Skip Navigation


Human Molecular Genetics Advance Access originally published online on April 6, 2005
Human Molecular Genetics 2005 14(10):1305-1314; doi:10.1093/hmg/ddi141
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Material
Right arrow All Versions of this Article:
14/10/1305    most recent
ddi141v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (10)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Maas, K.
Right arrow Articles by Aune, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Maas, K.
Right arrow Articles by Aune, T.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2005. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oupjournals.org

Shared gene expression profiles in individuals with autoimmune disease and unaffected first-degree relatives of individuals with autoimmune disease

Kevin Maas1,2, Heidi Chen3, Yu Shyr3, Nancy J. Olsen1,2 and Thomas Aune1,2,*

1Division of Rheumatology, Department of Medicine, 2Department of Microbiology and Immunology and 3Department of Biostatistics, Vanderbilt University School of Medicine, Nashville, TN 37232, USA

* To whom correspondence should be addressed at: MCN T3219, Vanderbilt University Medical Center, 1161 21st Avenue South, Nashville, TN 37232, USA. Tel: +1 6153437353; Fax: +1 6153226248; Email: thomas.aune{at}vanderbilt.edu

Received January 11, 2005; Accepted March 22, 2005


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Patients with autoimmune disorders exhibit highly reproducible gene expression profiles in their peripheral blood mononuclear cells. These signatures may result from chronic inflammation, other disease manifestations, or may reflect family resemblance. To test the latter hypothesis, we determined gene expression profiles in unaffected first-degree relatives of individuals with autoimmune disease. Gene expression profiles in unaffected first-degree relatives resembled the profiles found in individuals with autoimmune diseases. A high percentage of differentially expressed genes in unaffected first-degree relatives were previously identified as autoimmune signature genes. Examination of the linear regression relationship of gene transcript levels between parent–offspring pairs revealed that autoimmune signature genes display high levels of family resemblance. Taken together, these results support the hypothesis that these variations in gene transcript levels are associated with family resemblance rather than clinical manifestations of disease.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Autoimmune diseases affect ~5% of the human population (1Go). Although there is considerable heterogeneity among these disorders, their manifestations are believed to arise from immune-mediated attack against self-antigens. In some cases, the immune response targets specific organs and tissues. For example, in multiple sclerosis (MS), the immune system targets the protective myelin sheaths surrounding nerves (2Go). In type I diabetes (IDDM), the immune system destroys the insulin-producing ß cells but spares the other cells within islets of Langerhans in the pancreas (3Go). Clinical manifestations of autoimmune disease may also lead to more systemic effects. The major symptoms of rheumatoid arthritis (RA) arise through immune-mediated destruction of peripheral joints; however, these features are typically accompanied by systemic complications such as rheumatoid nodules and vasculitis (4Go). Many of the pathologic aspects of systemic lupus erythematosus (SLE) arise from vascular deposition of immune complexes, with resulting immune-mediated injury occurring at numerous sites (5Go).

Despite their clinical heterogeneity, recent studies examining gene expression profiles in peripheral blood mononuclear cells (PBMC) of individuals with autoimmune disorders reveal common features that are either shared within a disease group or among disease groups. For example, two independent groups have reported that SLE patients with aggressive disease exhibit similar gene expression profiles (6Go,7Go). Similarly, we have identified gene expression signatures in RA that vary as a function of disease duration (8Go). We have also found a common gene expression signature that is present in PBMC of all patients with RA, SLE, IDDM and MS (9Go). This expression signature is composed of both under- and over-expressed genes and perfectly discriminates between patients with autoimmune disorders and normal controls and appears to be independent of disease severity, duration, symptoms and treatment.

Substantial evidence points toward a heritable component for many autoimmune disorders. Epidemiological data suggest that there are underlying genetic contributions to disease susceptibility. The relative risk of autoimmune disease in siblings ({lambda}S) and monozygotic twins ({lambda}MZ) when compared with the general population is elevated in a number of autoimmune diseases. For example in SLE, the {lambda}MZ has been estimated to be around 250 and {lambda}S is around 20, risks that are substantially greater than those for the general population (10Go,11Go). Similar results have been reported for RA, MS, ankylosing spondylitis and IDDM (10Go,12Go–14Go). Further epidemiological support for an underlying genetic contribution to autoimmunity comes from studies in autoimmune prone families. In these families, multiple individuals suffer from a variety of autoimmune disorders suggesting a common underlying genetic predisposition to autoimmunity (15Go–17Go).

Genetic linkage studies provide additional evidence for a genetic contribution to disease susceptibility. A large number of linkage studies have been performed in MS (18Go–20Go), IDDM (21Go–23Go), SLE (24Go–27Go) and RA (28Go,29Go). These studies have revealed that the strongest linkage for susceptibility to autoimmunity resides in the HLA locus. Additional non-MHC loci have also been identified. A substantial number of these non-MHC susceptibility loci are shared among different autoimmune diseases suggesting that there are at least two types of autoimmune susceptibility loci; those that are shared among autoimmune diseases and those that are specific for an individual autoimmune disease (30Go,31Go).

