The Journal of Allergy and Clinical Immunology
Volume 116, Issue 1 , Pages 56-59, July 2005

Allergy-related genes in microarray: An update review

  • Hirohisa Saito, MD, PhD

      Affiliations

    • Corresponding Author InformationReprint requests: Hirohisa Saito, MD, PhD, Department of Allergy and Immunology, National Research Institute for Child Health and Development, 2-10-1 Okura, Setagaya-ku 157-8535 Tokyo, Japan.
    • ,
  • Jun Abe, MD, PhD
    • ,
  • Kenji Matsumoto, MD, PhD

From the Department of Allergy and Immunology, National Research Institute for Child Health and Development

Received 28 January 2005; received in revised form 6 March 2005; accepted 11 March 2005. published online 24 May 2005.

Tokyo, Japan

Article Outline

Microarrays have attracted tremendous interest among biologists. However, questions have been raised regarding the reproducibility between experiments performed in different laboratories using different platforms. Here, we discuss these problems and reach the following conclusions. First, the reproducibility between different platforms of arrays is low, but bioinformatics may permit compensation at least among oligonucleotide microarrays. Second, it is hard to interpret microarray results generated using mixed cell populations. Hierarchical clustering may be applied to identify whether upregulated transcripts in an inflammatory tissue were caused simply by an increased number of inflammatory cells.

Key words: Microarray, DNA chip, transcriptome, bioinformatics

Abbreviation used: T7-primer, Oligo (dT) primer containing a T7 RNA polymerase promoter

 

Microarrays, also called genome arrays or DNA chips, were lauded by former President Clinton in his State of the Union address on January 27, 1998, and have been expected to be a powerful research tool.1 They have attracted tremendous interest among biologists. However, questions were raised in 2002 regarding the reproducibility of studies performed using microarrays.2 Thus, a document was generated outlining the minimal information that should be reported about a microarray experiment to enable its unambiguous interpretation and reproduction when submitting a microarray study to most journals.3

Microarray technologies can be classified into 2 types. One consists of oligonucleotide microarrays that use synthesized oligonucleotide as probes, whereas the other consists of cDNA microarrays that use whole cDNA molecules with irregular lengths as probes. The microarray guidelines3 aim mainly at reproducing the results obtained bay using cDNA microarrays handmade in individual laboratories. Compared with oligonucleotide microarrays, the results of cDNA microarray assays are less reproducible because both the target RNA and probe sizes are irregular, resulting in variable binding capacity.4 However, questions were raised again regarding the reproducibility between results obtained even when using 3 different oligonucleotide microarray platforms.5 Here, we discuss the pros and cons regarding microarray technology by focusing on studies related to allergy and asthma.

Back to Article Outline

Pros of microarray 

Most successful microarray-based results, especially in allergy-related topics, have been obtained in studies dealing with well-controlled animal model experiments. For example, Karp et al6 obtained 8 new strains of mice by crossing A/J mice, which tend to have antigen-specific airway hypersensitivity, with C3H/HeJ mice. They used GeneChip (Affymetrix, Santa Clara, Calif), an oligonucleotide microarray platform, to examine >7000 genes expressed in the lungs after sensitization and challenge with allergens. They found that a genetic polymorphism of complement C5a, known to induce mast cell degranulation, was associated with the allergen-induced airway hypersensitivity. Zimmerman et al7 created an asthma model by sensitizing Balb/c mice with ovalbumin and Aspergillus fumigatus and used GeneChip to analyze the lungs of these mice and controls after allergen challenge. As a result, they found 291 genes whose expression levels were commonly increased by challenge with both antigens. Among those genes, arginase I, arginase II, and cationic acid transporter 2 were found to be dramatically increased. They suggested the importance of these molecules in the pathology of asthma by showing that they were also highly expressed in human asthmatic airway epithelium. It should be noted that these 2 groups have confirmed their results to be highly reproducible by comparing 2 samples after allergen challenge from each experimental condition.

