SEARCH

SEARCH BY CITATION

Introduction

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES

As with other forms of chronic inflammatory disease, autoimmune or otherwise, patients with juvenile idiopathic arthritis (JIA) may have only a limited family history of the JIA phenotype but a more evident family history of autoimmune disease in general (1). Based on this family history and reports of HLA associations in JIA, the generation of risk of developing the disease is likely based on complex trait genetics (for review, see ref.2).

Our understanding of the genetic contribution has not progressed as quickly for JIA as compared with other autoimmune arthropathies, including rheumatoid arthritis and ankylosing spondylitis. A major hindrance to progress has been the change in nomenclature, from 3 subtypes in the familiar juvenile rheumatoid arthritis system to 7 subtypes in the JIA system (3). This new nomenclature, while now commonly used, has not been adopted by the American College of Rheumatology. Consistent with observations in the clinic, the JIA criteria provide increases in phenotype homogeneity that can be translated to a corresponding need for studies to consider 7 subtypes as separate entities. This splitting of the patient population into smaller groups provides challenges for recruiting patient cohorts of adequate size within any one subtype with sufficient analytical power for genomic studies. Findings of recent gene expression studies suggest that additional heterogeneity remains even within the JIA classification.

Differences in phenotype between children and adults with arthritis

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES

The observation that there are phenotype features present in JIA that are not found in adults with arthritis (4) suggests that examining pediatric arthritis at the molecular level will yield new information related to autoimmunity. Examples of these unusual phenotype features include the observation that oligoarticular JIA, the most common JIA subtype, does not have a postadolescence or adult-onset counterpart. In fact, in females, the onset of oligoarticular JIA after the age of 6 or 7 years is relatively rare. Thus, the age at onset is an important element of categorization, as supported by the recent study by Hollenbach et al (5).

Another important difference from adult-onset arthritis is that, with the exception of a small group of patients with late-onset polyarticular disease, the vast majority of JIA patients do not have antibodies to cyclic citrullinated peptides, nor do they have IgM rheumatoid factor (IgM-RF), although antinuclear antibodies (ANAs) may be variably present.

HLA associations with JIA

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES

In the past, the sizes of the available JIA cohorts, both national and international, have proven less than robust for many analytical purposes, except perhaps for HLA studies (6). But even in the case of HLA analyses, further increases in cohort sizes and HLA typing resolution provide an opportunity to more precisely define susceptibility alleles and haplotypes (see Tables 3 and 4 in the article by Hollenbach et al [5]). In addition, the availability of larger cohorts also allows definitive statistical exclusion of other HLA alleles that have been variably, but not consistently, associated with JIA in some, but not all, studies. The recent study by Hollenbach et al (5) included a well-characterized and substantial North American JIA cohort and a well-matched community-based control population. Typing of 8 HLA loci (HLA–A, B, C, DRB1, DPA1, DPB1, DQA1, and DQB1) at a 4-digit level of resolution was included and thus provided a comprehensive data set for analysis. From this data set and considering class II loci, it was found that HLA–DR, rather than HLA–DQ, is important in the pathogenesis of JIA. This finding was consistent with the conclusions made by other investigators (7).

The scope of the data set examined by Hollenbach et al also led to novel findings that included the identification of susceptibility and protective HLA–DRB;DQA;DQB haplotypes, age-at-onset–related effects, as well as novel HLA class I associations, including HLA–C. The association with HLA–C and its known role as a ligand for natural killer cell receptors suggest avenues for the investigation of innate immunity in JIA. In addition, this robust experimental design confirms the presence of a protective haplotype (DRB1*1501;DQA1*0602;DQB1*0102) that is shared by both the oligoarticular and the polyarticular JIA subtypes tested. Interestingly, DRB1*1501 is strongly associated with disease risk in both systemic lupus erythematosus and multiple sclerosis, and as in JIA, this allele is protective against IgA immunodeficiency disease; thus, the importance of DRB1*1501 has been consistently noted across multiple forms of autoimmunity (8).

Genome-wide association studies identify additional risk factors for JIA

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES

Acquiring multidimensional data in large cohorts enhances the value of the cohort. Genome-wide single-nucleotide polymorphism (SNP) data for a cohort include ancestry-informative markers for consideration of population substructure as well as SNPs for fine-mapping associations within the major histocompatibility complex (MHC). In this context, SNP data for the 1,000-strong Cincinnati Genomic Cohort, combined with high-resolution HLA typing available for a large subset as a result of the work reported in Hollenbach et al (5), now comprise a particularly valuable resource of locally matched controls.

The importance of additional variation in the MHC region beyond the associated HLA loci needs to be reexamined in JIA. In the data set described by Hollenbach et al (5), this can be done by conditioning on the associated JIA HLA haplotypes and focusing on SNPs that are not in linkage disequilibrium with any of the HLA loci. Most recently, the analysis of fixed sequences derived from exon 2 in HLA class II proteins (exon 2) has added another route for association studies and has expanded their potential (9).