Recent studies examining the genetics of gene expression in a range of eukaryotic organisms have revealed that differences in gene transcript levels can arise through inheritance (32Go,33Go). In this sense, unique gene expression profiles in autoimmune patients may also arise from genetic factors associated with susceptibility to autoimmunity. In a preliminary attempt to address this question, we examined gene expression profiles in unaffected first-degree relatives of patients with RA or SLE. Unaffected first-degree relatives of individuals with autoimmune disease exhibited gene expression patterns distinct from normal control individuals but similar to autoimmune patients. Many of the differentially expressed genes (DEGs) present in unaffected family members were the same genes found in patients with autoimmune disease. Furthermore, studies examining correlation of gene expression levels among parent–offspring pairs revealed that the expression levels of ‘autoimmune signature’ genes displayed higher levels of family resemblance than did non-autoimmune signature genes.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Similarities in gene expression profiles between individuals with autoimmune disease and unaffected first-degree relatives of individuals with autoimmune disease
We wanted to determine if gene expression profiles from unaffected first-degree relatives of autoimmune patients were more closely related to profiles of control individuals or autoimmune patients. We used standard microarray analysis to collect gene expression profiles from the PBMC of control individuals, autoimmune patients and unaffected first-degree relatives of individuals with autoimmune disease. Each of the six unaffected first-degree relatives of autoimmune individuals was unrelated to any of the eight autoimmune individuals or six healthy control individuals in this preliminary analysis. Initially, a hierarchical clustering algorithm was used to compare unaffected family members to control individuals by overall relatedness of gene expression profiles (using all 4000+genes). As illustrated by a representative node of genes that reflect the hierarchical clustering of all over 4000 genes, this analysis grouped all control individuals into a single branch and all unaffected family members into separate branches (Fig. 1A).



View larger version (29K):
[in this window]
[in a new window]
  		Figure 1. Hierarchical clustering of total gene expression profiles. A hierarchical clustering algorithm was employed to group individuals on the basis of similarities in overall gene expression profiles. Representative nodes are showed to provide sample genes that discriminate between branches. Hybridization intensities for individual genes across the profiles are represented as a range from black (no expression) to red (high expression level). Genes are represented by their gene symbols. Clustering and nodes for. (A) control individuals (Cont) and unaffected first-degree relatives (family number- F no.), (B) unaffected first-degree relatives and individuals with autoimmune disease (RA, SLE) and (C) control individuals, unaffected first-degree relatives and autoimmune individuals.

 
As unaffected first-degree relatives of autoimmune patients displayed expression patterns distinct from control individuals, we wanted to determine whether the profiles of unaffected family members were also distinct from unrelated patients with autoimmune disease. The same hierarchical clustering algorithm was employed to group individuals and genes by overall similarities in expression patterns. As with the previous clustering, a representative node that reflects the clustering of the total genes was selected. In contrast to the earlier results, the program did not group unaffected family members and individuals with autoimmune disease into separate branches (Fig. 1B). We were unable to find gene nodes that accurately discriminated between autoimmune patients and unaffected family members using unsupervised hierarchical clustering.

Next, we compared individuals in all three groups using the same clustering algorithm. All control individuals were segregated into a single branch. All unaffected family members and autoimmune individuals were segregated into the remaining branches (Fig. 1C). Unaffected family members and autoimmune individuals did not segregate into separate branches. Rather, each branch with autoimmune individuals also contained unaffected first-degree relatives. As described previously, we chose a node of genes that reflected the differences in the clustering profiles (Fig. 1C). Of note, many of the genes present in this node have been previously defined as part of an under-expressed autoimmune gene expression signature (9Go).

Autoimmune signature genes are differentially expressed in unaffected first-degree relatives
To determine whether DEGs from unaffected family members overlapped with those from autoimmune patients, we performed k-means clustering analysis. We compared the average gene expression levels among the control, unaffected first-degree relative and autoimmune patient groups (Fig. 2). We used the same profiles for the control individuals examined in Figure 1 (n=6, control). As an additional control, we included gene expression profiles from the same control individuals after immunization with influenza vaccine (n=6, MidIMM) (9Go). The average gene expression levels of the control groups were compared to the average gene expression profiles of the previously examined autoimmune patients (n=8) (RA, n=4 and SLE, n=4; autoimmune) and unaffected first-degree relatives (n=6, unaffected FAM).



View larger version (31K):
[in this window]
[in a new window]
  		Figure 2. Percentage of DEGs in unaffected first-degree relatives that are autoimmune signature genes. Using an approach similar to that outlined in Figure 1, we determined the percent contribution of autoimmune genes to the most differentially expressed clusters identified in unaffected family members compared to random controls. Gene expression data are presented as the natural alogarithm of ratio (control/control, immune/control, unaffected first-degree relative/control and autoimmune/control). (A) Percent of autoimmune signature genes found in DEGs (P<0.01, Chen test) for unaffected family members. (B) Percent of autoimmune signature genes found in DEGs (P<0.001, Chen test).

 
As a group, we found that the most DEG clusters (Chen test, P<0.01) in autoimmune patients were differentially expressed to the same extent in unaffected family members (Fig. 2). The over-expressed autoimmune/unaffected family cluster contained a total of 127 genes and the under-expressed cluster contained 74 genes. Approximately 52% (66/127) of the total over-expressed and 55% (41/74) of the total under-expressed genes were autoimmune genes, revealing that a large percentage of the most DEGs were autoimmune signature genes (Fig. 2A). When the stringency for significance was increased to 99.9% (Chen test, P<0.001), >90% of the remaining DEGs were previously identified autoimmune signature genes (9Go) (Fig. 2).

We also performed analyses to determine the contribution of the autoimmune signature to the DEGs in individual profiles. To do this, we first used the k-means clustering algorithm to group genes with similar expression patterns among control individuals, immune response individuals and an individual unaffected family member (Fig. 3A). The ratio of gene expression (experimental/control, LN) was determined for each of the over 4000 genes for the control group, the immune response group and each unaffected family member or autoimmune individual (RA and SLE). Differences in gene expression among the groups that did not achieve statistical significance (Chen test, P<0.01) were excluded from further analysis (Fig. 3B). Both the major over- and under-expressed gene clusters were isolated (Fig. 3Ca and b). We further restricted these DEG clusters to the autoimmune signature genes (Fig. 3Da and b) identified in the previous studies.