In a recent issue of the Journal, Wang et al8 and Okumura et al9 used GeneChip to analyze IgE receptor–activated mast cell transcripts in vitro. These 2 groups separately identified amphiregulin as a transcript that is markedly increased following aggregation of FcεRI. Amphiregulin induced tissue remodeling, ie, proliferation of lung fibroblasts and marked induction of MUC5AC transcripts in a human respiratory epithelial cell line. Both groups showed that amphiregulin-positive mast cells are increased in the airways of patients with asthma. These investigators employed culture-derived human mast cells for microarray as the first screening and confirmed the in vivo expression and function of this upregulated gene using human tissue samples. Thus, microarray experiments seem to be highly reproducible when well-controlled animal models or pure cells are used.

Back to Article Outline

Cons of microarray reproducibility between different platforms 

Because cDNA microarray analysis has many difficulties in yielding reproducible results among arrays made in different laboratories, we discuss the subsequent issues by focusing only on oligonucleotide microarray systems such as GeneChip.

In the GeneChip standard protocol, the mRNA contained in more than 5 μg total RNA is reverse-transcribed into cDNA and amplified using an oligo (dT) primer containing a T7 RNA polymerase promoter (T7-primer), followed by transcription into cRNA. In a newly developed small sample protocol, this amplification step using a T7-primer is repeated. Thus, only 50 ng total RNA is sufficient to reproduce the results with this protocol. The correlation between the 2 different protocols using the same sample is not very high (r2 < 0.7). This is because the T7-primer binds to the 3′ flanking region of mRNA and sometimes fails to amplify a sufficient amount of 5′ sequences in the small-sample protocol. Microarray data are usually normalized against the mRNA concentration of certain housekeeping genes, total mRNA quantity, or other empirical baselines considered universal among samples. However, such relative measurements can not be applied beyond different sample protocols or different microarray platforms. To overcome this problem, the Toxicogenomics Projects at National Institute of Health Sciences10 established a bioinformatics system called Percellome that will normalize the gene expression values on a per 1 cell basis. Briefly, they have established a system using internal standards and a compensation program for each transcriptional level. Once normalized, the data from all samples and studies can be expressed as a copy number per genomic DNA level. This system is primarily for use with the Affymetrix GeneChip but can be expanded to other platforms. However, degenerated RNA, found, for example, in human allergic clinical samples rich in eosinophils that contain RNAse-rich granules, has less 5′ sequence complementary RNA and causes irreproducibility even when using the same sample protocol. This is one of the problems facing clinical application of microarrays. Fresh RNA samples are definitely preferred whenever possible.

Back to Article Outline

Use of bioinformatics for interspecies comparison 

Comparison of the gene expression levels between different microarray platforms may be somewhat similar to interspecies comparison of orthologous genes. It is theoretically possible to compare orthologous genes by using a suitable correction formula and by using similar high-quality RNA from both species.

Animal disease models have long been used as surrogates for human diseases and have been informative. Controversy does exist, however, about the relevance of these models for allergic diseases such as asthma. Rodent mast cells are also commonly used experimental tools but are often different from their human counterparts in such things as their responses to certain cytokines and antiallergic drugs.11 Nakajima et al12 have comparatively examined the genome-wide gene expression in human and mouse mast cells. The expression levels of several cytokine and chemokine genes (eg, CCL5, IL-5, CSF1, TNF, SERPINE1, CCL4, IL-3, and CCL1) were markedly upregulated both in human and mouse mast cells after stimulation via FcεRI. They concluded that the regulatory mechanisms of these genes are highly conserved between these species. To facilitate interspecies comparisons, an interspecies comparison database has recently been constructed online (bio.mki.co.jp/en/results/_notes/comparativeDB_index.html), which includes human and mouse mast cell transcriptomes for orthologous genes. Studies on the function of molecules that are highly expressed only in mouse cells must be interpreted carefully with regard to their relevance to potential function in humans. Another implication of the interspecies comparison database for gene expression is that we may be able to correct the information obtained for the mouse transcriptome so that it can be extrapolated to the human transcriptome by using the Percellome bioinformatics system. Mouse disease models are often still required to investigate the pathogenesis of human diseases. Interspecies comparison of functional genomics might then become particularly useful.

