4766.Full.Pdf

Identification of Novel Th2-Associated Genes in T Memory Responses to Allergens Anthony Bosco, Kathy L. McKenna, Catherine J. Devitt, Martin J. Firth, Peter D. Sly and Patrick G. Holt This information is current as of October 1, 2021. J Immunol 2006; 176:4766-4777; ; doi: 10.4049/jimmunol.176.8.4766 http://www.jimmunol.org/content/176/8/4766 Downloaded from References This article cites 119 articles, 43 of which you can access for free at: http://www.jimmunol.org/content/176/8/4766.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 1, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Identiﬁcation of Novel Th2-Associated Genes in T Memory Responses to Allergens1

Anthony Bosco, Kathy L. McKenna, Catherine J. Devitt, Martin J. Firth, Peter D. Sly, and Patrick G. Holt2

Atopic diseases are associated with hyperexpression of Th2 cytokines by allergen-speciﬁc T memory cells. However, clinical trials with recently developed Th2 inhibitors in atopics have proven disappointing, suggesting underlying complexities in atopy pathogenesis which are not satisfactorily explained via the classical Th1/Th2 paradigm. One likely possibility is that additional Th2- associated genes which are central to disease pathogenesis remain unidentiﬁed. The aim of the present study was to identify such novel Th2-associated genes in recall responses to the inhalant allergen house dust mite. In contrast to earlier human microarray studies in atopy which focused on mitogen-activated T cell lines and clones, we concentrated on PBMC-derived primary T cells stimulated under more physiological conditions of low dose allergen exposure. We screened initially for allergen-induced gene Downloaded from activation by microarray, and validated novel genes in independent panels of subjects by quantitative RT-PCR. Kinetic analysis of allergen responses in PBMC revealed an early wave of novel atopy-associated genes involved in signaling which were coexpressed with IL-4 and IL-4R, followed by a later wave of genes encoding the classical Th2 effector cytokines. We further demonstrate that these novel activation-associated Th2 genes up-regulate in response to another atopy-associated physiological stimulus bacterial superantigen, but remain quiescent in nonphysiological responses in primary T cells or cell lines driven by potent mitogens, which may account for their failure to be detected in earlier microarray studies. The Journal of Immunology, 2006, 176: http://www.jimmunol.org/ 4766–4777.

t is well-established from studies in humans and animal mod- number of differentially expressed genes identiﬁed, however, the els that Th2 lymphocytes secreting signature cytokines such lack of consistency in the gene lists reported by each study is I as IL-4, IL-5, IL-9, and IL-13 are central to the development striking. of atopy (1, 2). Accordingly, these cytokines and their downstream In the current study, a different approach was undertaken. We products have become major foci for drug development for control reasoned that while the pattern of T cell gene expression induced of diseases such as atopic asthma. However, the level of clinical in vivo at sites such as the airway mucosa is ultimately controlled

efficacy achieved in recent trials with newly developed Th2 an- by local tissue microenvironmental factors, significant elements of by guest on October 1, 2021 tagonists including anti-IgE (3), rIL-12 (4), anti-IL-4 (5), and anti- the potential gene response “program” of allergen-specific T mem- IL-5 (6) have been disappointing, suggesting that additional (as yet ory cells (exemplified by the IL-4/IL-5-dominant cytokine profile unrecognized) components of the Th2 cascade which escape reg- of T cell clones from atopics) can be accurately revealed by low- ulation via these approaches play key roles in atopy pathogenesis. intensity in vitro stimulation of recirculating memory cells har- The likelihood that additional atopy-associated genes remain to vested from peripheral blood. In adopting this approach, we have be identified can be inferred from recent experiences reported in minimized in vitro manipulations and avoided the use of strong the immunological literature following the introduction of mi- activation stimuli, which have the potential to distort patterns of croarray technology. Collectively, these studies indicate that indi- gene expression in T cells (22, 23), and report for the first time the vidual immune responses typically involve up-regulation of mul- results of microarray analysis of PBMC-derived T cell responses tiple hundreds of genes (7–9). This technology has also been to the house dust mite (HDM)3 allergen in short-term primary applied in the allergy field to study Th2 responses and related culture. The fidelity of the experimental system was evident by the signaling pathways in freshly isolated T cells (10), polyclonally parallel detection of Th2 cytokine hyperexpression in atopics as stimulated T cells (11, 12), allergen-specific T cell clones and lines pooled samples by microarray, and individually by ELISA. In ad- (13, 14), polarized Th1/Th2 cell lines (15–20), and in animal mod- dition, several novel atopy-associated genes were identified, and els (21). A consistent finding in the studies was the substantial the preferential expression of these genes in atopics was validated in HDM-stimulated CD4ϩ T cells from an additional panel of atopic patients and nonatopic controls, both as pooled samples by Telethon Institute for Child Health Research, and Centre for Child Health Research, Faculty of Medicine and Dentistry, University of Western Australia, Perth, Western microarray and individually by quantitative RT-PCR. We addi- Australia tionally demonstrate differential expression of these novel genes in Received for publication November 1, 2005. Accepted for publication January an unrelated T cell response system, notably activation which is 30, 2006. driven by the bacterial superantigen staphylococcal enterotoxin B The costs of publication of this article were defrayed in part by the payment of page (SEB). charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by the National Health and Medical Research Council of Australia and Pfizer Pharmaceuticals. 3 Abbreviations used in this paper: HDM, house dust mite; SEB, staphylococcal en- 2 Address correspondence and reprint requests to Prof. Patrick G. Holt, Division of terotoxin B; PPD, purified protein derivative; qRT-PCR, quantitative RT-PCR; QT, Cell Biology, Telethon Institute for Child Health Research, P.O. Box 855, West Perth, quality threshold; MB, multivariate empirical Bayes; SR, stimulation ratio; DTH, WA 6872, Australia. E-mail address: [email protected] delayed-type hypersensitivity.

Materials and Methods the novel Th2-associated genes detected by microarray screening, using Subjects more precise qRT-PCR methodology. Microarray data were analyzed in the open-source statistical software R Subjects were volunteers aged 11–58 years. Atopic status to HDM was (͗www.r-project.org/͘), using several additional add-on packages from the determined by skin prick test (wheal Ն5 mm) and/or positive serum HDM- Bioconductor Project (͗www.bioconductor.org/͘) (35) including affy, specific IgE (Ն0.35 kU/L). The study was approved by our institutional affyQCReport, affyPLM, annotate, hgu133a, hgu133plus2, limma, q human ethics committee. value, and time course. The probe-level model algorithm (Refs. 34 and 36, ͗http://stat-www.berkeley.edu/users/bolstad/Dissertation/ Cell preparation and culture methodologies Bolstad_2004_Disseration.pdf͘), which is based on the robust multi- array average algorithm (37), was used for background subtraction, nor- PBMC were cultured as detailed (24) with medium alone or containing ␮ malization, and summarization of probe set intensities. All microarrays optimal stimulatory concentrations of 10 g/ml HDM (Dermatophagoides within each experiment were of comparable quality as assessed using the pteronyssinus; CSL) or 10 ␮g/ml purified protein derivative (PPD; Myco- ␮ affyQCReport package. Microarray scan images were checked for spatial bacterium tuberculosis; CSL), 200 ng/ml SEB (Sigma-Aldrich), 1 g/ml defects/artifacts using diagnostic plots from the affyPLM package. RNA PHA (Murex Biotech), or soluble anti-CD3 and 20 U/ml rIL-2 (Cetus). integrity was confirmed by the Affymetrix RNA degradation controls Optimal concentrations of in vitro stimuli were established in forerunner (GAPDH and actin, 3Ј–5Ј ratios Ͻ3). dose response experiments using panels of PBMC from healthy donors and To identify differentially expressed genes, gene expression intensity was sung lymphoproliferation as a readout. Where specified, CD8ϩ T cells ϩ compared in HDM-stimulated cells with unstimulated cells (each time followed by CD4 T cells were isolated from Ag/allergen-stimulated point was analyzed independently where applicable) using a moderated t PBMC by positive selection using Dynabeads (Dynal Biotech) to Ն92 and Ն test (38) using an empirical Bayes approach which is more powerful than 99% purity, respectively. Selection of dividing cells in HDM-stimulated standard t tests when the number of replicates is low (39). To account for cultures using the dye CFSE was based on methodology in Turcanu et al. multiple testing, the false discovery rate (proportion of false positives (25). Analysis of cytokine protein secretion in parallel 48 h cultures was among the tests called significant) was estimated from p values derived routinely performed as an internal control to confirm the Th2 polarity of from the moderated t test statistics (40) using the method of Storey and Downloaded from atopic responses, before RNA extraction for microarray analyses. Forerun- Tibshirani (41). A false discovery rate of 0.05 (i.e., q value Ͻ0.05) was ner studies (24) within the batch of HDM used in this study demonstrated selected as the significance level for differential expression. A total number that the in vitro stimulatory effects of this allergen on PBMC cytokine of 1482 genes were up-regulated in atopics and nonatopics, and these data production were not influenced significantly by covert LPS contamination. were then analyzed by hierarchical clustering (Pearson correlation, average Affymetrix methodologies linkage; Ref. 42) and quality threshold (QT) clustering (Pearson correlation, maximum cluster diameter 0.5; Ref. 43).