View larger version (24K):
[in this window]
[in a new window]
  		Figure 3. Schematic of analytic method to determine overlap between DEGs and autoimmune signature genes. (A) The average ratio of gene expression relative to the control group was calculated for each group or individual and expressed on a natural alogarithm scale. (B) Genes that were not significantly differentially expressed (P<0.01, Chen test) were removed from the data set. Remaining DEGs were grouped into 10 clusters using a k-means clustering algorithm. (C) Over- and under-expressed clusters were isolated for further analysis. (D) Isolated clusters were further restricted to previously identified autoimmune signature genes. The genes that remained were used to calculate the percent of the total autoimmune signature gene present in the clusters (Table 1).

 
Using results from this type of analysis, we determined the percentage of autoimmune genes present in the DEGs of unaffected family members. For example, after clustering and statistical treatment, the unaffected first-degree relative F1-M had two major DEG clusters: an under-expressed cluster of 246 genes and an over-expressed cluster of 245 genes. We previously defined the autoimmune signature genes (RA, SLE, IDDM and MS) as an under-expressed cluster of 113 genes and an over-expressed cluster of 96 genes from a total of more than 4000 genes on the microarray (9Go). When we restricted the DEG clusters in individual F1-M to autoimmune genes, the under-expressed cluster retained 75 genes and the over-expressed cluster retained 86 genes. Therefore, after clustering, statistical treatment and restriction of non-autoimmune genes, we found that individual F1-M contained 66% (75/113) of the under-expressed and 90% (86/96) of the over-expressed autoimmune signature genes in their total DEG profile.

Results have been summarized in Table 1 for control individuals, autoimmune patients and unaffected family members. DEGs in control individuals contained only a small percentage of genes from the autoimmune signature. Conversely, individuals diagnosed with autoimmune disease contained significantly higher percentages of both the over- and under-expressed autoimmune signature genes (47%±44, P<0.02 and 54%±15, P<10–4, respectively). Unaffected first-degree relatives also contained significantly higher numbers of autoimmune genes represented in the under-expressed cluster when compared with control individuals (44%±24, P<0.01). The over-expressed cluster in family members had a higher percentage of autoimmune genes as well, but these differences did not achieve statistical significance.


View this table:
[in this window]
[in a new window]
  		Table 1. Conservation of autoimmune gene expression profiles in unaffected first-degree relatives
 
In addition to the total autoimmune signature, we previously identified a core set of 35 ESTs that were consistently under-expressed in autoimmune individuals (9Go). We utilized the same method as outlined previously (Fig. 3) to determine the representation of this core in unaffected family members (Table 1). Of these 35 ESTs, only 29 ESTs map to known genes using current genome databases. Therefore, we elected to restrict this analysis to those 29 ESTs that have been aligned with known genes in the human genome database. Control individuals had few of these core genes present among their most DEGs (7%±10). DEGs for both autoimmune patients and unaffected first-degree relatives were highly represented (90%±12 and 76%±25, respectively) and were statistically significant when compared with controls (P<10–7 and P<10–3, respectively). There was no statistically significant difference between the percent of these core genes present in autoimmune patients and unaffected family members. Overall, these data demonstrate that genes differentially expressed in individuals with autoimmune disease are also differentially expressed in unaffected first-degree relatives.

Given the fact the core genes were so highly represented in unaffected family members, we wanted to see how the individual profiles would cluster if we restricted the microarray data set down to the 29 gene autoimmune core. Profiles from control individuals sorted with a remarkable degree of precision to a single branch (Fig. 4), whereas both autoimmune and unaffected family member profiles segregated into the other branches. This analysis further supports the hypothesis that gene expression profiles of individuals with autoimmune disease and unaffected first-degree family members are highly similar to each other and that both groups are highly distinct from control individuals.



View larger version (52K):
[in this window]
[in a new window]
  		Figure 4. Hierarchical clustering using core autoimmune genes. Microarray data were restricted to 29 previously identified core autoimmune genes (9Go). Profiles for control individuals, unaffected family members and autoimmune individuals were subjected to hierarchical clustering. Hybridization intensities are represented as a range from black (no expression) to red (high expression).

 
Unaffected family members do not display signs of autoimmune disease, however, they contain a high number of autoimmune DEGs. Therefore, we wanted to determine whether there were DEGs that could discriminate unaffected first-degree relatives from individuals with autoimmune disease. Using a variety of approaches, we were unable to identify combinations of DEGs that successfully discriminated between all unaffected family members and all individuals with autoimmune disease (data not shown). This is not meant to imply that other computational methods would not be able to identify combinations of DEGs that successfully discriminate between these two groups.

Autoimmune signature genes display high levels of family resemblance
Our findings indicate that the gene expression profile of an unaffected first-degree relative is more similar to an unrelated individual with autoimmune disease than an individual from our control group. This argues that transcript levels of a proportion of the autoimmune signature genes may be determined by family resemblance rather than by disease activity. We wanted to use a combination of computational techniques and statistical analyses to determine the degree of gene expression resemblance among family members. We reasoned that if expression levels of autoimmune genes demonstrated a high family resemblance, they should be similar between a parent and offspring but may vary among parent–offspring pairs.

