769)||3.90706e−08||3.67264e−07||GO:0008152 :metabolic process (BP)|
|18 genes||99 (37 435)||18 (1769)||8.26542e−07||6.47458e−06||GO:0006950 :response to stress (BP)|
|47 genes||505 (37 435)||47 (1769)||9.00404e−06||6.04557e−05||GO:0006810 :transport (BP)|
|21 genes||177 (37 435)||21 (1769)||0.00010213||0.000600016||GO:0006464 :protein modification process (BP)|
We matched 13.032 transcripts with Affymetrix genes, using official gene symbol agreement, and used in the factor score analysis. Split-half correlations were computed for each of the 50 factor scores as an indication of independence of factor scores of individual genes.
Among the 50 TSRs described by Fehrmann et al. (10), the TSRs 1, 2, 4, 5, 6, 7, 8, 10, 12, 13, 38, 46, 49, and 50 were most different between patients and controls. In a second FA, two uncorrelated (orthogonal) factors (nr 2 and 4) were identified that both differentiated between cases and controls). The factors 2 and 4 provided the best discriminative properties of predicting the presence of mastocytosis. The main function indicated by the TSRs and KEGG pathway (9) analysis are MAPK signalling pathway, focal and cell adhesion, calcium signalling pathway, neuroactive ligand–receptor interaction, ribosome, cytokine–cytokine receptor interactions, regulation of actin cytoskeleton, and oxidative phosphorylation.
Using leucocyte-specific transcripts described by Liu et al. (11), we analysed the expression profiles of the leucocyte-specific genes characteristic for dendritic cells, B cells, effector memory T cells, mast cells, and basophils. Statistically significant differences in expression between patients and controls were found for the following genes expressed in mast cells: ATP6VOA1, LOC348262, RFESD, OSBPL6, T cells: GALK2, IL32, KLF12, IL12A, SOS1 dendritic cells: C2orf64, CD1B, ZFP3 and B cells: MEF2C, and MS4A1. The genes identified by D’Ambrosio et al. (12) studying gene expression analysis of bone marrow mononuclear cells found similar changes in expression in four of 10 described genes (CPA3, GATA2, KIT, and MAF).
We subsequently went on to build the prediction model which could be used in diagnosing indolent systemic mastocytosis based on the gene expression in the peripheral blood cells. We built the prediction model using a Naïve Bayes classifier based on the most discriminative 29 genes with P < 10−10 corrected for multiple testing (Table 4). The sensitivity of this predictive model was 100% with a specificity of 97%, a negative predictive value of 100%, and a positive predictive value of 96%. Clustering analysis divided those genes into two clusters based on the similarities in gene expression pattern (Fig. 1).
|Gene symbol||Gene name||FC|
|Correlation P||P||Gene function|
|RAB27A||rab27a, member ras oncogene family||0.39||2.58||3e−8||1.3e−11||Neutrophil secretion and shape (26, 27)|
|ETS1||v-ets erythroblastosis virus e26 oncogene homologue 1 (avian)||2.32||0.43||1.2e−8||3e−12||Carcinogenesis (20)|
|ITGB1||Integrin, beta 1 (fibronectin receptor, beta polypeptide, antigen cd29 includes mdf2, msk12)||2.69||0.37||5.2e−8||3.4e−11||Carcinogenesis (18)|
|ARL16||adp-ribosylation factor-like 16||0.38||2.66||1.4e−9||1.7e−11||Signal transduction (8, 9)|
|LRAP||Leucocyte-derived arginine aminopeptidase||0.47||2.11||5.3e−11||3.3e−15||Carcinogenesis (22)|
|MLL3||Myeloid/lymphoid or mixed-lineage leukaemia 3||5.07||0.20||1.9e−8||6.2e−12||Carcinogenesis (28)|
|PLEC1||Plectin 1, intermediate filament binding protein 500 kDa||0.42||2.37||7.8e−8||6.6e−11||Carcinogenesis (17)|
|HSPC268||Hypothetical protein hspc268||0.37||2.69||5.2e−8||3.6e−11||Unknown|
|C3ORF34||Chromosome 3 open reading frame 34||0.41||2.44||3.4e−8||2e−11||Unknown|
|SERTAD2||Serta domain containing 2||2.02||0.49||1.2e−8||3.2e−12||Carcinogenesis (25)|
|ITGAV||Integrin, alpha v (vitronectin receptor, alpha polypeptide, antigen cd51)||4.29||0.23||9.2e−8||7.9e−11||Carcinogenesis (30)|
|SIAH1||Seven in absentia homologue 1 (drosophila)||2.06||0.49||1.2e−8||3.4e−12||Carcinogenesis (21)|
|CCR2||Chemokine (c-c motif) receptor 2||4.53||0.22||7.2e−8||5.6e−11||Carcinogenesis (29)|