Total RNA was extracted using TRIzol (Invitrogen Life Technologies) To rank differentially expressed genes according to their nonconsistency http://www.jimmunol.org/ followed by RNeasy (Qiagen). RNA integrity was assessed on the Bio- of expression in atopics and nonatopics over time, a multivariate empirical analyser (Agilent Technologies). For gene expression studies of unfrac- Bayes (MB statistic) (Ref. 44, ͗www.stat.berkeley.edu/users/yuchuan/pre- tionated PBMC, cultures from each individual were set-up with the same prints/667.pdf͘) analysis was used. The MB statistic is based on the uni- initial cell number to ensure equivalent amounts of RNA from each subject variate models proposed in Refs. 38 and 45) which use moderated/Bayes- were added to the pool. For gene expression studies of T cell subsets, RNA ian estimates of the variance to overcome the lack of precision due to the samples were quantitated by spectrophotometry and equal amounts of low number of replicates. In addition, the MB statistics also take into ac- RNA from each individual were added to the RNA pool. Pooled RNA count any temporal correlation structure typical of time-course experi- samples (ϳ2 ␮g) were labeled using the one cycle labeling kit (Affymetrix) ments. The MB statistics were plotted against their null distribution (quan- according to manufacturer’s instructions except: 1) 20 U RNase inhibitor tile-quantile plot) to select an appropriate cut-off to allow discrimination of (Geneworks) was added to first-strand cDNA synthesis; 2) reactions were genes which had the same expression pattern/level in atopics and nona- heat denatured (70°C, 10 min) after first-strand cDNA synthesis; 3) sam- topics from those which were different. The null distribution was generated by guest on October 1, 2021 ples were incubated for 10 min with T4 DNA ligase after second-strand in a permutation manner, whereby for each gene at each time point, the cDNA synthesis. Fragmentation of cRNA, hybridization to Affymetrix mi- atopic status labels were randomly permuted among the experimental croarrays, washing, staining, and scanning was performed according to groups. The MB statistics were then recalculated for this permuted data and manufacturer’s instructions. In the experiment in Fig. 5, total RNA was the process was reiterated 100 times. If no genes varied in expression amplified using the small sample labeling protocol, version 2, from Af- between atopics and nonatopics over the time course then a qq plot of the fymetrix before hybridization. real data and this randomly permuted data would fall upon a straight line whereas deviations from such a line identify genes varying in expression Quantitative (q)RT-PCR between atopics and nonatopics over time. Volcano plots (46) (used in Fig. 3) were derived as follows. Stimu- RNA was reverse-transcribed with Omniscript (Qiagen) according to man- lation ratios (SR; the expression level in HDM-stimulated cells divided ufacturer’s instructions. Primer sequences were obtained from a database by that in unstimulated cells) were calculated for atopics and nonatop- (͗http://pga.mgh.harvard.edu/primerbank/͘; Ref. 26), designed in-house us- ics. SR in atopics and nonatopics were then compared using the mod- ing Primer Express software (Applied Biosystems), or purchased from Qia- erated t test, and the t test statistics were plotted on the vertical axis. gen. Reverse-transcribed RNA samples were quantitated using QuantiTect The ratio of the SR in atopics and nonatopics was calculated and plotted SYBR Green (Qiagen) on the ABI Prism 7900HT (Applied Biosystems). on the horizontal axis. Relative standard curves were prepared from serially diluted RT-PCR Microarray plots (used in Fig. 5) were derived as follows. SR were products or plasmid standards and data were normalized to the stably ex- calculated and plotted on the vertical axis. The average signal (average of pressed gene EEF1A1 (27). Data were expressed as expression level above the expression intensity in unstimulated and stimulated cells) was plotted background (level in stimulated cells minus unstimulated cells) and mul- on the horizontal axis. tiplied by a scaling factor to obtain whole numbers. Statistical analyses by Mann-Whitney U test were performed in SPSS Study design and statistical analyses software. The paired unequal variance t test, moderated t test, MB statistics, and hierarchical clustering were performed in R software. QT clustering Extensive physiological variation in gene expression has been reported in was performed in MeV software (47). multiple species (28–31), making it difficult to discriminate between do- nor-specific changes in gene expression and changes associated with the phenotype of interest (32). Therefore, a large number of individuals were Results included in the study and a pooling strategy was adopted for the initial Kinetic analysis of inhalant allergen-driven gene expression in microarray screens. The rationale for this approach is the fact that vari- PBMC: cluster analysis reveals early vs late Th2-associated ability due to biological variation is much greater than that due to technical genes variation (33, 34); this was accordingly emphasized in the study design by creating multiple independent RNA pools and hybridizing them once rather Recirculating T memory cells are present in peripheral blood at than performing multiple hybridizations of a single pool. Hybridization of frequencies usually below 1 in 1000 PBMC (48). Accordingly, multiple independent RNA pools minimizes the chance of an outlying individual biasing the results, and more importantly separate RNA samples published studies on microarray analysis of gene expression in were reserved from each culture to evaluate subject-to-subject variation in allergen-specific T memory cells have used a variety of tissue cul- gene expression patterns and to determine the consistency of expression of ture techniques to enrich for the cells of interest, before extraction 4768 GENE EXPRESSION IN INHALANT ALLERGY of RNA. Prolonged in vitro manipulation is known to alter gene the RNA samples were combined into six different pools (three expression profiles of T cells, and in this study we sought to min- atopic: AT1 and AT2 from study group 1, AT3 from study group imize such potential artifact by studying gene expression in freshly 2; and their nonatopic counterparts NA1 and NA2 from study activated memory cells in short-term primary cultures of PBMC. group 1 and NA3 from study group 2) as detailed in Fig. 1, and The results of differential gene expression studies on HDM- hybridized to Affymetrix series U133 plus 2.0 microarrays. The stimulated unfractionated PBMC from two groups of subjects, half pooling strategy was adopted to reduce the variance (50, 51) due of each group being nonatopic controls and half atopic and skin to normal physiological subject-to-subject variation in gene ex- prick test-positive to HDM, are illustrated in Fig. 1. PBMC from pression (see Materials and Methods); separate individual RNA all these individuals (total n ϭ 40) were cultured in the presence or samples were reserved for follow-up studies by qRT-PCR (see absence of optimal stimulatory levels of HDM for 12, 24, and 48 h. below). These time points were selected because they bracket the period To identify genes triggered in response to HDM exposure, gene during which expression of T cell activation markers is maximal expression intensity in allergen-stimulated cells was compared following in vitro stimulation (49). Total RNA was extracted from with that in unstimulated cells at each time point using a moder- PBMC at the termination of these cultures. At each time point for ated t test (see Materials and Methods). After accounting for mul- each treatment (unstimulated or HDM-stimulated), aliquots from tiple testing, 926 genes were statistically significantly up-regulated Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 1. Identification of novel Th2-associated clusters in HDM-stimulated PBMC. PBMC from 20 atopics and 20 nonatopics were cultured in the presence or absence of HDM for the indicated time points (12, 24, or 48 h). Total RNA was extracted and pooled into six separate groups: AT1 (pool of 5 atopics), AT2 (pool of 5 atopics), AT3 (pool of 10 atopics), NA1 (pool of 5 nonatopics), NA2 (pool of 5 nonatopics), NA3 (pool of 10 nonatopics). Pooled RNA samples were labeled and hybridized to Affymetrix Plus 2.0 microarrays. Differentially expressed genes were identified by comparing gene expression intensities in HDM-stimulated vs unstimulated cells at each time point using a moderated t test. At a false discovery rate of 0.05, 926 genes were up-regulated in atopics, and 1344 genes were up-regulated in nonatopics (a total of 1482 genes were induced by HDM stimulation). Genes were then divided into groups with similar expression patterns over the time course using the QT clustering algorithm. The clustering algorithm identified 29 clusters, 12 major clusters (i.e., contained Ն16 genes), and 2 of the 12 clusters were Th2 associated. The expression patterns of the Th2-associated clusters and the 10 other major clusters are illustrated in A and B, respectively. The former comprised an early Th2 cluster including IL-4 and IL-4R which peaked at 12–24 h, and a late Th2 cluster peaking at 48 h which included the Th2 effectors IL-5, IL-9, IL-13, and CCL17. The line plots depict the average (across replicates AT1-AT3, NA1-NA3) normalized (subtracted by the mean and divided by the SD) expression level of all the genes in the cluster over the time course. A subset of the most highly expressed novel Th2-associated genes from each cluster were selected for further study, and the heat map displays the gene expression intensity for respective replicates at each time point. Data are expressed as fold change (HDM-stimulated cells/unstimulated cells). The Journal of Immunology 4769 at the q Ͻ 0.05 level in atopics and 1344 in nonatopics, of which not shown). Previous microarray studies in model systems such as 788 were common to both groups (Venn diagram in Fig. 1A). A Saccharomyces cerevisiae and Caenorhabditis elegans have indi- comparable number of genes were down-regulated in the groups cated that clusters of genes are generally involved in the same but these were not considered further. The nature of the response biological function/pathway (Ref. 52; reviewed in Ref. 53). It was within the nonatopics is discussed in more detail below. With re- therefore not surprising that a distinct functional bias was also spect to the atopic responses, the fidelity and sensitivity of this observed within the Th2-associated clusters. The early Th2 cluster short-term culture model was demonstrated by the consistent de- was enriched for genes known to play a role in signal transduction, tection of hyperexpression of the classical Th2 index genes IL-4, whereas the late Th2 gene cluster was enriched for genes mainly IL-4R, IL-5, IL-9, and IL-13 in the atopic cultures. This finding was involved in immune effector function (functions detailed in Table I). consistent with the presence of relatively high levels of Th2 cyto- The hypothesis-driven data analysis strategy based on clustering kine protein (IL-5 and IL-13) in 48 h cell culture supernatants of could potentially miss important discriminators of HDM responses atopics but not nonatopics (data not shown). in atopics and nonatopics, therefore an alternative analysis was Hyperexpression of Th2 cytokines has consistently been asso- sought to investigate this possibility. The MB statistic derived by ciated with asthma and allergy in vitro and in vivo, and genes Tai and Speed (44) was chosen for this analysis because it ranks which are coexpressed with these Th2 cytokine genes over the genes according to the nonconsistency of their expression across time course are thus potential candidates for an effector role in time and biological conditions taking into account any temporal disease pathogenesis. To identify such coexpressed genes, we used correlation structure in the data (see Materials and Methods). QT cluster analysis, which divides the data into groups of genes Ranking the top 1482 HDM-induced genes in the expression pro- with similar expression patterns (see Materials and Methods). The files of atopics and nonatopics (Venn diagram in Fig. 1A)bythe Downloaded from algorithm identified 29 clusters, 12 of which contained Ն16 genes. MB The expression pattern of these 12 major clusters in atopics and statistic revealed that the top ranking gene in the genome dis- nonatopics is illustrated in Fig. 1. The QT algorithm identified two criminating atopic and nonatopic responses to HDM was IL-9 from separate Th2-associated gene clusters demonstrating peak expres- the late Th2 cluster, followed by IL-4R and the novel genes sion early vs late in the response. The early cluster comprised a set DACT1, GNG8, MAL, and NDFIP2 from the early Th2 cluster of genes peaking in expression at 12–24 h in conjunction with the (data not shown). It is noteworthy that expression of the Th1 principal Th2 growth factor IL-4 and its receptor IL-4R, together marker genes IFNG, LTA, GZMB, STAT1, and IRF1 (15) and ad- http://www.jimmunol.org/ with a series of novel genes including DACT1, MAL, NDFIP2, ditional genes downstream of IFNG including TAP1, TAP2, GBP2, GNG8, RAB27B, DPP4, and NSMCE1 which were selected for FAS, CXCL9, ICAM1, SOCS1, and INDO (54) was a feature com- further study (early Th2 cluster in Fig. 1A). Expression of the other mon to both atopics and nonatopics; these genes were all expressed principal Th2 effector cytokines IL-5, IL-9, and IL-13 peaked later late in respective responses (cluster 1 in Fig. 1B), and expression at 48 h and clustered with additional genes of interest notably levels within the two groups were not statistically significantly IL-17RB, CISH, and CCL17 (TARC) (late Th2 cluster in Fig. 1A; different. Additionally, IFNG protein levels in cell culture super- gene nomenclature detailed in Table I). It is noteworthy that these natants after 48 h of HDM stimulation did not differ significantly same clusters were also identified by hierarchical clustering (data between atopics and nonatopics (data not shown). by guest on October 1, 2021

Table I. Functional annotation of atopy-associated genes described in Resultsa

Symbol Name/alternative names Function

DACT1 Dapper, antagonist of ␤-catenin, homolog 1; DPR1; Inhibits disheveled activation of JNK and ␤-catenin FRODO; HDPR1 signaling pathways MAL Mal, T cell differentiation protein; T lymphocyte Formation and stabilization of lipid raft domains; potential maturation-associated protein; myelin and role in vesicular protein trafficking lymphocyte protein NDFIP2 Nedd4 family interacting protein 2; N4WBP5A Promotes NF-␬B signaling; interacts with Nedd4; localizes to multivesicular bodies and may play a role in protein trafficking RAB27B RAB27B, member RAS oncogene family Similar functions to RAB27A, role in exocytosis of lytic granules containing GZMB DPP4 Dipeptidylpeptidase 4; CD26 Ag; adenosine deaminase- Costimulatory role in T cell activation; regulates the complexing protein 2 biological activity of several chemokines. GNG8 Guanine nucleotide-binding protein (G protein), ␥ 8 Modulator/transducer in various transmembrane-signaling systems; contains GTPase activity NSMCE1 Non-SMC element 1 homolog; NSE1 Ubiquitin-protein ligase activity CISH Cytokine inducible SH2-containing protein; CIS; CIS-1; Negative regulation of STAT5 cytokine-signaling pathway G18; SOCS IL17RB IL-17R B; CRL4; EVI27; IL-17BR; IL-17RH1 Receptor for the proinflammatory cytokines IL-17B and IL- 17E (IL-25) PLXDC1 Plexin domain-containing 1; TEM3; TEM7 Interacts with cortactin; contains a domain found in plexins, semaphorins, and integrins PECAM1 Platelet/endothelial cell adhesion molecule; CD31 Ag Cell adhesion molecule which regulates calcium signaling EndoCAM and inhibits T cell activation GZMB Granzyme B; CCPI; CGL-1; CGL1; CSP-B; CSPB; Mediates apoptosis of target cells CTLA1; CTSGL1; HLP; SECT CAMK2D Calcium/calmodulin-dependent protein kinase II ␦ Activated by calcium/calmodulin signaling pathway and regulates activity of key transcription factors SYTL3 Synaptotagmin-like 3; SLP3; exophilin-6. Binds to RAB27A; role in exocytosis and protein transport

a Annotation obtained from Genecards ([www.genecards.org/index.shtm]) and the literature. 4770 GENE EXPRESSION IN INHALANT ALLERGY