To perform this analysis, we determined the correlation coefficients of expression levels for individual genes in eight parent–offspring pairings as an indication of the level of family resemblance for autoimmune signature genes and non-autoimmune signature genes present on the microarray data set. In five of these pairings, one individual suffered from a previously examined autoimmune disorder (RA, MS, IDDM or SLE), whereas the other individual was unaffected. In the three remaining pairs, neither parent nor offspring had one of the four major autoimmune diseases we have previously studied. Because of small sample size, we utilized the non-parametric Spearman rank correlation coefficient to determine the degree of parent–offspring relatedness for a given gene. Genes and expression data were divided into one of three major categories: over-expressed autoimmune genes (94 genes), under-expressed autoimmune genes (111 genes) and non-autoimmune genes (3924 genes). The average Spearman correlation coefficient for each category was calculated. Examples of the autoimmune signature genes displaying the highest average levels of correlation are provided in Tables 2 and 3.


View this table:
[in this window]
[in a new window]
  		Table 2. Most correlated under-expressed autoimmune signature genes
 

View this table:
[in this window]
[in a new window]
  		Table 3. Most correlated over-expressed autoimmune signature genes
 
The results of the Kruskal–Wallis test suggested that the correlations among the three gene categories were not the same (P<0.0001). The pairwise comparisons were completed using a permutation test (34Go), a non-parametric test without the assumption of a normal distribution. The permutation test was applied to the t-statistics of the rank score of Spearman correlation to determine which of the categories (over-, under- or non-autoimmune genes) were significantly different. SAS/STAT MULTTEST, which controls the family-wise error rate for multiple hypothesis testing was used for the analysis of the pairwise comparisons. The results revealed that, on average, expression levels for the both over- and under-expressed autoimmune genes were more highly correlated between parent and offspring than to the non-autoimmune genes (Fig. 5A). These differences were highly significant for both over-expressed autoimmune genes versus non-autoimmune genes and under-expressed autoimmune genes versus non-autoimmune genes (P<0.0001, permutation t-test). Comparison of the average correlation for the over- and under-expressed autoimmune genes was not significant (P=0.83, permutation t-test).



View larger version (10K):
[in this window]
[in a new window]
  		Figure 5. Comparison of parent–offspring average gene expression correlation coefficients. The average Spearman correlation coefficients from eight parent–offspring pairs for the following categories: non- (non-autoimmune), over- (over-autoimmune) and under-expressed (under-autoimmune) autoimmune signature genes. Significance was established using a permutation t-test.

 
In addition to examining the average degree of correlation for the autoimmune and non-autoimmune gene expression signature groups, we determined whether genes displaying significant parent–offspring correlation were distributed differently among autoimmune and non-autoimmune signature genes. To perform this analysis, we determined the number of genes in the non-autoimmune, under-expressed and over-expressed autoimmune signature groups, which showed evidence of significant family resemblance (Table 4). Transcript levels of 33% of the non-autoimmune genes showed evidence of significant parent–offspring resemblance, Rs>0.7, P≤0.05. We used a similar approach to compare transcript levels of autoimmune signature genes. Of the over-and under-autoimmune signature genes, 67 and 62% of the genes, respectively, demonstrated significant correlation Rs>0.7, P≤0.05 in the family comparison. An odds ratio analysis of the correlation data was used to further determine whether autoimmune genes were more likely to be significantly correlated than the non-autoimmune genes (Table 4). Both over-expressed and under-expressed autoimmune genes were more likely to be significantly correlated (Rs>0.7) than non-autoimmue genes in the pairwise family analysis because the over-expressed genes have 4.27 times (P<0.0001) higher odds than the odds of non-autoimmune genes. Similarly, the under-expressed genes have 3.45 times (P<0.0001) higher odds compared with the non-autoimmune genes.


View this table:
[in this window]
[in a new window]
  		Table 4. Comparison of correlation between gene groups
 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 REFERENCES
 
We have used several different approaches to compare gene expression profiles among control individuals, unaffected family members and autoimmune patients. Hierarchical clustering of gene expression profiles shows that unaffected family members more closely resemble unrelated autoimmune patients than unrelated control individuals. Examination of the DEGs in unaffected family members reveals that a significant portion of these genes are previously defined autoimmune signature genes. This finding is confirmed by examining the contribution of autoimmune signature genes to the DEGs in individual unaffected family members. The finding that unaffected first-degree relatives carry a significant portion of the autoimmune signature suggests that this expression pattern may reflect a family trait. Although the sample size in this study is not large enough to accurately estimate heritability of gene expression levels, we present the regression relationship of gene expression levels among first-degree relatives as an initial assessment of familial resemblance. We consider a statistically significant slope of the regression line to be a preliminary evidence that first-degree relatives have more similar gene expression levels than would be expected for unrelated subjects.

To begin examining the heritability of gene expression in the context of human autoimmune disease, we derived correlation coefficients for the gene expression levels from parent–offspring pairs of unaffected and autoimmune families. This pairwise analysis revealed that gene expression levels of ~33% of the genes present in our microarray data set demonstrate significant familial correlation, a percentage similar to those reported examining gene expression in extended families (33Go). We also find that expression levels of previously identified autoimmune signature genes demonstrate significantly higher levels of familial correlation than the remaining genes on the microarrays. This finding provides further support for the notion that the autoimmune gene expression pattern reflects a heritable trait or traits.