Back to Article Outline

Cons of microarray experiments using mixed cell population 

The gene expression levels obtained using GeneChip are usually expressed between 1 and approximately 25,000. However, the expression levels of certain abundantly or scantly expressed genes do not show linear relationships, so that the actual dynamic range is considered to be only 102. In our early GeneChip study using human clinical samples, we identified several increased transcripts in PBMCs obtained from severe atopic patients by calculating the ratio of their expression levels.13 However, the ratio should have been more carefully calculated when the values were small even when accompanied by the presence call, which means an analysis that is statistically significant. Increased transcripts in PBMC included defensin-1 and ribosomal protein L37. However, these genes may have been derived from contamination of the PBMC population by basophils, because our recent results regarding cell type-specific transcriptome data14 found these genes to be highly expressed by basophils. Thus, contamination by a very small population of a different cell type having a certain highly expressed transcript may cause an artifactual presence of the transcript in the whole population even where the major cell type lacks it.

Changes in cell populations are especially crucial in microarray analysis of tissues characterized by recruitment of inflammatory cells. Indeed, arginase I is highly expressed by neutrophils among the human leukocyte types. Upregulation of transcriptional levels in a crude tissue is often caused simply by an increase in recruitment of a certain inflammatory cell type. Therefore, to avoid difficulty in interpreting the results, it is important, for comparison, to purify the target cell population as much as possible. However, mRNA is unstable, meaning that complicated procedures for purification should be avoided. Guajardo et al15 have recently succeeded in identifying nasal epithelial–specific upregulated genes by differentiating the cell types obtained by scraping the nasal membranes of patients with allergic rhinitis. However, computational identification of the cell type specificity may be preferable when a certain important transcript is found in crude tissue samples when cell types can not be identified. For this purpose, a cell type–specific transcriptome database has been developed.14

Back to Article Outline

Cell type–specific gene expression 

A cell type–specific transcriptome database is useful for determining whether an increase of a certain transcript in inflammatory tissues reflects an increase in its expression level or an increase in the number of inflammatory cells that express the transcript at a high level (Fig 1). This database may be also applied efficiently to select safer drug targets.14

  • View full-size image.
  • Fig 1. 

    Hierarchical clustering analysis using public database: gene expression profiles of nasal mucosa tissues with or without inflammation and various inflammatory cells. All data were obtained from our Web site (www.nch.go.jp/imal/GeneChip/public.htm). Hierarchical clustering was applied based on Eisen's Gene Cluster and Tree View (rana.lbl.gov/EisenSoftware.htm) to identify whether 774 upregulated (2-fold) transcripts in the inflammatory tissue (nasal mucosa from allergic rhinitis vs normal control) were caused by the increased number of inflammatory cells (eosinophils, neutrophils, mast cells, CD4 cells, and CD14 cells). The normalized expression index for each transcript sequence (rows) in each sample (columns) is indicated by a color code (red, highest; blue, lowest). Computations were clearly separated into normal tissue–derived transcripts and inflammatory tissue–derived transcripts. Normalization of copy number of transcripts by using the bioinformatics system Percellome in the future should help clarify whether upregulated transcripts are caused by increased expression or simply by the increased number of inflammatory cells.

Mast cells, eosinophils, and basophils are crucially involved in allergic reactions and inflammation. Genes specific for mast cell, basophil, and/or eosinophil could be potential therapeutic targets for allergic diseases because these cells play crucial roles in allergic inflammation. A certain gene that is highly expressed in mast cells, for example, can be easily found by searching the mast cell transcriptome. Comparison with the transcriptomes expressed by other cell types, however, may reveal that the said gene is not truly mast cell–specific. Thus, it is especially important to elucidate information regarding the cell type specificity when developing a new drug. Such information may help to predict drug-related adverse events caused by interaction of a drug with responsible molecules present in important organs. In the future, the safety of candidate drugs could be evaluated by comparing their efficacy on these granulocytes with their toxicity to vital organs.

Back to Article Outline

Future microarray applications 

The increasing prevalence of allergic diseases in developed countries is considered by many to be caused, at least in part, by rapid improvement of hygiene. In human beings, the immune system developed as an ingenious device for defending against frequent attacks by microbes. Therefore, our immune system seems to have become deranged in our recent, unprecedented hygienic environment. It is now necessary to understand the total functional elements making up the immune system, not just a single molecule present in an immunocyte working in our immune system. Here, we have mainly discussed pros and cons of microarrays as a high-throughput assay method. However, it should be stressed that microarrays can detect whole transcripts present in a cell and can be used for understanding system biology.16 It is anticipated that continued advances in this and related technologies will help our understanding of deranged human immunity as a system.