Expression of the “novel” Th2-associated genes in these clusters did not contain probe sets for NDFIP2, GNG8, NSMCE1,or has not been reported previously in the context of specific allergen- RAB27B. Expression of the Th2 index cytokines IL-4, IL-5, IL-9, triggered T cell memory responses in atopics. Some have been and IL-13 was clearly evident in both T cell subsets in the atopics observed previously in experiments broadly related to Th2 immu- (Fig. 2), clustering into early and late groups as in Fig. 1A. Ex- nity, notably 1) CISH was induced by IL-4 in mouse T cells (21) pression of these genes appeared more consistent and of higher and was also detected in human Th2 cells (55); 2) IL-17RB was intensity within the CD4ϩ compartment (see also study group 3 reported to be anti-CD3/CD28 up-regulated in restimulated Rye- below). The results in Fig. 2 also demonstrate elevated expression specific T cells (14) and in Th2 cell lines (19); 3) MAL was re- of DACT1 and MAL in both the CD4ϩ and CD8ϩ T cell compart- ported in a differential display screen of resting peripheral blood T ments of atopics relative to nonatopics. In addition, the kinetics of cells from atopics with dermatitis and asthma, but follow-up PCR gene expression was similar to that in unfractionated PBMC viz experiments failed to confirm these observations (56); 4) DPP4 DACT1 and MAL expression peaked early in the time course, was detected in bronchial biopsies of allergic asthmatics (57) and whereas the main Th2 effector cytokines peaked late. Several ad- in Th1 and Th2 cells (58). To our knowledge, this is the first report ditional genes of interest were identified in the experiments of Fig. of any associations of DACT1, NDFIP2, GNG8, NSMCE1, and 2, presumably unmasked as a result of the improvement in signal: RAB27B with atopy or Th2 responses. noise ratio on the microarrays resulting from enrichment of responder cells by subset purification. PECAM1 and PLXDC1 genes Gene expression profiling of inhalant allergen driven gene ϩ ϩ were reciprocally expressed in atopics and nonatopics, and sub- expression in CD4 and CD8 T cells purified from stantial GZMB expression was also detected in the CD4ϩ com- HDM-stimulated PBMC

partment in both groups. Up-regulation of PECAM1 has been pre- Downloaded from A second series of experiments was performed to identify the cel- viously reported in anti-CD3/CD28-activated CD4ϩ T cells from lular source(s) of HDM-induced gene expression within PBMC. atopic asthmatics (11), and also in a microarray screen of Th2-like ϩ Of particular interest was the potential contribution of CD4 vs CD8 T cells in the mouse (16). GZMB expression has also been ϩ CD8 T cells, as these subsets within PBMC are known to have previously reported in Th2 cells (17, 60). distinct gene expression profiles (59). As part of the experiments involving generation of the RNA samples used in Fig. 1, an additional series of replicate PBMC Gene expression profiling of inhalant allergen-driven gene http://www.jimmunol.org/ cultures was set up in parallel from the individuals (the 10 atopics expression in CD4ϩ T cells from an independent group of and 10 nonatopics from the subgroups AT3 and NA3 of study atopic patients and nonatopic controls group 2) at each time point, and CD4ϩ and CD8ϩ T cells were To confirm the preferential expression of the novel genes identified purified from HDM-stimulated and unstimulated PBMC at the end ϩ ϩ in Figs. 1 and 2 in allergen-stimulated CD4 and CD8 T cells of the cultures. Total RNA was extracted, pooled into two groups from atopics, PBMC from an additional independent panel of (AT3, pool of 10 atopics, NA3 pool of 10 nonatopics) and ana- (study group 3) subjects comprising 10 atopic and 10 nonatopic lyzed on Affymetrix microarrays (Fig. 2). For logistical reasons, it individuals were cultured in the presence or absence of HDM for

was necessary to use U133a microarrays for this part of the study by guest on October 1, 2021 24 h. T cell subsets were purified from these cultures, and total (as opposed to U133 plus 2.0 series in Fig. 1), and the U133a series RNA was extracted and pooled into four separate groups of subjects (two groups of five atopics, two groups of five nonatopics), and hybridized to Affymetrix U133 Plus 2.0 microarrays. Separate RNA samples were reserved for follow-up studies by qRT-PCR. To identify differences in the expression patterns of atopics and nonatopics, the data was visualized on a volcano plot (Fig. 3). The plot displays differential gene expression according to magnitude along the horizontal axis, and statistical significance along the vertical axis (see Materials and Methods). In the plot, genes which are located in the top right quadrant are hyperexpressed in atopics relative to nonatopics. Strikingly, the analysis revealed that all of the novel Th2-associated genes (except for GNG8) which were identified in the experiments of Figs. 1 and 2 were also hyperexpressed in the responses of CD4ϩ T cells in the atopics but not the nonatopics from this independent panel of subjects. The novel FIGURE 2. Microarray analysis of allergen-stimulated CD4ϩ and ϩ genes appear as extreme outliers on the plot indicating their im- CD8 T cell subsets in atopics and nonatopics. Additional samples of portance relative to all other genes in the genome. Analysis of PBMC from the 10 atopics and 10 nonatopics comprising the RNA pools ϩ AT3 and NA3, respectively, from Fig. 1 were cultured in the presence or these purified CD4 cells allowed the identification of additional absence of HDM for the indicated time points (12, 24, or 48 h). CD4ϩ (A) atopy-associated genes including CAMK2D and SYTL3 which and CD8ϩ T cells (B) were purified by positive selection and total RNA were not detected in unfractionated PBMC. was extracted, pooled into two groups (AT3, pool of 10 atopics, NA3, pool Several genes including DACT1, MAL, CISH, IL-9, IL-4R, of 10 nonatopics), and analyzed on U133a microarrays, using a single GNG8, and DPP4 were also hyperexpressed in the CD8ϩ com- microarray for each pool. Data expressed as stimulation ratios (expression partment of atopics. Other genes including NSMCE1, IL-5, PE- level in HDM-stimulated cells/unstimulated cells). Note that the genes CAM1, PLXDC1, and NDFIP2 were not prominent in the CD8ϩ NDFIP2, RAB27B, GNG8, and NSMCE1 from Fig. 1 are not represented responses of the atopics, however, it is not clear from these ex- on the U133a microarrays, but these were followed up in qRT-PCR studies periments whether this is due to relative insensitivity of the mi- (Table II) and in additional microarray studies (Fig. 3). Additional highly ϩ expressed genes PECAM1, PLXDC1, and GZMB, which were not identi- croarrays or to the lack of expression within the CD8 compart- fied in unfractionated PBMC in Fig. 1 were identified here and selected for ment. This and related issues were accordingly addressed using further study. more sensitive qRT-PCR methodology (see below). The Journal of Immunology 4771 p Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 26.3 0.063 0.49 3.37 0.029 0.46 0.52 0.007 1.24 0.002 0.38 0.003 0.37 0.11 0.54 0.002 0.07 0.002 0.41 0.015 0.01 0.02 0.02 0.64 0.30 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ ϩ ϭ ( n CD8 Non atopic Ϫ 69.1 Ϫ 0.62 Ϫ 4.62 Ϫ 0.18 Ϫ 1.09 Ϫ 3.47 Ϫ 0.19 Ϫ 0.05 Ϫ 0.23 28.3 0.37 4.80 1.67 1.03 1.101.89 0.10 0.78 0.43 1.02 1.22 0.75 0.60 1.05 0.96 0.01 0.86 0.03 1.71 0.06 0.81 0.003 0.94 1.22 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ ϭ Atopic ( n p Ͻ 0.001 1.58 Ͻ 0.001 2.23 Ͻ 0.001Ͻ 0.001 4.42 3.63 Ͻ 0.001 3.45 Ͻ 0.001 3.24 Ͻ 0.001 5.37 Ͻ 0.001 2.07 Ͻ 0.001 2.88 Ͻ 0.001 2.56 Ͻ 0.001 2.90 Ͻ 0.001 8.18 Ͻ 0.001 3.43 Ͻ 0.001 6.59 15.5 0.007 8.2 0.37 1.01 1.4 0.28 0.70.29 0.001 8.26 0.37 1.10 0.38 1.20 0.08 0.37 0.88 0.35 0.38 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ ϩ ϭ ( n Non atopic Ϫ 8.9 CD4 Ϫ 49.1 Ϫ 3.21 Ϫ 1.17 Ϫ 0.06 Ϫ 1.70 Downloaded from 22.6 1.78 0.99 3.235.6 0.55 9.9 2.1 1.78 1.345.30 1.54 4.40 3.76 0.78 8.50 2.19 0.17 4.49 0.62 1.51 1.57 2.80 0.35 3.61 0.96 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ ϭ Atopic ( n p http://www.jimmunol.org/ 22.5 0.529 31.8 0.34 0.004 12.15 0.67 0.247 10.61 0.222.86 0.029 0.123 14.74 21.4 3.11 0.796 33.8 0.55 0.063 nd nd 6.60 0.30 0.0290.331.24 8.74 0.035 0.089 8.28 22.70 0.19 0.105 17.08 0.29 0.165 48.40 0.06 0.043 12.04 0.23 0.123 20.25 0.47 0.105 15.22 0.03 0.009 16.64 0.83 0.035 19.38 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ ϭ ϩ ( n Non atopic Ϫ 33.6 Ϫ 1.05 Ϫ 1.23 Ϫ 0.21 Ϫ 7.60 Ϫ 0.94 Ϫ 0.48 CD8