The incidence of autoimmune diseases in families with an affected individual is increased compared with the general population. In particular, previous examination of the inheritance of an autoimmune phenotype in families with at least one affected individual shows interesting parallels with our microarray results (35Go). To perform this analysis, researchers extended the definition of an autoimmune phenotype to include either the presence of an autoimmune disease (SLE) or the presence of serological markers used to diagnose SLE. These include anti-nuclear antibodies, anti-single stranded DNA antibodies and a false positive blood test for syphilis. Analysis of these families using this extended autoimmune phenotype suggests that the autoimmune phenotype is inherited as an autosomal dominant trait with greater penetrance in females than in males. As gene expression levels may be even more closely linked to the biochemical processes that give rise to them, analysis of gene transcript levels as a genetic trait may provide even greater resolving power than assessment of sub-clinical traits such as these (33Go).

In model systems, several recent studies have addressed the question of the inheritance of gene transcript levels. Through the combined approach of microarray profiling and genetic linkage analysis, researchers have shown that gene expression patterns are heritable under a variety of conditions across a range of species (32Go,33Go,36Go–41Go). In particular, several recent studies have examined the genetics of gene expression in cell lines derived from several extended human families (36Go–38Go). These studies have universally found that variations in transcript levels of many genes have a heritable component. One power of this approach is that linkage analysis also identifies loci that regulate variation in transcript levels and knowledge of the gene encoding the transcript will undoubtedly stimulate identification of ‘candidate genes’. We hypothesize that this type of approach in autoimmune disease will dramatically enhance the search for candidate genes in the loci identified by linkage analyses.

As screening patients with a range of autoimmune disorders identified this gene expression signature, we believe it reflects a common property of autoimmune disease. In addition to the inheritance of an autoimmune phenotype, more recent family (15Go,42Go) and genetic linkage studies (28Go,30Go) suggest that there are common genetic regions linked to multiple autoimmune disorders. As gene expression profiles can be a product of genotype, we believe that this autoimmune gene expression signature may arise through a heritable trait or traits that may represent a risk factor for developing autoimmune disease.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
REFERENCES
 
Patient populations
The study consisted of the following groups as described in the text:

  1. Six control individuals without active infection or family history of autoimmunity, before and after receiving influenza vaccine;
  2. Four individuals with RA and four individuals with SLE;
  3. Five families were also included in this study:
    1. one unaffected parent, two daughters both with SLE,
    2. both unaffected parents, son with SLE,
    3. one parent with RA, unaffected daughter,
    4. one unaffected parent, daughter with SLE and
    5. grandmother, mother, son and daughter, all unaffected.

All autoimmune patients satisfied established ACR criteria for diagnosis of their respective disease. Human subject studies were approved by the committee for the protection of human subjects of the Vanderbilt University Institutional Review Board.

Sample preparation and RNA isolation
All blood samples were processed within 2 h of collection. In order to reduce variability from handling, a single individual performed all RNA extractions and microarray hybridizations. PBMCs were isolated from heparinized blood by centrifugation on a Histopaque gradient (Sigma, St Louis, MO, USA). Flow cytometry was used to estimate the relative leukocyte distribution in representative samples from control, normal immune response and autoimmune disease groups. All groups were composed of ~75% T lymphocytes, 10% B lymphocytes, 5% monocytes and <1% contaminating neutrophils. Tri-Reagent (Molecular Research Ctr. Inc., Cincinnati, OH, USA) was used to isolate total RNA according to the manufacturer's protocol.

Probe synthesis and microarray hybridization
Total RNA was reversed transcribed with Superscript II reverse transcriptase (Gibco BRL Life Technologies, Rockville, MD, USA) in the presence of 33GoP-dCTP, to yield radio-labeled cDNA probes. The cDNA probes were hybridized to the Research Genetics GF-211 gene filters (Research Genetics, Huntsville, AL, USA) according to the manufacturer's protocol. After overnight incubation, unhybridized probe was successively washed from the filters using 2xSSC buffer/1% SDS (2x20 min) and 0.5xSSC buffer/1% SDS (1x10 min). The filters were exposed to a phosphorimaging screen for 24 h, scanned and digitally imported for computer analysis.

Clustering data analysis of DEGs
Images were digitally imported and processed using Research Genetic's Pathways 3.0 software package. Data were normalized to yield an average intensity of 1.0 for each clone (4329) represented on the cDNA microarray. Hierarchical clustering analysis was performed using Eisen's Cluster and Treeview software. Total microarray gene expression intensities were entered into a tab-delimited database and analyzed using the Cluster software package. To aid in the visualization of the clustered total database, a single representative node of genes was arbitrarily selected that reflected the overall clustering of the analyzed samples. For further clustering analysis, data sets were restricted to previously defined ‘core’ autoimmune genes and subjected to hierarchical clustering as described previously.

Identification of autoimmune genes in DEG clusters
Research Genetics Pathways 3.0 program was used to identify autoimmune genes present among the most DEGs in individual profiles. Microarrays from previously compiled control groups were separated into two reference conditions: control (control individuals) and post-immunization (6–9 days after immunization with flu vaccine). The gene intensities for each control condition group were averaged. Individual expression profiles for control individuals, RA and SLE patients and unaffected first-degree family members were compared against the compiled control conditions. Expression levels were represented graphically by plotting the natural logarithm of the ratio of the gene expression intensity in the experimental group versus the control group for each condition. Genes that did not change significantly as determined by the Chen test over any of the conditions were removed from the database (43Go,44Go). The Chen test is a statistical analysis that determines whether two sampled intensities are different on the basis of a specified confidence interval (99% for all tests, unless otherwise specified). Unlike the t-test, this test is specifically designed for pair-wise comparisons of microarray data. After removal of genes that did not display significant difference, the remaining genes in the data set were clustered using an unsupervised k-means clustering algorithm with 10 centroids. The major under- and over-expressed clusters in the individual profiles were isolated and the total number of genes in the respective clusters determined. The differentially expressed clusters were restricted to previously identified over-expressed, under-expressed and core autoimmune genes using Pathways 3.0 Paths command.