|LRRFIP1||Leucine-rich repeat (in flii) interacting protein 1||0.41||2.42||2.9e−8||1.2e−11||Carcinogenesis (23)|
|C9ORF72||Hypothetical protein flj11109||5.23||0.19||5.2e−8||3.5e−11||Unknown|
|ATM||Ataxia telangiectasia mutated (includes complementation groups a, c and d)||2.74||0.37||9.7e−8||8.9e−11||Carcinogenesis (19)|
|PTP4A2||Protein tyrosine phosphatase type iva, member 2||0.47||2.14||2.6e−9||3.7e−13||Carcinogenesis (8, 9)|
|NCOA2||Nuclear receptor coactivator 2||7.36||0.14||6.6e−11||5.4e−15||Carcinogenesis (29)|
|LBR||Lamin B receptor||4.36||0.23||3e−8||1.4e−11||Granulopoiesis maturation of neutrophils (31)|
|MATR3||Matrin 3||4.80||0.21||9.7e−8||8.8e−11||Transcription (33)|
|ACTR2||arp2 actin-related protein 2 homologue (yeast)||2.14||0.47||6.8e−8||5.1e−11||Carcinogenesis (14)|
|PRR5||rho gtpase-activating protein 8||0.50||2.01||7.8e−8||6.6e−11||Carcinogenesis (16)|
|CHRNA5||Cholinergic receptor, nicotinic, alpha 5||0.41||2.42||3.9e−11||1.6e−15||Carcinogenesis (15)|
|ZMAT3||p53 target zinc finger protein||0.46||2.18||1e−11||7.2e−12||Carcinogenesis (13)|
|ZFAND5||Zinc finger, a20 domain containing 2||2.51||0.40||3.6e−8||2.2e−11||Carcinogenesis (8, 9)|
|SGPP1||Sphingosine-1-phosphate phosphatase 1||5.07||0.20||2.2e−8||8.8e−12||Cell proliferation, apoptosis (32)|
|C14ORF153||Chromosome 14 open reading frame 153||0.38||2.65||4.4e−10||4.5e−14||Unknown|
The results of the present study show strong and very significant differences in gene expression in peripheral blood cells between patients with indolent systemic mastocytosis and healthy controls. Additionally, 29 gene expression profile differentiating patients from controls was created based on the most differently expressed transcripts.
We were able to show that gene expression differences are found in other cells than mast cells solely. It confirms the finding that expression effects of the specific KIT mutation of mastocytosis may be found in other cell lineages than mast cells. In addition, it shows a difference in expression of 2330 other transcripts (3).
We also used TSR profiling in the data analysis. The regulation of transcription is a complex mechanism; however, the overlap between diseases and tissues analysed led to the conclusion that a large part of differences in transcription may be explained by a network of co-regulated gene clusters. The studies by Fehrmann et al. (10) showed that the number of orthogonal factors needed to explain most of the variability in expression may be limited to 50 TSRs. Studies made on 17.550 human microarray experiments led to identify 50 TSRs capturing 64% of the variability in gene expression (10). The TSR analysis in our study showed profound transcriptosome abnormalities in patients with mastocytosis. The results of the gene expression analysis made in GeneSpring and by the TSR method indicate similar processes and pathways involved in mastocytosis. The biological complexities of these systems suggest networks of co-regulated genes. The results of the TSR and FA analysis suggest, that the abnormalities in gene expression in indolent systemic mastocytosis are related to biological processes also found in other diseases. This approach, which analyses transcriptional mechanisms common across tissues and diseases, allows analysing of gene expression in whole blood without cell sorting and reduces the probability of finding gene expression patterns related to the experimental conditions and sample studied. The functional analysis reveals processes responsible for neoplastic cell transformation (pathways in cancer, MAPK, Jak-STAT signalling, p53 signalling pathway, cell cycle, and apoptosis). The finding of abnormally expressed genes and pathways may also lead to the application of novel drugs in systemic mastocytosis.
In the next step, we went on to create a gene expression profile which could be of help in diagnosing patients suffering from indolent systemic mastocytosis. We suggested a set of 29 most differently expressed genes divided in two clusters according to the pattern of expression (Fig. 1).