ϩ 30.0 0.70 11.32 1.95 1.36 0.801.56 0.91 0.77 1.26 1.24 1.21 1.39 3.03 0.73 0.29 1.14 0.10 1.48 0.40 0.61 2.17 1.44 0.82 0.01 FIGURE 3. Microarray analysis of allergen-stimulated CD4 and 0.68 3.00 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ ϩ Ϯ CD8 T cell subsets in an independent panel of atopics and nonatopics. ϭ Atopic by guest on October 1, 2021 PBMC from a different group of 10 atopics and 10 nonatopics were cul- ( n Ϫ 4.60 tured in the presence or absence of HDM for 24 h. CD4ϩ (top panel) and Ϫ 15.40 CD8ϩ T cells (bottom panel) were purified by positive selection and total ϭ RNA was extracted, pooled into four groups (two atopics pools of n 5, p 0.0290.0090.001 2.71 0.043 2.08 3.44 0.043 3.50 0.0030.029 2.16 0.87 two nonatopic pools of n ϭ 5), and analyzed on U133 Plus 2.0 microarrays. Ͻ 0.001 1.02 Data are expressed as a volcano plot (see Materials and Methods). The b b b horizontal axis displays the difference (ratio) in the stimulation ratios of b e e c 11.6 0.009 0.52 0.71 0.32 2.38 1.81 0.280 6.64 0.63 0.18 0.13 0.63 0.27 0.019 2.09 0.42 0.105 1.06 0.05 0.015 2.39 0.09 0.011 1.48 0.87 0.009 6.51 0.11 0.003 1.79 0.28 0.005 4.76 atopics and nonatopics (positive values indicate higher expression level in 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ ϩ atopics, negative values indicate higher expression levels in nonatopics). ϭ ( n 1.35 1.63 0.53 2.59 Non atopic Ϫ 31.7 Ϫ 2.23 Ϫ 1.60 Ϫ 0.46 Ϫ 2.25 The vertical axis displays the statistical significance (moderated t test sta- Ϫ 0.11 Ϫ 10.77 tistic) of the difference in the SR of atopics and nonatopics (higher values indicate increasing significance). Genes appearing in or round about the top b b c c c b b 1.33 21.0 1.36 5.71 2.73 0.91 3.67 0.83 2.19 0.39 3.67 2.92 0.09 1.27 0.25 4.32 3.04 1.89 0.02 1.21 2.80 3.79 1.13 right quadrant are clearly hyperexpressed in atopics relative to nonatopics, 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ ϭ and these genes were selected for further study. Data expressed on the log2 Atopic ( n scale. 5.43 a p Ͻ 0.001 7.62 Ͻ 0.001 11.93 Ͻ 0.001 7.04 Ͻ 0.001 3.47 qRT-PCR validation of microarray findings Ͻ 0.001 6.41 The microarray experiments reported above were performed on 0.58 0.015 32.4 0.73 0.47 1.13 0.003 1.33 0.47 0.54 0.46 0.002 17.73 0.09 0.20 0.004 10.81 0.51 0.005 7.15 pooled RNA samples and therefore do not provide information on 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ nd subject-to-subject variation in gene expression patterns. To obtain ϭ ( n 12 h 12 h 24 h 24 h 24 h Non atopic PBMC CD4 more detailed information in this regard, and to formally demon- Ϫ 1.58 strate statistically significant differences in gene expression pat- Patient Group 1 Patient Group 2 Patient Group 2 Patient Group 3 Patient Group 3 0.79 1.81 0.30 1.30 0.92 2.75 2.45 2.87 0.65 1.03 1.09 1.12 0.82 0.70 0.10 0.46 0.46 terns between atopics and nonatopics, reserved individual RNA 1.42 0.72 10) Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ nd samples were assayed by qRT-PCR. The results of these studies ϭ Atopic ( n 5.42 5.01

are summarized in Table II. In the table, study group 1 refers to the 10.21 test. Patient group 1 refers to the individual RNA samples from the subjects pooled into the groups AT1, AT2, NA1, NA2 (from Fig. 1); patient group 2 refers to individuals from RNA pools AT3, and NA3 (from Figs. 1 and individual RNA samples from the subjects pooled into the groups U

AT1, AT2, NA1, NA2 (from Fig. 1). Study group 2 refers to the qRT-PCR validation of microarray results individuals from RNA pools AT3, and NA3 used in the experi- Data are presented as gene expression levels above background after HDMTime stimulation, point normalized was to 24 the h, stably not expressed 12 EEF1A1. h; Data nd, are not expressed done. as mean and SE. Note that negative data in the table indicates . Statistical analysis by Time point was 48 h, not 12 h. Cell Type Time point Individuals GZMB nd nd 5.43 PECAM1 1.05 CAMK2DSYTL3 nd nd nd nd 4.38 0.65 IL17RB nd nd 1.87 NSMCE1CISH nd nd nd nd 5.46 3.31 PLXDC1 7.70 IL-4 6.44 GNG8 12.44 IL-5 IL-13 DPP4 MAL DACT1 2.17 NDFIP2 2.10 RAB27B 5.64 ments from Figs. 1 and 2. Study group 3 refers to the subjects used a b c 2); patient group 3 refers to the subjects used in the experiments in Fig. 3. Table II. in the experiments in Fig. 3. Mann-Whitney 4772 GENE EXPRESSION IN INHALANT ALLERGY

the latter are associated selectively with activation of Th2 memory cells. However, these genes have escaped detection in previous microarray studies which have typically used methodology radi- cally different to the present study (14–20), including initial enrichment of putative Th2 cells via initial culture under conditions which favor survival of the cells of interest, followed by reactivation with mitogenic stimuli. One such approach which is standard in this area involves flow cytometric positive selection of cells which have been driven through repeated cycles of division by culture with specific allergen (thus revealing their allergen-specific memory phenotype), and subsequently comparing gene expression FIGURE 4. Relative expression of IL-5 and IL-13 mRNA in HDM- patterns in these selected cells at rest vs following reactivation by stimulated CD4ϩ and CD8ϩ T cell subsets from atopics. Data expressed as brief exposure to the potent mitogen PMA/ionomycin. We repro- gene expression levels above background after HDM stimulation. Data are normalized to the stably expressed gene EEF1A1. Statistical analysis by duced this experimental system in the study in Fig. 5, which il- Mann-Whitney U test. lustrates gene expression ratio (stimulated/unstimulated) in atopics, plotted against average signal strength. Consistent with earlier The data of primary interest in Table II are those derived from reports (e.g., Ref. 25), clear differential up-regulation of the Th2 CD4ϩ and CD8ϩ populations in which enrichment for responder reference genes (IL-4, IL-5, IL-9, and IL-13) can be seen in the cells may be expected to reduce the background noise relative to atopics together with a range of additional pan-specific T cell ac- Downloaded from unfractionated PBMC (59, 61). Our principal focus was upon tivation markers (IL-2, IL-3, IL-17, IFNG, TNF, CSF2, and IL- 12/24 h time points which encompass the peak of the early Th2- 2RA). However, the novel Th2-associated genes which were prom- associated cluster, and where specified genes which did not vali- inent in our primary T cell cultures stimulated only with allergen, date at these time points were followed up further in 48-h samples. most notably the highly expressed early genes DACT1, MAL, Consistent with the published literature (62), an allergen-specific DPP4, IL-17RB, and PLXDC1, were not up-regulated in the cytokine response was detected in the CD8ϩ compartment as well CFSE-selected T cell lines (Fig. 5). as in the CD4ϩ T cells, characterized by the consistent up-regu- We hypothesized that the explanation for this discrepancy could http://www.jimmunol.org/ lation of the Th2 index genes IL-4, IL-5, and IL-13. However, the be that the more physiological methods for allergen-induced T cell CD8ϩ signal was attenuated relative to that in CD4ϩ T cells (study activation used in the present study used TCR-associated activa- group 3; Fig. 4), and this was reflected in the qRT-PCR data in tion pathway(s) that are bypassed in the cell line approaches, in Table II in terms of correspondingly lower and more variable p particular, those encompassing the use of high-intensity activation values for atopic:nonatopic comparisons across many (but not all) stimuli which have a lower requirement for (or are often indepen- of the CD8ϩ gene list. The notable exceptions were DACT1, dent of) help from APCs. To investigate this issue further, we PLXDC1, and DPP4, which displayed an equally high level of analyzed expression of the novel Th2-associated signaling genes in primary T cells after stimulation with a panel of alternative specific activity and selectivity within both cellular compartments by guest on October 1, 2021 of the atopic response, and GNG8 which validated in the PBMC activating agents comprising the polyclonal mitogens soluble and CD8ϩ compartments, but not in the CD4ϩ compartment. It is noteworthy that the rare nonatopic individuals who produced detectable HDM-specific Th2 responses (around 10% of the nonatopics studied) also exhibited elevated expression of the novel genes (data not shown), indicating their association with the Th2 cytokine response per se as opposed to some other aspect of the atopic phenotype. Moreover, preliminary studies with PBMC from panels of peanut-allergic subjects and controls stimulated with mixed peanut allergen also revealed expression of the novel genes which was restricted to the atopic responses (data not shown), suggesting that these findings may be broadly applicable in human atopy. It is additionally of note that a further follow-up case/control study was performed with archived RNA samples from a randomly selected subgroup of 24 HDM skin prick test-positive atopics and 24 nonatopic 11-year olds from a recently published cohort study FIGURE 5. Microarray analysis of gene expression in allergen-driven (24). These comprised 48-h samples derived from PBMC cultured cell lines after polyclonal restimulation. PBMC from two atopic donors with/without HDM allergen. Atopy-specific expression of the were labeled with CFSE and stimulated with HDM for 6 days, and dividing novel genes was confirmed for DACT1 ( p ϭ 0.012 by Mann- effector memory cells (CFSElow) were isolated by cell sorting. The cells Whitney), MAL ( p ϭ 0.024), PLXDC1 ( p ϭ 0.008), CISH ( p Ͻ were rested overnight, split into replicate wells, and one well was restim- 0.001), IL-17RB ( p Ͻ 0.001), and RAB27B ( p ϭ 0.039), but not ulated with PMA and ionomycin. Total RNA was extracted, amplified, and for NDF1P2 ( p ϭ 0.514) or GNG8 ( p ϭ 0.101), suggesting that analyzed on U133a microarrays, using a single microarray for each donor. Data expressed as microarray plot; the vertical axis displays the stimulation expression of the latter two genes is more tightly linked to the ratio (expression level in restimulated cells/unstimulated cells) and the hor- earlier phase of T cell activation. izontal axis displays the average signal (see Materials and Methods). Note Expression of the Th2 signaling cluster depends on the nature of the log2 scale on both axes. Positive and negative values on the y-axis the stimulus received by the T cell indicate genes which are up-regulated and down-regulated, respectively, after polyclonal restimulation. The horizontal lines on the plot indicate a The consistent correlations between the panel of Th2 reference 2-fold change on the linear scale. The data are the average results obtained genes and the novel genes identified above strongly suggest that in cell lines generated from two atopic donors. The Journal of Immunology 4773 anti-CD3/IL2 and PHA, and the oligoclonal T cell stimulant bacterial superantigen (SEB) which activates T cells expressing a defined range of TCRV␤ molecules (63). PBMC from six patients with severe atopic dermatitis were cultured for 12 h in the presence or absence of the selected stimuli. The T cells responded to the mitogen panel with a mixed Th1/ Th2 profile as demonstrated by IL-4, IL-5, IL-13, and IFNG production (Fig. 6). Strikingly, DACT1, MAL, DPP4, and NDFIP2 were all up-regulated after SEB stimulation, but only NDFIP2 was up-regulated in response to PHA. Furthermore, MAL was down- regulated after PHA stimulation, and not significantly up-regulated after anti-CD3 stimulation ( p ϭ 0.11). Additionally, consistent with earlier findings with CFSE-selected cell lines, a follow-up FIGURE 7. Selective attenuation of a subset of proinflammatory genes ϩ study on a panel of primary PBMC T cells stimulated with PMA/ in Th1-like HDM responses in CD4 T cells from nonatopics relative to ionomycin also demonstrated up-regulation of Th2 effector genes PPD responses: qRT-PCR validation of microarray findings. PBMC cultures from a group of eight healthy adults were stimulated with PPD for in the absence of accompanying activation of these novel genes ϩ (data not shown). 48 h. CD4 T cells were purified, RNA was extracted and pooled, and analyzed on a single U133a microarray. The expression profile was com- ϩ Nature of the immune response in nonatopics pared with those of HDM-stimulated CD4 T cells from the nonatopics

used in Fig. 2 (at the 48-h time point). A broad range of differentially Downloaded from We finally sought to characterize the T cell response to HDM in expressed Th1-associated genes including IL-1B, -3, -6, -8, -12RB, -17, nonatopic individuals in more detail. This response is currently -22, CCL1, -2, -3, -7, -8, -20, -22; CXCL1, -3, -5, 9, -10, TNFRSF6, assumed to be Th1-like (64, 65), and as noted above a significant SOCS1, LTA, TNF, IRF1, and STAT1 appeared 3- to 10-fold more highly number of Th1 marker genes including GZMB, LTA, and IFNG expressed in the PPD response relative to the response to HDM in nona- were detected in HDM-stimulated PBMC from the nonatopics in topics (data not shown). The striking exception identified by the microarray Ͼ the microarray experiments in Figs. 1 and 2. Similar findings have screen was IFNG, which appeared 100-fold higher for PPD, and to a