Hybridization intensity correlation coefficient calculations
Spearman rank correlation measured the degree of correspondence between the ranks of the sample observations rather than between the observations themselves. The procedure to calculate the Spearman correlation was as follows:

  1. Rank the observation for each variable.
  2. Obtain the difference in ranks for the paired observation.
  3. Estimate Spearman correlation (Rs) by the equation below:

    where di is the difference for the ith pair and n is the number of d's.

  4. If the number of pairs (n) is large, the Spearman correlation maybe tested using statistics

    where t is distributed as Student's t-test with n–2 degrees of freedom.


   ACKNOWLEDGEMENTS
 
We thank Drs Theodore Pincus, Howard Fuchs, Victor Byrd, Tuulikki Sokka, Lloyd King and their patients for access to their clinics and for providing blood samples. We would also like to thank Bo Yelverton, Sukumar Narasimhulu and Xuan Li for technical assistance with experiments. This work was supported by grants from the National Institutes of Health (RO1 DK 58765 and RO1 AI 44924) and the Lupus Foundation.

Conflict of Interest statement: None declared.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 

  1. Marrack, P., Kappler, J. and Kotzin, B.L. (2001) Autoimmune disease: why and where it occurs. Nat. Med., 7, 899–905.[CrossRef][Web of Science][Medline]

  2. Brey, R.L., R.J., B. and Amato, A.A. (2000) Autoimmune nervous system disease. In Lahita, R.G. (ed.), Textbook of the Autoimmune Diseases. Lippincott Williams & Wilkins, Philadelphia, pp. 351–375.

  3. Atkinson, M.A. and Maclaren, N.K. (1994) The pathogenesis of insulin-dependent diabetes mellitus. N. Engl. J. Med., 331, 1428–1436.[Free Full Text]

  4. Weyand, C.M. and Goronzy, J.J. (2000) Rheumatoid arthritis. In Lahita, R.G. (ed.), Textbook of the Autoimmune Diseases. Lippincott Williams & Wilkins, Philadelphia, pp. 573–594.

  5. Lahita, R.G. (2000) Systemic lupus erythematosus. In Lahita, R.G. (ed.), Textbook of the Autoimmune Diseases. Lipinncott Williams & Wilkins, Philadelphia, pp. 537–547.

  6. Bennett, L., Palucka, A.K., Arce, E., Cantrell, V., Borvak, J., Banchereau, J. and Pascual, V. (2003) Interferon and granulopoiesis signatures in systemic lupus erythematosus blood. J. Exp. Med., 197, 711–23.[Abstract/Free Full Text]

  7. Baechler, E.C., Batliwalla, F.M., Karypis, G., Gaffney, P.M., Ortmann, W.A., Espe, K.J., Shark, K.B., Grande, W.J., Hughes, K.M., Kapur, V. et al. (2003) Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus. Proc. Natl Acad. Sci. USA, 100, 2610–2615.[Abstract/Free Full Text]

  8. Olsen, N., Sokka, T., Seehorn, C.L., Kraft, B., Maas, K., Moore, J. and Aune, T. (2004) Gene expression signature for recent onset rheumatoid arthritis in peripheral blood mononuclear cells. Ann. Rheum. Dis., 63, 1387–1392.[Abstract/Free Full Text]

  9. Maas, K., Chan, S., Parker, J., Slater, A., Moore, J., Olsen, N. and Aune, T.M. (2002) Cutting edge: molecular portrait of human autoimmune disease. J. Immunol., 169, 5–9.[Abstract/Free Full Text]

  10. Hochberg, M.C. (1987) The application of genetic epidemiology to systemic lupus erythematosus. J. Rheumatol., 14, 867–869.[Web of Science][Medline]

  11. Gregerson, P.K. (2000) Genetics of autoimmune disease. In Lahita, R.G. (ed.), Textbook of the Autoimmune Diseases, 1st edn. Lipincott Williams & Wilkins, Philadelphia, Vol. 1, pp. 37–45.

  12. Vyse, T.J. and Todd, J.A. (1996) Genetic analysis of autoimmune disease. Cell, 85, 311–318.[CrossRef][Web of Science][Medline]

  13. Seldin, M.F., Amos, C.I., Ward, R. and Gregersen, P.K. (1999) The genetics revolution and the assault on rheumatoid arthritis. Arthritis Rheum., 42, 1071–1079.[CrossRef][Web of Science][Medline]

  14. Brown, M.A., Kennedy, L.G., MacGregor, A.J., Darke, C., Duncan, E., Shatford, J.L., Taylor, A., Calin, A. and Wordsworth, P. (1997) Susceptibility to ankylosing spondylitis in twins: the role of genes, HLA, and the environment. Arthritis Rheum, 40, 1823–1828.[Web of Science][Medline]

  15. Heinzlef, O., Alamowitch, S., Sazdovitch, V., Chillet, P., Joutel, A., Tournier-Lasserve, E. and Roullet, E. (2000) Autoimmune diseases in families of French patients with multiple sclerosis. Acta. Neurol. Scand., 101, 36–40.[CrossRef][Web of Science][Medline]

  16. Shamim, E.A. and Miller, F.W. (2000) Familial autoimmunity and the idiopathic inflammatory myopathies. Curr. Rheumatol. Rep., 2, 201–211.[Medline]