Back to Article Outline

References 

  1. Editorial: getting hip to the chip. Nat Genet. 1998;18:195–197
  2. Editorial: microarray standards at last. Nature. 2002;419:323
  3. Brazma A, Hingamp P, Quackenbush J, Sherlock G, Spellman P, Stoeckert C, et al. Minimum information about a microarray experiment (MIAME)-toward standards for microarray data. Nat Genet. 2001;29:365–371
  4. Southern E, Mir K, Shchepinov M. Molecular interactions on microarrays. Nat Genet. 1999;21:5–9
  5. Marshall E. Getting the noise out of gene arrays. Science. 2004;306:630–631
  6. Karp CL, Grupe A, Schadt E, Ewart SL, Keane-Moore M, Cuomo PJ, et al. Identification of complement factor 5 as a susceptibility locus for experimental allergic asthma. Nat Immunol. 2000;1:221–226
  7. Zimmermann N, King NE, Laporte J, Yang M, Mishra A, Pope SM, et al. Dissection of experimental asthma with DNA microarray analysis identifies arginase in asthma pathogenesis. J Clin Invest. 2003;111:1863–1874
  8. Wang SW, Oh CK, Cho SH, Hu G, Martin R, Demissie-Sanders S, et al. Amphiregulin expression in human mast cells and its effect on the primary human lung fibroblasts. J Allergy Clin Immunol. 2005;115:287–294
  9. Okumura S, Sagara H, Fukuda T, Saito H, Okayama Y. FcεRI-mediated amphiregulin production by human mast cells increases mucin gene expression in epithelial cells. J Allergy Clin Immunol. 2005;115:272–279
  10. Kanno J, Aisaki K, Ono A, Igarashi K. Toxicogenomics using “Percellome” and “Millefeuille” data system. [abstract] Toxicol Appl Pharmacol. 2004;197:265
  11. Okayama Y, Church MK. Comparison of the modulatory effect of ketotifen, sodium cromoglycate, procaterol and salbutamol in human skin, lung and tonsil mast cells. Int Arch Allergy Immunol. 1992;97:216–225
  12. Nakajima T, Inagaki N, Tanaka H, Tanaka A, Yoshikawa M, Tamari M, et al. Marked increase in CC chemokine gene expression in both human and mouse mast cell transcriptomes following Fcε receptor I cross-linking: an interspecies comparison. Blood. 2002;100:3861–3868
  13. Heishi M, Kagaya S, Katsunuma T, Nakajima T, Yuki K, Akasawa A, et al. High-density oligonucleotide array analysis of mRNA transcripts in peripheral blood cells of severe atopic dermatitis patients. Int Arch Allergy Immunol. 2002;129:57–66
  14. Nakajima T, Iikura M, Okayama Y, Matsumoto K, Uchiyama C, Shirakawa T, et al. Identification of granulocyte subtype-selective receptors and ion channels by high-density oligonucleotide probe array. J Allergy Clin Immunol. 2004;113:528–535
  15. Guajardo JR, Schleifer KW, Daines MO, Ruddy RM, Aronow BJ, Wills-Karp M, et al. Altered gene expression profiles in nasal respiratory epithelium reflect stable versus acute childhood asthma. J Allergy Clin Immunol. 2005;115:243–251
  16. Kitano H. Systems biology: a brief overview. Science. 2002;295:1662–1664

 Disclosure of potential conflict of interest: H. Saito received grants/research support in part from the Pharmaceutical and Medical Devices Agency with the Ministry of Health, Labour and Welfare (Millennium Genome Project, MPJ-5) and is employed by the Japanese Ministry of Health, Labour and Welfare only. No commercial association.

PII: S0091-6749(05)00721-9

doi:10.1016/j.jaci.2005.03.048

The Journal of Allergy and Clinical Immunology
Volume 116, Issue 1 , Pages 56-59, July 2005

Article Tools

* Image Usage & Resolution