CCL8 GZMB http://www.jimmunol.org/ been reported by many groups, however, the lack of evidence for lesser extent and . Analysis of gene expression in the individual RNA samples by qRT-PCR confirmed the selective down-regulation expression of Th1-associated delayed-type hypersensitivity (DTH) of IFN-␥ and CCL8 in the nonatopic responses relative to other proinflam- to this or other allergens among nonatopics, despite lifetime aller- matory genes. Data expressed as the ratio of gene expression after PPD gen exposure, has led investigators to question the applicability of stimulation (average of 8 individuals) to gene expression after HDM stim- the murine Th1/Th2 paradigm in this context (66). We speculated ulation (average of 10 individuals). that this apparent paradox may be explicable by covert quantitative or possibly qualitative differences between the allergen-specific Th1-like HDM responses of nonatopics, and classical Th1 immu- expression profiles compared with those of HDM-stimulated nity associated with DTH expression exemplified by the Mantoux by guest on October 1, 2021 ϩ response to mycobacterial Ag. To test this possibility, we com- CD4 T cells from the nonatopics used in Fig. 2. A broad range of pared gene expression profiles in nonatopic HDM responses with differentially expressed Th1-associated genes that were common to those in Th1-polarized responses to mycobacterial PPD. Pooled both responses were identified as per Fig. 1 (data not shown). An RNA from PPD-stimulated CD4ϩ T cells from a group of eight initial comparison of signal strength on the microarray was made healthy adults with a previous bacillus Calmette-Gue´rin vaccina- on a Th1 gene-by-gene basis between the two responses, and this tion history were screened by microarray as per Fig. 2, and their analysis revealed that almost all of the relevant genes appeared 3- to 10-fold more highly expressed in the PPD response relative to the response to HDM in nonatopics. The striking exception was IFNG which appeared Ͼ100-fold higher for PPD, and to a lesser extent CCL8 and GZMB. In the experiments in Fig. 7, a panel of the most highly expressed Th1-associated genes were selected for qRT-PCR validation, using individual RNA samples from the pools, and this confirmed selective enrichment of the IFNG and CCL8 components within the PPD response.

Discussion Although there is general acceptance that allergic disease is mediated via allergen-induced hyperexpression of Th2 cytokines at challenge sites, the full spectrum of Th2-associated genes which participate in this inflammatory cascade await identification. Ad- ditionally, the immunological basis for allergen responder phenotype (viz nonatopic vs atopic) and for allergy intensity among atopics, remains ill-defined. In particular, the basis for resistance to FIGURE 6. Expression of novel Th2-associated signaling genes is de- allergic sensitization (i.e., avoidance of generation of Th2-polar- pendent on the conditions of T cell activation. PBMC from six patients ized allergen-specific memory) is still debated (66), especially is- with atopic dermatitis were cultured in the presence or absence of soluble anti-CD3/IL-2, SEB, or PHA for 12 h. Total RNA was extracted and gene sues relating to the contribution of different forms of tolerance to expression was analyzed by qRT-PCR. Data were normalized to the stably clinical unresponsiveness in this group. expressed gene EEF1A1 and expressed as mean and SE. For statistical In the current study, microarray technology was used to obtain analysis, the data were log transformed, and significance was calculated more detailed information on patterns of allergen-induced gene p Ͻ 0.05; †, p Ͻ 0.01; ‡, p Ͻ 0.001; §, p Ͻ 0.0001. expression in T memory responses in atopics and nonatopics, using ,ء .using a paired t test 4774 GENE EXPRESSION IN INHALANT ALLERGY short-term primary cultures. Despite the low frequency of re- of several of these novel signaling-associated genes in a different sponder cells within unfractionated PBMC, we have demonstrated experimental setting, notably in oligoclonal and polyclonal Th2- consistent hyperexpression of classical Th2 reference genes in al- like responses initiated via stimulants which mediate T cell acti- lergen-stimulated PBMC from atopics, in particular IL-4 which is vation by APC-dependent interactions with the TCR (anti-CD3 expressed at only low levels. This suggests that the microarray and SEB). However, these same genes appear redundant in T cell technology used here has sufficient sensitivity to yield cogent in- responses initiated by more potent mitogenic stimuli such as PHA formation from mixed PBMC samples in this experimental setting. and PMA/ionomycin, indicating that alternative intracellular-sig- Microarray analyses also revealed that an additional Th1-like gene naling pathways mediate T cell activation under these less physi- expression signature which includes LTA, IFNG, and a number of ological conditions of stimulation. IFNG-induced genes which was qualitatively similar to that ob- This latter finding has important practical implications as it served in responses to mycobacterial PPD, was present in the demonstrates how the validity of T cell-related drug target identi- HDM responses (data not shown), regardless of the atopic status of fication by microarray screening can be potentially compromised the PBMC donors. This is consistent with findings from a range of by the choice of stimulants applied to the cells being screened. It laboratories (reviewed in Ref. 67) which indicate that the cytokine is additionally of interest to note in this context that SEB has been phenotype of Th-memory responses to inhalant allergens in atopics suggested to play an important ancillary role in driving atopic dis- is typically of the mixed Th1/Th2 or Th0 phenotype. Recent evi- ease pathogenesis in both the skin (72) and the upper respiratory dence from our laboratory (24) and others (68) (reviewed in Ref. tract (73), and the potency of this molecule as an oligoclonal T cell 69) suggests that Th1-like mechanisms superimposed on Th2 re- stimulant capable of mimicking the effects of specific allergen on sponses in atopics may exacerbate allergic disease expression, and atopic Th2 memory cells, may account for this facet of its biolog- further microarray based studies targeted specifically at this class ical activity. Downloaded from of genes have potential to shed further light on this issue. Further studies are required to determine the importance and The substantial overlap between the PPD and nonatopic HDM function of the novel Th2-associated genes in allergic responses, responses was noteworthy. However, unlike PPD, the Th1-like re- however some hints as to their potential roles can be deduced from sponse against HDM is not associated with potential for DTH, what is known in other areas of the literature: which suggests that some form of covert internal regulation atten- DACT1 has not previously been implicated in T cell regulation. uates the overall intensity of the nonatopic HDM response. It is However, studies in mesenchymal cells have demonstrated that http://www.jimmunol.org/ feasible that this may be due (at least in part) to a comparatively overexpression of DACT1 blocks disheveled-mediated activation lower frequency of HDM responsive T memory cells in vivo com- of the JNK and ␤-catenin-signaling pathways (74) and by infer- pared with those participating in PPD responses, and consistent ence, DACT1 may also regulate these pathways in T cells. ␤-cate- with this possibility, we report a generalized 3- to 10-fold increase nin is required for normal T cell development, and ␤-catenin-de- in Th1 gene expression levels in the latter. However, if variations ficient T cells show defective proliferation in response to anti-CD3 in responder cell frequencies accounted entirely for the differences mAb (75). Moreover, increased expression of the ␤-catenin target in the intensity of in vivo reactivity, then there should be broad genes LEF-1 and TCF-1 has also been reported in peripheral T cell consistency in relative expression levels across the entire spectrum lymphomas expressing cell surface markers characteristic of Th1 by guest on October 1, 2021 of inflammation-associated genes, comparing the two in vitro T cells, but is absent in lymphomas expressing Th2-associated mark- memory responses. Instead, as illustrated in Fig. 7, a small subset ers (76). Regulation of the JNK-signaling pathway by DACT1 is a of proinflammatory genes (most notably IFNG and CCL8) are ex- plausible possibility because JNK regulates NFAT activity, which pressed on a log-fold higher scale in T memory responses to the in turn regulates expression of multiple target genes including the classical DTH stimulant PPD. This suggests that these potent Th2 cytokines (reviewed in Ref. 77). JNK1-deficient T cells also proinflammatory genes are preferentially attenuated in the HDM preferentially differentiate into Th2 cells (78). responses of the nonatopics. A precedent for this form of selective MAL localizes to lipid raft domains in T cells in association with regulation is the “modified Th2 response” to cat allergen recently several TCR-signaling molecules including Lck, suggesting a role described by Platts-Mills et al. (70), in which a proportion of nona- in T cell activation (Ref. 79, reviewed in Ref. 80). It is also note- topics exhibit ongoing production of IL4-dependent IgG4 Ab worthy that MAL is induced by the ICOS-signaling pathway (81), against cats in the absence of parallel IgE production or skin prick which is required for optimal Th2 effector function (82). MAL has reactivity to the same allergen. Both these response patterns to been shown to play a role in raft-dependent protein trafficking environmental allergens may reflect forms of split tolerance (71) (Ref. 83, reviewed in Ref. 84), a process which is central to T cell resulting from chronic exposure, and this possibility will be ad- activation (85). By inference, MAL may thus play a role in estab- dressed in follow-up studies. lishment of the immunological synapse (86) during the activation More detailed analysis of the HDM response in atopics using process. QT clustering identified several genes that exhibited expression DPP4 localizes to lipid raft domains in T cells, and interaction patterns similar to the Th2 reference genes. Two major waves of with adenosine deaminase anchored to dendritic cells provides a synchronous gene expression were identified in the atopics, an costimulatory signal augmenting T cell activation (87, 88). It is of early group peaking in activity between 12 and 24 h, which was interest to note that in cell lines MAL has been demonstrated to enriched for genes involved in signaling, and a late cluster peaking mediate intracellular trafficking of DPP4 (83), and interactions be- at 48 h enriched for genes associated with Th2 effector function. tween these molecules may thus serve to modulate T cell activa- Quantitative RT-PCR validation experiments confirmed the asso- tion threshold. In the direct context of atopy, DPP4 has also been ciation of these genes with atopy in independent panels of subjects, reported to augment allergen-specific IgE production, regulate che- and cell separation experiments demonstrated that all of the genes mokine activity, and influence T cell recruitment in a rat model of (except for GNG8) were induced in purified CD4ϩ and CD8ϩ T asthma (89, 90). cells, although gene expression levels in the latter were generally NDFIP2 interacts with Nedd4, an E3 ubiquitin ligase (91) and lower. Elevated expression of the GNG8 gene was also associated promotes NF-␬B signaling (92). In anergic T cells, the E3 ubiq- with atopy, however it was restricted to the CD8ϩ T cell subset. uitin ligases Itch and Nedd4 translocate to lipid raft domains and Additionally, we have shown large-scale and consistent activation flag signaling proteins for degradation, thereby inhibiting T cell The Journal of Immunology 4775 activation (93). NDFIP2 regulation of Nedd4 activity (91), may CAMKII is also activated by disheveled signaling (122), suggest- thus influence T cell activation and signaling. NDFIP2 also has ing a functional linkage with DACT1. been reported to localize with multivesicular bodies and may also NSMCE1 and GNG8 have no known functions in T cells. play a role in protein trafficking (94). It is noteworthy that the most Note added in proof. During the manuscript submission and re- commonly used family of drugs used for treatment of chronic in- view process, MAL was reportedly induced by IL-4 during the halant allergy, the corticosteroids, target NF-␬B signaling (95). early polarization of human Th2 cells (Lund, R. H. Ahlfors, RAB27B is highly homologous to RAB27A, and they have sim- E. Kainonen, A. M. Lahesmaa, C. Dixon, and R. Lahesmaa. 2005. ilar functions, however, they also have distinct expression patterns Identification of genes involved in the initiation of human Th1 or and are therefore not redundant (96). Mutations in the latter cause Th2 cell commitment. Eur. J. Immunol. 35: 3307–3319). Griscelli syndrome, a disease which is characterized by inter alia attenuated CD8ϩ T cell (CTL) cytolytic activity resulting from Acknowledgments defective secretion of lytic granules containing GZMB (97, 98). We thank Barbara Holt and Jenny Tizard for help and advice with bio- The high level of GZMB expression observed in atopic CD4 T banking and methodology for cellular immunology, and Matt Wikstro¨m for cells in the current study was noteworthy. Studies in gene-targeted cell sorting. and mutant mice have demonstrated that CD4 and CD8 T cells mainly use the FAS and granule exocytosis pathways (perforin Disclosures dependent), respectively, to mediate cytotoxic effector functions The authors have no financial conflict of interest. (99–101). However, recent evidence in human systems has demonstrated that CD4 T cells express GZMB and display perforin- References dependent cytolytic activity (102, 103). CD4 T cells expressing 1. Romagnani, S. 2000. The role of lymphocytes in allergic disease. J. Allergy Clin. Downloaded from Immunol. 105: 399–408. both GZMB and perforin have also been detected in lesions from 2. Wills-Karp, M. 2000. Murine models of asthma in understanding immune dys- atopic dermatitis patients (104). RAB27A has been reported to in- regulation in human asthma. Immunopharmacology 48: 263–268. 3. Sole´r, M., J. Matz, R. Townley, R. Buhl, J. O’Brien, H. Fox, J. Thirlwell, teract with several synaptotagmin-like proteins including SYTL3, N. Gupta, and G. Della Cioppa. 2001. The anti-IgE antibody omalizumab reduces which are required for lysosomal secretion (98, 105). exacerbations and steriod requirement in allergic asthmatics. Eur. Respir. J. 18: CISH is a member of the suppressor of cytokine signaling pro- 254–261.