  17. Lin, J.P., Cash, J.M., Doyle, S.Z., Peden, S., Kanik, K., Amos, C.I., Bale, S.J. and Wilder, R.L. (1998) Familial clustering of rheumatoid arthritis with other autoimmune diseases. Hum. Genet., 103, 475–482.[CrossRef][Web of Science][Medline]

  18. Ebers, G.C., Kukay, K., Bulman, D.E., Sadovnick, A.D., Rice, G., Anderson, C., Armstrong, H., Cousin, K., Bell, R.B., Hader, W. et al. (1996) A full genome search in multiple sclerosis. Nat. Genet., 13, 472–476.[CrossRef][Web of Science][Medline]

  19. Sawcer, S., Jones, H.B., Feakes, R., Gray, J., Smaldon, N., Chataway, J., Robertson, N., Clayton, D., Goodfellow, P.N. and Compston, A. (1996) A genome screen in multiple sclerosis reveals susceptibility loci on chromosome 6p21 and 17q22. Nat. Genet., 13, 464–468.[CrossRef][Web of Science][Medline]

  20. Haines, J.L., Ter-Minassian, M., Bazyk, A., Gusella, J.F., Kim, D.J., Terwedow, H., Pericak-Vance, M.A., Rimmler, J.B., Haynes, C.S., Roses, A.D. et al. (1996) A complete genomic screen for multiple sclerosis underscores a role for the major histocompatability complex. The Multiple Sclerosis Genetics Group. Nat. Genet., 13, 469–471.[CrossRef][Web of Science][Medline]

  21. Davies, J.L., Kawaguchi, Y., Bennett, S.T., Copeman, J.B., Cordell, H.J., Pritchard, L.E., Reed, P.W., Gough, S.C., Jenkins, S.C., Palmer, S.M. et al. (1994) A genome-wide search for human type 1 diabetes susceptibility genes. Nature, 371, 130–136.[CrossRef][Medline]

  22. Hashimoto, L., Habita, C., Beressi, J.P., Delepine, M., Besse, C., Cambon-Thomsen, A., Deschamps, I., Rotter, J.I., Djoulah, S., James, M.R. et al. (1994) Genetic mapping of a susceptibility locus for insulin-dependent diabetes mellitus on chromosome 11q. Nature, 371, 161–164.[CrossRef][Medline]

  23. Rowe, R.E., Wapelhorst, B., Bell, G.I., Risch, N., Spielman, R.S. and Concannon, P. (1995) Linkage and association between insulin-dependent diabetes mellitus (IDDM) susceptibility and markers near the glucokinase gene on chromosome 7. Nat. Genet., 10, 240–242.[CrossRef][Web of Science][Medline]

  24. Gaffney, P.M., Ortmann, W.A., Selby, S.A., Shark, K.B., Ockenden, T.C., Rohlf, K.E., Walgrave, N.L., Boyum, W.P., Malmgren, M.L., Miller, M.E. et al. (2000) Genome screening in human systemic lupus erythematosus: results from a second Minnesota cohort and combined analyses of 187 sib-pair families. Am. J. Hum. Genet., 66, 547–556.[CrossRef][Web of Science][Medline]

  25. Moser, K.L., Neas, B.R., Salmon, J.E., Yu, H., Gray-McGuire, C., Asundi, N., Bruner, G.R., Fox, J., Kelly, J., Henshall, S. et al. (1998) Genome scan of human systemic lupus erythematosus: evidence for linkage on chromosome 1q in African–American pedigrees. Proc. Natl. Acad. Sci. USA, 95, 14869–14874.[Abstract/Free Full Text]

  26. Shai, R., Quismorio, F.P., Jr., Li, L., Kwon, O.J., Morrison, J., Wallace, D.J., Neuwelt, C.M., Brautbar, C., Gauderman, W.J. and Jacob, C.O. (1999) Genome-wide screen for systemic lupus erythematosus susceptibility genes in multiplex families. Hum. Mol. Genet., 8, 639–644.[Abstract/Free Full Text]

  27. Lindqvist, A.K., Steinsson, K., Johanneson, B., Kristjansdottir, H., Arnasson, A., Grondal, G., Jonasson, I., Magnusson, V., Sturfelt, G., Truedsson, L. et al. (2000) A susceptibility locus for human systemic lupus erythematosus (hSLE1) on chromosome 2q. J. Autoimmun., 14, 169–178.[CrossRef][Web of Science][Medline]

  28. Jawaheer, D., Seldin, M.F., Amos, C.I., Chen, W.V., Shigeta, R., Monteiro, J., Kern, M., Criswell, L.A., Albani, S., Nelson, J.L. et al. (2001) A genomewide screen in multiplex rheumatoid arthritis families suggests genetic overlap with other autoimmune diseases. Am. J. Hum. Genet., 68, 927–936.[CrossRef][Web of Science][Medline]

  29. Cornelis, F., Faure, S., Martinez, M., Prud'homme, J.F., Fritz, P., Dib, C., Alves, H., Barrera, P., de Vries, N., Balsa, A. et al. (1998) New susceptibility locus for rheumatoid arthritis suggested by a genome-wide linkage study. Proc. Natl Acad. Sci. USA, 95, 10746–10750.[Abstract/Free Full Text]

  30. Becker, K.G., Simon, R.M., Bailey-Wilson, J.E., Freidlin, B., Biddison, W.E., McFarland, H.F. and Trent, J.M. (1998) Clustering of non-major histocompatibility complex susceptibility candidate loci in human autoimmune diseases. Proc. Natl Acad. Sci. USA, 95, 9979–9984.[Abstract/Free Full Text]