The genes composing cluster 1 were described previously in the pathogenesis of both solid tumours and haematological malignancies. We observed both upregulation of proto-oncogenes and downregulation of tumour suppressor genes. Three of the genes were described in lung cancer, namely ZMAT3 (p53 target zinc finger protein) (13), arp2 actin-related protein 2 (ACTR2) (14), and cholinergic receptor, nicotinic alpha 5 (CHRNA5) (15), and two others in breast cancer, namely rho gtpase-activating protein 8 (PRR5) (16) and plectin 1 (PLEC1) (17). Plectin 1 was also described in ovarian cancer (17) and PRR5 (rho gtpase-activating protein 8) (16) in colorectal cancer. Four other genes were also described in acute myeloid leukaemia, namely integrin beta 1 (ITGB1) (18), ataxia telangiectasia mutated (ATM) (19), and v-ets erythroblastosis virus e26 oncogene homologue 1 (ETS1) (20) and seven in absentia homologue (SIAH1) (21). Leucocyte-derived arginine aminopeptidase (LRAP) plays a role in the development of lymphoma (22). Leucine-rich repeat interacting protein (LRRFIP1) contributes to the pathology of myelodysplastic syndrome (23) and glioblastoma (24). Multiple tumours including lymphomas and solid tumours are related to overexpression of SERTAD2 (serta domain containing) (25). RAB27A gene product is a protein member of the ras oncogene family involved in neutrophil secretion (26) and melanocyte shape (27).
All genes in cluster 2 were upregulated in mastocytosis. Myeloid/lymphoid or mixed-lineage leukaemia (3MLL3) (28), nuclear receptor coactivator 2 (NCOA2), and eosinophilic leukaemia CCR2 [chemokine (c-c motif) receptor 2] (29) also play a role in the pathology of myeloid leukaemia. Integrin alpha v (ITGAV) involvement was described in laryngeal and hypopharyngeal carcinomas (30). For the three genes lamin B receptor (LBR), SGPP1, and MATR3 involvement in cancer was not described, but their function may contribute to carcinogenesis. Lamin B receptor plays a role in the morphological maturation of neutrophils and granulopoiesis (31). Sphingosine-1 phosphate phosphatase (SGPP1) is important in the regulation of cell proliferation, angiogenesis and apoptosis (32). MATR3 plays a role in the regulation of transcription (33).
The analysis of gene expression is becoming a popular diagnostic method in neoplastic and inflammatory diseases (34). Also, gene profiling is a standard procedure in assessing the need for chemotherapy in breast cancer (35). To date, the diagnosis of systemic mastocytosis is based on WHO criteria (3–5). The large differences we observed between cases and controls suggest that a gene expression based test could be developed that would improve the reliability of current diagnostic methods. A potential role for the described gene profile may be in the differential diagnosis of patients with myeloproliferative or myelodysplastic disorders co-existing or masking mastocytosis, or in patients refusing bone marrow examination.
Furthermore, in contrast to the study by D’Ambrosio et al. (21), we analysed RNA isolated from whole blood without prior cell separation. This approach (1) reduces the effect of sample handling and (2) is a simple and standardized method which may be used in the clinical practice in the future, and furthermore (3) peripheral blood sampling is less of a burden to patients in comparison with a bone marrow biopsy. Additionally, RNA isolation and gene expression analysis used in this protocol have become standardized methods which avoid human laboratory errors and may be further adapted to clinical practice in the future.
Some aspects of our study warrant comment. A necessary next step is the validation of the abnormal gene profile in an independent group of patients with indolent systemic mastocytosis to confirm our findings, and in cases with other haematological diseases to see whether such RNA profiles are specific for mastocytosis. Further studies may also answer the questions whether analysis of gene expression profiles may be used in clinical practice to (1) establish the diagnosis of mastocytosis and its clinical variants, (2) assess the risk of anaphylaxis in patients with mastocytosis and (3) assess the risk of progressive disease or to develop a non–mast cell haematological malignancy in these patients. The results of the present study, although limited, may open a new area of research.
In conclusion, we were able to find abnormalities in gene expression in peripheral blood cells of patients with indolent systemic mastocytosis and to construct a specific gene expression profile which may be useful in further research and possibly in clinical practice.
The authors would like to thank Professor Cisca Wijmenga Head of the Department of Genetics UMCG and Professor Dirkje Postma from the Department of Pneumonology UMCG for help in all steps of the scientific work, the nurses from the Department of Allergology for help in collecting the blood samples, Pieter van der Vlies and all colleagues from the Department of Genetics UMCG for help in laboratory work, Agata Somla from the Medical University of Gdansk for help in financial logistics.
The research was supported by the Foundation for Polish Science, and grant of the Polish Ministry of Science and Higher Education, no. N402085934 and N40201031.