4. Bryan, S. A., B. J. O’Connor, S. Matti, M. J. Leckie, V. Kanabar, J. Khan, http://www.jimmunol.org/ tein family and inhibits STAT5 signaling (reviewed in Ref. 106). In S. J. Warrington, L. Renzetti, A. Rames, J. A. Bock, et al. 2000. Effects of transgenic mice, CISH overexpression promotes Th2 differentia- recombinant human interleukin-12 on eosiniphils, airway hyper-responsiveness, tion (107). and the late asthmatic response. Lancet 356: 2149–2153. 5. Hart, T. K., M. N. Blackburn, M. Brigham-Burke, K. Dede, N. Al-Mahdi, IL-17RB is the receptor for IL-17B and IL-17E (IL-25). Treat- P. Z. Zia-Amirhosseini, and R. M. Cook. 2002. Preclinical efficacy and safety of ment of mice with IL-17E induced the expression of IL-4, IL-5, pascolizumab (SB 240683): a humanized anti-interleukin-4 antibody with ther- apeutic potential in asthma. Clin. Exp. Immunol. 130: 93–100. and IL-13, and the development of allergic symptoms such as eo- 6. Leckie, M. J., A. ten Brinke, J. Khan, Z. Diamant, B. J. O’Connor, C. M. Walls, sinophilia and mucous hypersecretion (108). A. K. Mathur, H. C. Cowley, K. F. Chung, R. Djukanov, et al. 2000. Effects of PECAM1 (CD31) is an adhesion molecule expressed on plate- an interleukin-5 blocking monoclonal antibody on eosinophils, airway hyper- responsiveness, and the late asthmatic response. Lancet 356: 2144–2148. lets and leukocytes. PECAM1 contains immunoreceptor tyrosine 7. Tang, Z., and A. Saltzman. 2004. Understanding human dendritic cell biology by guest on October 1, 2021 inhibitory motifs and regulates Ag-induced T cell and B cell ac- through gene profiling. Inflamm. Res. 53: 424–441. tivation (reviewed in Ref. 109), and PECAM1 knockout mice are 8. van der Pouw Kraan, T. C., P. V. Kasperkovitz, N. Verbeet, and C. L. Verweij. 2004. Genomics in the immune system. Clin. Immunol. 111: 175–185. prone to develop autoimmunity (110, 111). A study in Jurkat T 9. Saito, H., J. Abe, and K. Matsumoto. 2005. Allergy-related genes in microarray: cells reported that PECAM1 inhibited early calcium mobilization an update review. J. Allergy Clin. Immunol. 116: 56–59. 10. Hansel, N. N., S. C. Hilmer, S. N. Georas, L. M. Cope, J. Guo, R. A. Irizarry, and from intracellular stores (112). It is known that transient (as op- G. B. Diette. 2005. Oligonucleotide-microarray analysis of peripheral-blood lym- posed to sustained) calcium flux favors Th2 differentiation and phocytes in severe asthma. J. Lab. Clin. Med. 145: 263–274. signaling pathways (Ref. 113; reviewed in Ref. 114), and attenu- 11. Schmidt-Weber, C. B., J. G. Wohlfahrt, and K. Blaser. 2001. DNA arrays in allergy and immunology. Int. Arch. Allergy Immunol. 126: 1–10. ation of calcium mobilization during T cell activation via 12. Wohlfahrt, J. G., S. Kunzmann, G. Menz, W. Kneist, C. A. Akdis, K. Blaser, and PECAM1 may thus favor expression of the Th2 response pheno- C. B. Schmidt-Weber. 2003. T cell phenotype in allergic asthma and atopic der- type. PECAM1 also regulates ␤-catenin levels, suggesting a func- matitis. Int. Arch. Allergy Immunol. 131: 272–282. 13. Li, X. D., D. M. Essayan, M. C. Liu, T. H. Beaty, and S. K. Huang. 2001. tional link between DACT1 and PECAM1. Profiling of differential gene expression in activated, allergen-specific human Th2 PLXDC1 has recently been demonstrated to bind cortactin cells. Genes Immun. 2: 88–98. 14. Lindstedt, M., A. Schiott, A. Bengtsson, K. Larsson, M. Korsgren, L. Greiff, and (115), a protein that activates Arp2/3 complexes (reviewed in Ref. C. A. K. Borrebaeck. 2005. Genomic and functional delineation of dendritic cells 116). These complexes initiate actin polymerization, which is im- and memory T cells derived from grass pollen-allergic patients and healthy in- portant for many facets of T cell biology, including activation, dividuals. Int. Immunol. 17: 401–409. 15. Rogge, L., E. Bianchi, M. Biffi, E. Bono, S. Y. Chang, H. Alexander, C. Santini, adhesion, and migration. G. Ferrari, L. Sinigaglia, M. Seiler, et al. 2000. Transcript imaging of the devel- CAMKII is a prominent kinase of the CNS which is important in opment of human T helper cells using oligonucleotide arrays. Nat. Genet. 25: cognitive functions including learning and memory (117). Four 96–101. 16. Chtanova, T., R. A. Kemp, A. P. Sutherland, F. Ronchese, and C. R. Mackay. closely related genes (␣, ␤, ␦, and ␥) encode distinct isoforms of 2001. Gene microarrays reveal extensive differential gene expression in both CAMKII, and alternative splicing of the mRNA transcripts derived CD4ϩ and CD8ϩ type 1 and type 2 T cells. J. Immunol. 167: 3057–3063. 17. Lund, R., T. Aittokallio, O. Nevalainen, and R. Lahesmaa. 2003. Identification of from these genes generates a considerable number of variants novel genes regulated by IL-12, IL-4, or TGF-␤ during the early polarization of which assemble into homo- or heteromultimers of 8–12 catalytic CD4ϩ lymphocytes. J. Immunol. 171: 5328–5336. subunits (117). A role for CAMKII variants encoded by the ␥ and 18. Lu, B., P. Zagouras, J. E. Fischer, J. Lu, B. Li, and R. A. Flavell. 2004. Kinetic ␤ analysis of genomewide gene expression reveals molecule circuitries that control genes has been previously reported in T cells (118, 119), how- T cell activation and Th1/2 differentiation. Proc. Natl. Acad. Sci. USA 101: ever, the function of CAMK2D which is encoded by the ␦ gene is 3023–3028. unknown in T cells. Calmodulin kinases are activated by the cal- 19. Chtanova, T., S. G. Tangye, R. Newton, N. Frank, M. R. Hodge, M. S. Rolph, and C. R. Mackay. 2004. T follicular helper cells express a distinctive transcriptional cium/calmodulin-signaling pathway and regulate gene expression profile, reflecting their role as non-Th1/Th2 effector cells that provide help for B by regulating the activity/expression of transcription factors and cells. J. Immunol. 173: 68–78. ␤ 20. Cao, W., Y. Chen, S. Alkan, A. Subramaniam, F. Long, H. Liu, R. Diao, coactivators including CREB, CREB-binding protein, C/EBP , se- T. Delohery, J. McCormick, R. Chen, et al. 2005. Human T helper (Th) cell rum response factor, and AP1 (120, 121); it is noteworthy that lineage commitment is not directly linked to the secretion of IFN-␥ or IL-4: 4776 GENE EXPRESSION IN INHALANT ALLERGY