  31. Jawaheer, D., Seldin, M.F., Amos, C.I., Chen, W.V., Shigeta, R., Etzel, C., Damle, A., Xiao, X., Chen, D., Lum, R.F. et al. (2003) Screening the genome for rheumatoid arthritis susceptibility genes: a replication study and combined analysis of 512 multicase families. Arthritis Rheum., 48, 906–916.[CrossRef][Web of Science][Medline]

  32. Brem, R.B., Yvert, G., Clinton, R. and Kruglyak, L. (2002) Genetic dissection of transcriptional regulation in budding yeast. Science, 296, 752–755.[Abstract/Free Full Text]

  33. Schadt, E.E., Monks, S.A., Drake, T.A., Lusis, A.J., Che, N., Colinayo, V., Ruff, T.G., Milligan, S.B., Lamb, J.R., Cavet, G. et al. (2003) Genetics of gene expression surveyed in maize, mouse and man. Nature, 422, 297–302.[CrossRef][Medline]

  34. Westfall, P.H. and Young, S.S. (1993) Resampling-based multiple testing: examples and methods for p-value adjustment. John Wiley & Sons, Inc. New York, pp. 62–72.

  35. Bias, W.B., Reveille, J.D., Beaty, T.H., Meyers, D.A. and Arnett, F.C. (1986) Evidence that autoimmunity in man is a Mendelian dominant trait. Am. J. Hum. Genet., 39, 584–602.[Web of Science][Medline]

  36. Monks, S.A., Leonardson, A., Zhu, H., Cundiff, P., Pietrusiak, P., Edwards, S., Phillips, J.W., Sachs, A. and Schadt, E.E. (2004) Genetic inheritance of gene expression in human cell lines. Am. J. Hum. Genet., 75, 1094–1105.[CrossRef][Web of Science][Medline]

  37. Cheung, V.G., Conlin, L.K., Weber, T.M., Arcaro, M., Jen, K.Y., Morley, M. and Spielman, R.S. (2003) Natural variation in human gene expression assessed in lymphoblastoid cells. Nat. Genet., 33, 422–425.[CrossRef][Web of Science][Medline]

  38. Morley, M., Molony, C.M., Weber, T.M., Devlin, J.L., Ewens, K.G., Spielman, R.S. and Cheung, V.G. (2004) Genetic analysis of genome-wide variation in human gene expression. Nature, 430, 743–747.[CrossRef][Medline]

  39. Cavalieri, D., Townsend, J.P. and Hartl, D.L. (2000) Manifold anomalies in gene expression in a vineyard isolate of Saccharomyces cerevisiae revealed by DNA microarray analysis. Proc. Natl. Acad. Sci. USA, 97, 12369–12374.[Abstract/Free Full Text]

  40. Sandberg, R., Yasuda, R., Pankratz, D.G., Carter, T.A., Del Rio, J.A., Wodicka, L., Mayford, M., Lockhart, D.J. and Barlow, C. (2000) Regional and strain-specific gene expression mapping in the adult mouse brain. Proc. Natl. Acad. Sci. USA, 97, 11038–11043.[Abstract/Free Full Text]

  41. Primig, M., Williams, R.M., Winzeler, E.A., Tevzadze, G.G., Conway, A.R., Hwang, S.Y., Davis, R.W. and Esposito, R.E. (2000) The core meiotic transcriptome in budding yeasts. Nat. Genet., 26, 415–423.[CrossRef][Web of Science][Medline]

  42. Prahalad, S., Shear, E.S., Thompson, S.D., Giannini, E.H. and Glass, D.N. (2002) Increased prevalence of familial autoimmunity in simplex and multiplex families with juvenile rheumatoid arthritis. Arthritis Rheum., 46, 1851–1856.[CrossRef][Web of Science][Medline]

  43. Kim, S., Dougherty, E.R., Chen, Y., Sivakumar, K., Meltzer, P., Trent, J.M. and Bittner, M. (2000) Multivariate measurement of gene expression relationships. Genomics, 67, 201–209.[CrossRef][Web of Science][Medline]

  44. Chen, Y., Dougherty, E.R. and Bittner, M. (1997) Ratio-based decisions and the quantitative analysis of cDNA microarray images. J. Biomed. Opt., 2, 364–374.[CrossRef]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Ann Rheum DisHome page
E J M Toonen, P Barrera, T R D J Radstake, P L C M van Riel, H Scheffer, B Franke, and M J H Coenen
Gene expression profiling in rheumatoid arthritis: current concepts and future directions
Ann Rheum Dis, December 1, 2008; 67(12): 1663 - 1669.
[Abstract] [Full Text] [PDF]

Home page
 J. Mol. Diagn.Home page
S. C. Fossey, C. L. Vnencak-Jones, N. J. Olsen, S. Sriram, G. Garrison, X. Deng, P. S. Crooke III, and T. M. Aune
Identification of Molecular Biomarkers for Multiple Sclerosis
J. Mol. Diagn., April 1, 2007; 9(2): 197 - 204.
[Abstract] [Full Text] [PDF]

Home page
 Hum Mol GenetHome page
Z. Liu, K. Maas, and T. M. Aune
Identification of gene expression signatures in autoimmune disease without the influence of familial resemblance
Hum. Mol. Genet., February 1, 2006; 15(3): 501 - 509.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Material
Right arrow All Versions of this Article:
14/10/1305    most recent
ddi141v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (10)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Maas, K.
Right arrow Articles by Aune, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Maas, K.
Right arrow Articles by Aune, T.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?