characterization of Th cells isolated by FACS based on IFN-␥ and IL-4 secretion. ferentiates memory from naive CD4ϩ T cells and is a molecular mechanism for Eur. J. Immunol. 35: 2709–2717. enhanced cellular response of memory CD4ϩ T cells. J. Immunol. 166: 21. Chen, Z., R. Lund, T. Aittokallio, M. Kosonen, O. Nevalainen, and R. Lahesmaa. 7335–7344. 2003. Identification of novel IL-4/Stat6-regulated genes in T lymphocytes. J. Im- 50. Agrawal, D., T. Chen, R. Irby, J. Quackenbush, A. F. Chambers, M. Szabo, munol. 171: 3627–3635. A. Cantor, D. Coppola, and T. J. Yeatman. 2002. Osteopontin identified as lead 22. Imada, M., F. E. R. Simons, F. T. Jay, and K. T. HayGlass. 1995. Antigen marker of colon cancer progression, using pooled sample expression profiling. mediated and polyclonal stimulation of human cytokine production elicit quali- J. Natl. Cancer Inst. 94: 513–521. tatively different patterns of cytokine gene expression. Int. Immunol. 7: 229–237. 51. Kendziorski, C., R. A. Irizarry, K. S. Chen, J. D. Haag, and M. N. Gould. 2005. 23. Ahmadzadeh, M., and D. L. Farber. 2002. Functional plasticity of an antigen- On the utility of pooling biological samples in microarray experiments. Proc. specific memory CD4 T cell population. Proc. Natl. Acad. Sci. USA 99: Natl. Acad. Sci. USA 102: 4252–4257. 11802–11807. 52. Kim, S. K., J. Lund, M. Kiraly, K. Duke, M. Jiang, J. M. Stuart, A. Eizinger, 24. Heaton, T., J. Rowe, S. Turner, R. C. Aalberse, N. de Klerk, D. Suriyaarachchi, B. N. Wylie, and G. S. Davidson. 2001. A gene expression map for Caenorhab- M. Serralha, B. J. Holt, E. Hollams, S. Yerkovich, et al. 2005. An immunoepi- ditis elegans. Science 293: 2087–2092. demiological approach to asthma: identification of in vitro T-cell response pat- 53. Staudt, L. M., and P. O. Brown. 2000. Genomic views of the immune system. terns associated with different wheezing phenotypes amongst 11 year olds. Lancet Annu. Rev. Immunol. 18: 829–859. 365: 143–150. 54. Schroder, K., P. J. Hertzog, T. Ravasi, and D. A. Hume. 2004. Interferon-␥:an 25. Turcanu, V., S. J. Maleki, and G. Lack. 2003. Characterization of lymphocyte overview of signals, mechanisms and functions. J. Leukocyte Biol. 75: 163–189. responses to peanuts in normal children, peanut-allergic children, and allergic 55. Anderson, P., A. Sundstedt, L. Li, E. J. O’Neill, S. Li, D. C. Wraith, and P. Wang. children who acquired tolerance to peanuts. J. Clin. Invest. 111: 1065–1072. 2003. Differential activation of signal transducer and activator of transcription 26. Wang, X., and B. Seed. 2003. A PCR primer bank for quantitative gene expres- (STAT)3 and STAT5 and induction of suppressors of cytokine signalling in Th1 sion analysis. Nucleic Acids Res. 31: e154. and Th2 cells. Int. Immunol. 15: 1309–1317. 27. Hamalainen, H. K., J. C. Tubman, S. Vikman, T. Kyrola, E. Ylikoski, 56. Matsumoto, Y., T. Oshida, I. Obayashi, Y. Imai, K. Matsui, N. L. Yoshida, J. A. Warrington, and R. Lahesmaa. 2001. Identification and validation of en- N. Nagata, K. Ogawa, M. Obayashi, T. Kashiwabara, et al. 2002. Identification dogenous reference genes for expression profiling of T helper cell differentiation of highly expressed genes in peripheral blood T cells from patients with atopic by quantitative real-time RT-PCR. Anal. Biochem. 299: 63–70. dermatitis. Int. Arch. Allergy Immunol. 129: 327–340. 28. Pritchard, C. C., L. Hsu, J. Delrow, and P. S. Nelson. 2001. Project normal: 57. van der Velden, V. H., A. F. Wierenga-Wolf, P. W. Adriaansen-Soeting, Downloaded from defining normal variance in mouse gene expression. Proc. Natl. Acad. Sci. USA S. E. Overbeek, G. M. Moller, H. C. Hoogsteden, and M. A. Versnel. 1998. 98: 13266–13271. Expression of aminopeptidase N and dipeptidyl peptidase IV in the healthy and 29. Brem, R. B., G. Yvert, R. Clinton and L. Kruglyak. 2002. Genetic dissection of asthmatic bronchus. Clin. Exp. Allergy 28: 110–120. transcriptional regulation in budding yeast. Science 296: 752–755. 58. Boonacker, E. P., E. A. Wierenga, H. H. Smits, and C. J. Van Noorden. 2002. 30. Schadt, E. E., S. A. Monks, T. A. Drake, A. J. Lusis, N. Che, V. Colinayo, CD26/DPPIV signal transduction function, but not proteolytic activity, is directly T. G. Ruff, S. B. Milligan, J. R. Lamb, G. Cavet, et al. 2003. Genetics of gene related to its expression level on human Th1 and Th2 cell lines as detected with expression surveyed in maize, mouse and man. Nature 422: 297–302. living cell cytochemistry. J. Histochem. Cytochem. 50: 1169–1177. 31. Morley, M., C. M. Molony, T. M. Weber, J. L. Devlin, K. G. Ewens, 59. McLaren, P. J., M. Mayne, S. Rosser, T. Moffatt, K. G. Becker, F. A. Plummer, http://www.jimmunol.org/ R. S. Spielman, and V. G. Cheung. 2004. Genetic analysis of genome-wide vari- and K. R. Fowke. 2004. Antigen-specific gene expression profiles of peripheral ation in human gene expression. Nature 430: 743–747. blood mononuclear cells do not reflect those of T-lymphocyte subsets. Clin. 32. Rogge, L. 2003. Gene profiling for defining targets for new therapeutics in au- Diagn. Lab. Immunol. 11: 977–982. toimmune diseases. Arthritis Res. Ther. 5: 47–50. 60. Lancki, D. W., C. S. Hsieh, and F. W. Fitch. 1991. Mechanisms of lysis by 33. Cheung, V. G., L. K. Conlin, T. M. Weber, M. Arcaro, K. Y. Jen, M. Morley, and cytotoxic T lymphocyte clones. Lytic activity and gene expression in cloned R. S. Spielman. 2003. Natural variation in human gene expression assessed in antigen-specific CD4ϩ and CD8ϩ T lymphocytes. J. Immunol. 146: 3242–3249. Nat. Genet. lymphoblastoid cells. 33: 422–425. 61. Whitney, A. R., M. Diehn, S. J. Popper, A. A. Alizadeh, J. C. Boldrick, 34. Bolstad, B. M., F. Collin, K. M. Simpson, R. A. Irizarry, and T. P. Speed. 2004. D. A. Relman, and P. O. Brown. 2003. Individuality and variation in gene ex- Experimental design and low-level analysis of microarray data. Int. Rev. Neuro- pression patterns in human blood. Proc. Natl. Acad. Sci. USA 100: 1896–1901. biol. 60: 25–58. ϩ 62. Ogg, G. S. 2003. T-cell immunotherapy of allergic disease: the role of CD8 T 35. Gentleman, R. C., V. J. Carey, D. M. Bates, B. Bolstad, M. Dettling, S. Dudoit, cells. Curr. Opin. Allergy Clin. Immunol. 3: 475–479. B. Ellis, L. Gautier, Y. Ge, J. Gentry, et al. 2004. Bioconductor: open software 63. Marrack, P., and J. Kappler. 1990. The staphylococcal enterotoxins and their by guest on October 1, 2021 development for computational biology and bioinformatics. Genome Biol. relatives. Science 248: 705–711. 5: R80. 64. Romagnani, S. 1991. Human TH1 and TH2 subsets: doubt no more. Immunol. 36. Bolstad, B. M. 2004. Low level analysis of high-density oligonucleotide array Today 12: 256–257. data: background, normalization and summarization. PhD dissertation, University of California, Berkeley, CA. 65. Wierenga, E. A., M. Snoek, C. de Groot, I. Chre´tien, J. D. Bos, H. M. Jansen, and M. L. Kapsenberg. 1990. Evidence for compartmentalization of functional sub- 37. Irizarry, R. A., B. Hobbs, F. Collin, Y. D. Beazer-Barclay, K. J. Antonellis, ϩ U. Scherf, and T. P. Speed. 2003. Exploration, normalization, and summaries of sets of CD4 T lymphocytes in atopic patients. J. Immunol. 144: 4651–4656. high density oligonucleotide array probe level data. Biostatistics 4: 249–264. 66. Borish, L., and L. Rosenwasser. 1997. Th1/Th2 lymphocytes: doubt some more. 38. Smyth, G. K. 2004. Linear models and empirical Bayes methods for assessing J. Allergy Clin. Immunol. 99: 161–164. differential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol. 3: 67. Holt, P. G. 2004. The role of genetic and environmental factors in the develop- article 3. ment of T-cell mediated allergic disease in early life. Paediatr. Respir. Rev. 39. Kooperberg, C., A. Aragaki, A. D. Strand, and J. M. Olson. 2005. Significance 5(Suppl. A): S27–S30. testing for small microarray experiments. Stat. Med. 24: 2281–2298. 68. Hansen, G., G. Berry, R. H. DeKruyff and D. T. Umetsu. 1999. Allergen-specific 40. Benjamini, Y., D. Drai, G. Elmer, N. Kafkafi, and I. Golani. 2001. Controlling the Th1 cells fail to counterbalance Th2 cell-induced airway hyperreactivity but false discovery rate in behavior genetics research. Behav. Brain Res. 125: cause severe airway inflammation. J. Clin. Invest. 103: 175–183. 279–284. 69. Castro, M., D. D. Chaplin, M. J. Walter, and M. J. Holtzman. 2000. Could asthma 41. Storey, J. D., and R. Tibshirani. 2003. Statistical significance for genomewide be worsened by stimulating the T-helper type 1 immune response? Am J. Respir. studies. Proc. Natl. Acad. Sci. USA 100: 9440–9445. Cell Mol. Biol. 22: 143–146. 42. Eisen, M. B., P. T. Spellman, P. O. Brown, and D. Botstein. 1998. Cluster anal- 70. Platts-Mills, T., J. Vaughan, S. Squillace, J. Woodfolk, and R. Sporik. 2001. ysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. USA Sensitisation, asthma, and a modified Th2 response in children exposed to cat 95: 14863–14868. allergen: a population-based cross-sectional study. Lancet 357: 752–756. ϩ 43. Heyer, L. J., S. Kruglyak, and S. Yooseph. 1999. Exploring expression data: 71. Otten, G. R., and R. N. Germain. 1991. Split anergy in CD8 T cell: receptor- identification and analysis of coexpressed genes. Genome Res. 9: 1106–1115. dependent cytolysis in the absence of interleukin-2 production. Science 251: 44. Tai, Y. C., and T. P. Speed. 2004. A multivariate empirical Bayes statistic for 1228–1231. replicated microarray time course data: technical report no. 667. Department of 72. Bunikowski, R., M. Mielke, H. Skarabis, U. Herz, R. L. Bergmann, U. Wahn, and Statistics, University of California, Berkeley. H. Renz. 1999. Prevalence and role of serum IgE antibodies to the Staphylococ- 45. Lonnstedt, I., and T. P. Speed. 2002. Replicated microarray data. Statist. Sinica. cus aureus-derived superantigens SEA and SEB in children with atopic derma- 12: 31–46. titis. J. Allergy Clin. Immunol. 103: 119–124. 46. Cui, X., and G. A. Churchill. 2003. Statistical tests for differential expression in 73. Shiomori, T., S.-I. Yoshida, H. Miyamoto, and K. Makishima. 2000. Relationship cDNA microarray experiments. Genome Biol. 4: 210. of nasal carriage of Staphylococcus aureus to pathogenesis of perennial allergic 47. Saeed, A. I., V. Sharov, J. White, J. Li, W. Liang, N. Bhagabati, J. Braisted, rhinitis. J. Allergy Clin. Immunol. 105: 449–454. M. Klapa, T. Currier, M. Thiagarajan, et al. 2003. TM4: a free, open-source 74. Cheyette, B. N., J. S. Waxman, J. R. Miller, K. Takemaru, L. C. Sheldahl, system for microarray data management and analysis. BioTechniques 34: N. Khlebtsova, E. P. Fox, T. Earnest, and R. T. Moon. 2002. Dapper, a 374–378. dishevelled-associated antagonist of ␤-catenin and JNK signaling, is required for 48. Upham, J. W., B. J. Holt, M. J. Baron-Hay, A. Yabuhara, B. J. Hales, notochord formation. Dev. Cell 2: 449–461. W. R. Thomas, R. K. Loh, P. T. O’Keeffe, L. Palmer, P. N. Le Souef, et al. 1995. 75. Xu, Y., D. Banerjee, J. Huelsken, W. Birchmeier, and J. M. Sen. 2003. Deletion Inhalant allergen-specific T-cell reactivity is detectable in close to 100% of atopic of ␤-catenin impairs T cell development. Nat. Immunol. 4: 1177–1182. and normal individuals: covert responses are unmasked by serum-free medium. 76. Dorfman, D. M., H. A. Greisman, and A. Shahsafaeri. 2003. Loss of expression Clin. Exp. Allergy 25: 634–642. of the WNT/␤-catenin-signaling pathway transcription factors lymphoid en- 49. Liu, K., Y. F. Li, V. Prabhu, L. Young, K. G. Becker, P. J. Munson, and hancer factor-1 (LEF-1) and T cell factor-1 (TCF-1) in a subset of peripheral T N. P. Weng. 2001. Augmentation in expression of activation-induced genes dif- cell lymphomas. Am. J. Pathol. 162: 1539–1544. The Journal of Immunology 4777

77. Hogan, P. G., L. Chen, J. Nardone, and A. Rao. 2003. Transcriptional regulation 100. Kagi, D., B. Ledermann, K. Burki, P. Seiler, B. Odermatt, K. J. Olsen, by calcium, calcineurin, and NFAT. Genes Dev. 17: 2205–2232. E. R. Podack, R. M. Zinkernagel, and H. Hengartner. 1994. Cytotoxicity me- 78. Dong, C., D. D. Yang, M. Wysk, A. J. Whitmarsh, R. J. Davis, and R. A. Flavell. diated by T cells and natural killer cells is greatly impaired in perforin-deficient 1998. Defective T cell differentiation in the absence of Jnk1. Science 282: mice. Nature 369: 31–37. 2092–2095. 101. Ju, S. T., H. Cui, D. J. Panka, R. Ettinger, and A. Marshak-Rothstein. 1994. ϩ 79. Millan, J., R. Puertollano, L. Fan, C. Rancano, and M. A. Alonso. 1997. The Participation of target Fas protein in apoptosis pathway induced by CD4 Th1 ϩ MAL proteolipid is a component of the detergent-insoluble membrane subdo- and CD8 cytotoxic T cells. Proc. Natl. Acad. Sci. USA 91: 4185–4189. mains of human T-lymphocytes. Biochem. J. 321: 247–252. 102. Yasukawa, M., H. Ohminami, J. Arai, Y. Kasahara, Y. Ishida, and S. Fujita. 80. Frank, M. 2000. MAL, a proteolipid in glycosphingolipid enriched domains: 2000. Granule exocytosis, and not the fas/fas ligand system, is the main pathway ϩ ϩ functional implications in myelin and beyond. Prog. Neurobiol. 60: 531–544. of cytotoxicity mediated by alloantigen-specific CD4 as well as CD8 cyto- 81. Riley, J. L., M. Mao, S. Kobayashi, M. Biery, J. Burchard, G. Cavet, toxic T lymphocytes in humans. Blood 95: 2352–2355. B. P. Gregson, C. H. June, and P. S. Linsley. 2002. Modulation of TCR-induced 103. Appay, V., J. J. Zaunders, L. Papagno, J. Sutton, A. Jaramillo, A. Waters, transcriptional profiles by ligation of CD28, ICOS, and CTLA-4 receptors. Proc. P. Easterbrook, P. Grey, D. Smith, A. J. McMichael, et al. 2002. Characteriza- ϩ Natl. Acad. Sci. USA 99: 11790–11795. tion of CD4 CTLs ex vivo. J. Immunol. 168: 5954–5958. 82. Nurieva, R. I., J. Duong, H. Kishikawa, U. Dianzani, J. M. Rojo, I. Ho, 104. Yawalkar, N., S. Schmid, L. R. Braathen, and W. J. Pichler. 2001. Perforin and R. A. Flavell, and C. Dong. 2003. Transcriptional regulation of th2 differentiation granzyme B may contribute to skin inflammation in atopic dermatitis and pso- by inducible costimulator. Immunity 18: 801–811. riasis. Br. J. Dermatol. 144: 1133–1139. 83. Martin-Belmonte, F., R. Puertollano, J. Millan, and M. A. Alonso. 2000. The 105. Kuroda, T. S., M. Fukuda, H. Ariga, and K. Mikoshiba. 2002. The Slp homol- MAL proteolipid is necessary for the overall apical delivery of membrane pro- ogy domain of synaptotagmin-like proteins 1–4 and Slac2 functions as a novel teins in the polarized epithelial Madin-Darby canine kidney and Fischer rat thy- Rab27A binding domain. J. Biol. Chem. 277: 9212–9218. roid cell lines. Mol. Biol. Cell 11: 2033–2045. 106. Elliott, J., and J. A. Johnston. 2004. SOCS: role in inflammation, allergy and 84. Alonso, M. A., and J. Millan. 2001. The role of lipid rafts in signallin and mem- homeostasis. Trends Immunol. 25: 434–440. brane trafficking in T lymphocytes. J. Cell Sci. 114: 3957–3965. 107. Matsumoto, A., Y. Seki, M. Kubo, S. Ohtsuka, A. Suzuki, I. Hayashi, K. Tsuji, 85. Millan, J., M. C. Montoya, D. Sancho, F. Sanchez-Madrid, and M. A. Alonso. T. Nakahata, M. Okabe, S. Yamada, and A. Yoshimura. 1999. Suppression of 2002. Lipid rafts mediate biosynthetic transport to the T lymphocyte uropod STAT5 functions in liver, mammary glands, and T cells in cytokine-inducible SH2-containing protein 1 transgenic mice. Mol. Cell. Biol. 19: 6396–6407.

subdomain and are necessary for uropod integrity and function. Blood 99: Downloaded from 978–984. 108. Fort, M. M., J. Cheung, D. Yen, J. Li, S. M. Zurawski, S. Lo, S. Menon, 86. Kupfer, A., and H. Kupfer. 2003. Imaging immune cell interactions and func- T. Clifford, B. Hunte, R. Lesley, et al. 2001. IL-25 induces IL-4, IL-5, and IL-13 tions: SMACs and the immunological synapse. Semin. Immunol. 15: 295–300. and Th2-associated pathologies in vivo. Immunity 15: 985–995. 109. Jackson, D. E. 2003. The unfolding tale of PECAM-1. FEBS Lett. 540: 7–14. 87. Ishii, T., K. Ohnuma, A. Murakami, N. Takasawa, S. Kobayashi, N. H. Dang, 110. Wilkinson, R., A. B. Lyons, D. Roberts, M. X. Wong, P. A. Bartley, and S. F. Schlossman, and C. Morimoto. 2001. CD26-mediated signaling for T cell D. E. Jackson. 2002. Platelet endothelial cell adhesion molecule-1 (PECAM-1/ activation occurs in lipid rafts through its association with CD45RO. Proc. Natl. CD31) acts as a regulator of B-cell development, B-cell antigen receptor (BCR)- Acad. Sci. USA 98: 12138–12143. mediated activation, and autoimmune disease. Blood 100: 184–193. 88. Pacheco, R., J. M. Martinez-Navio, M. Lejeune, N. Climent, H. Oliva, 111. Graesser, D., A. Solowiej, M. Bruckner, E. Osterweil, A. Juedes, S. Davis, http://www.jimmunol.org/ J. M. Gatell, T. Gallart, J. Mallol, C. Lluis, and R. Franco. 2005. CD26, adenosine N. H. Ruddle, B. Engelhardt, and J. A. Madri. 2002. Altered vascular perme- deaminase, and adenosine receptors mediate costimulatory signals in the immu- ability and early onset of experimental autoimmune encephalomyelitis in PE- nological synapse. Proc. Natl. Acad. Sci. USA 102: 9583–9588. CAM-1-deﬁcient mice. J. Clin. Invest. 109: 383–392. 89. Kruschinski, C., T. Skripuletz, S. Bedoui, T. Tschernig, R. Pabst, C. Nassenstein, 112. Newton-Nash, D. K., and P. J. Newman. 1999. A new role for platelet-endo- A. Braun, and S. von Horsten. 2005. CD26 (dipeptidyl-peptidase IV)-dependent thelial cell adhesion molecule-1 (CD31): inhibition of TCR-mediated signal recruitment of T cells in a rat asthma model. Clin. Exp. Immunol. 139: 17–24. transduction. J. Immunol. 163: 682–688. 90. Ohnuma, K., T. Yamochi, O. Hosono, and C. Morimoto. 2005. CD26 T cells in 113. Sloan-Lancaster, J., T. H. Steinberg, and P. M. Allen. 1997. Selective loss of the the pathogenesis of asthma. Clin. Exp. Immunol. 139: 13–16. calcium ion signaling pathway in T cells maturing toward a T helper 2 pheno- 91. Cristillo, A. D., L. Nie, M. J. Macri, and B. E. Bierer. 2003. Cloning and char- type. J. Immunol. 159: 1160–1168. acterization of N4WBP5A, an inducible, cyclosporine-sensitive, Nedd4-binding 114. Leitenberg, D., and K. Bottomly. 1999. Regulation of naive T cell differentiation protein in human T lymphocytes. J. Biol. Chem. 278: 34587–34597. by varying the potency of TCR signal transduction. Semin. Immunol. 11:

92. Matsuda, A., Y. Suzuki, G. Honda, S. Muramatsu, O. Matsuzaki, Y. Nagano, 283–292. by guest on October 1, 2021 T. Doi, K. Shimotohno, T. Harada, E. Nishida, et al. 2003. Large-scale identifi- 115. Nanda, A., P. Buckhaults, S. Seaman, N. Agrawal, P. Boutin, S. Shankara, ␬ cation and characterization of human genes that activate NF- B and MAPK M. Nacht, B. Teicher, J. Stampfl, S. Singh, et al. 2004. Identification of a bind- signaling pathways. Oncogene 22: 3307–3318. ing partner for the endothelial cell surface proteins TEM7 and TEM7R. Cancer 93. Heissmeyer, V., F. Macian, S. H. Im, R. Varma, S. Feske, K. Venuprasad, H. Gu, Res. 64: 8507–8511. Y. C. Liu, M. L. Dustin, and A. Rao. 2004. Calcineurin imposes T cell unre- 116. Fuller, C. L., V. L. Braciale, and L. E. Samelson. 2003. All roads lead to actin: sponsiveness through targeted pteolysis of signaling proteins. Nat. Immunol. 5: the intimate relationship between TCR signaling and the cytoskeleton. Immunol. 255–265. Rev. 191: 220–236. 94. Shearwin-Whyatt, L. M., D. L. Brown, F. G. Wylie, J. L. Stow, and S. Kumar. 117. Hudmon, A., and H. Schulman. 2002. Neuronal CA2ϩ/calmodulin-dependent 2004. N4WBP5A (Ndfip2), a Nedd4-interacting protein, localizes to multivesicu- protein kinase II: the role of structure and autoregulation in cellular function. lar bodies and the Golgi, and has a potential role in protein trafficking. J. Cell Sci. Annu. Rev. Biochem. 71: 473–510. 117: 3679–3689. 118. Bui, J. D., S. Calbo, K. Hayden-Martinez, L. P. Kane, P. Gardner, and 95. Barnes, P. J. 1997. Nuclear factor-␬B. Int J. Biochem. Cell Biol. 29: 867–870. S. M. Hedrick. 2000. A role for CaMKII in T cell memory. Cell 100: 457–467. 96. Barral, D. C., J. S. Ramalho, R. Anders, A. N. Hume, H. J. Knapton, 119. Lin, M. Y., T. Zal, I. L. Ch’en, N. R. Gascoigne, and S. M. Hedrick. 2005. A T. Tolmachova, L. M. Collinson, D. Goulding, K. S. Authi, and M. C. Seabra. pivotal role for the multifunctional calcium/calmodulin-dependent protein ki- 2002. Functional redundancy of Rab27 proteins and the pathogenesis of Griscelli nase II in T cells: from activation to unresponsiveness. J. Immunol. 174: syndrome. J. Clin. Invest. 110: 247–257. 5583–5592. 97. Trambas, C. M., and G. M. Griffiths. 2003. Delivering the kiss of death. Nat. 120. Means, A. R. 2000. Regulatory cascades involving calmodulin-dependent pro- Immunol. 4: 399–403. tein kinases. Mol. Endocrinol. 14: 4–13. 98. Stinchcombe, J., G. Bossi, and G. M. Griffiths. 2004. Linking albinism and im- 121. Hook, S. S., and A. R. Means. 2001. Ca2ϩ/CaM-dependent kinases: from ac- munity: the secrets of secretory lysosomes. Science 305: 55–59. tivation to function. Annu. Rev. Pharmacol. Toxicol. 41: 471–505. 99. Ramsdell, F., M. S. Seaman, R. E. Miller, T. W. Tough, M. R. Alderson, and 122. Sheldahl, L. C., D. C. Slusarski, P. Pandur, J. R. Miller, M. Kuhl, and D. H. Lynch. 1994. gld/gld mice are unable to express a functional ligand for Fas. R. T. Moon. 2003. Dishevelled activates Ca2ϩ flux, PKC, and CamKII in ver- Eur. J. Immunol. 24: 928–933. tebrate embryos. J. Cell Biol. 161: 769–777.