A variety of risk factors outside the MHC have been identified in different autoimmune diseases, including the arthropathies (http://www.genome.gov/gwastudies). Genome-wide association studies are based on the assumption that common genetic variants of modest effect are responsible for disease, with current technology relying on linkage disequilibrium between the measured SNP and the causal variant. From these studies, common themes have emerged across autoimmune diseases, revealing genes that are associated with susceptibility to more than one disease, as well as genes specific to a single disease (10). Consistent with this, recent studies of patients with JIA have identified genetic associations that are in common with those of autoimmune diseases including adult arthritis (11, 12).

Efforts are under way in several laboratories to identify the unique associations for JIA using genome-wide SNP technologies and patient cohorts exceeding 5,000 worldwide. The findings of JIA genome-wide studies that have been reported to date, although limited in the number of JIA cases examined (13) or the number of SNPs tested (14), illustrate the promise of deciphering the pathophysiology of JIA in the context of autoimmune disease. These expanding genome-wide association studies are vastly more extensive (and expensive) than originally anticipated but are necessary because of the heterogeneous nature of the JIA phenotype. The total samples in progress may not yet be sufficient to identify susceptibility and protective risk factors for each JIA subtype, even through collaborative efforts in the research community. Like other autoimmune diseases, there is, in general, much that is still to be learned as the actual causal variants are identified.

Even as the association analyses using SNP markers reveal the complexity of genetic risk in JIA and the JIA subtypes, no clear path for identifying the causal variants has emerged from studies in other autoimmune diseases. This has led to a proposal that multiple rare variants in a region act over large distances, thereby creating associations for the common variants that are detected in genome-wide association studies (15). The idea of rare variants may have implications for fine-mapping studies in the MHC, especially with regard to the high level of linkage disequilibrium present there.

Current microarray or microbead-based technologies may now assay up to 19,000 polymorphisms within the extended MHC and may provide for a more detailed mapping of MHC associations. In a genome-wide SNP data set corresponding to the HLA cohorts evaluated by Hollenbach et al (5), differences in the patterns of SNP associations between the oligoarticular and the polyarticular JIA subtypes were found (16). It is not clear, however, if a specific allele at a given locus will mediate consistent effects in different JIA populations, even within a JIA subtype.

Genome-wide gene expression analyses in JIA

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES

In addition to genome-wide SNP-based genotyping, genome-wide gene expression analyses may also contribute to better classification of JIA by demonstrating heterogeneity within classes and overlap between classes. Gene expression microarrays have been used to analyze samples from many of the subjects who were part of the JIA genomic studies discussed above. While these studies are large by gene expression standards, the cohorts are quite modest when compared with genomic cohorts. Nevertheless, gene expression profiles in peripheral blood mononuclear cells obtained prior to therapy with disease-modifying antirheumatic drugs can identify gene expression differences between groups of samples from patients with several JIA subtypes and those from normal controls (17). Further, gene expression analyses can identify phenotypically distinguishable subsets within JIA subtypes, including polyarticular JIA (18), systemic JIA (19), and oligoarticular JIA (20).

Preliminary analysis of samples from patients with oligoarticular JIA (20) demonstrate how gene expression profiling can be integrated with genetic associations to characterize JIA heterogeneity. Different HLA associations between groups of oligoarticular JIA patients with early disease onset (<6 years of age) and those with later disease onset (≥6 years) have been identified (5). Likewise, gene expression profiling can distinguish these same two groups of patients (early versus later disease onset) with persistent oligoarticular JIA. Unexpectedly, this same gene expression pattern can differentiate patients with early-onset versus late-onset RF-negative polyarticular JIA. In patients with systemic JIA, however, this pattern does not distinguish early-onset from late-onset disease, nor does it differentiate young (<6 years at time of sample) from old (≥6 years) normal control subjects.

Principal components analysis of this expression pattern shows that age at onset is more informative than JIA subtype (persistent oligoarticular or RF-negative polyarticular), which is consistent with the findings of the principal components analysis of HLA genetic associations in the study by Hollenbach et al (5). The early-versus-late gene expression pattern shows increased expression of many B cell–related genes in patients with early-onset JIA, which is consistent with previous observations of ANA positivity in patients with early-onset disease (21). The gene expression pattern also shows increased expression of many genes related to cells of the myeloid lineage in patients with late-onset disease. This list of genes overexpressed in late onset overlaps with one of the gene expression signatures described in a subset of patients with polyarticular JIA (18).

Childhood development and autoimmunity

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES

It is noteworthy that in the majority of patients, oligoarticular JIA is diagnosed between the ages of 0 and 6 or 7 years. Furthermore, adult onset of the oligoarticular subtype of JIA is rare or nonexistent. This marked age effect might suggest that the onset of JIA in the youngest children is part of a developmentally related phenotype. The same HLA–DRB1*0801;DQ haplotype is associated with oligoarticular and polyarticular JIA in patients of all ages, while this is not true for haplotypes that include HLA–DRB1*1103/4 and HLA–DRB1*1301, which are only associated with patients who have early-onset JIA.

Corresponding findings are apparent in the gene expression data showing that patients with early-onset oligoarticular and polyarticular disease have overlapping patterns of expression. The patients therefore evolve through periods of significant developmental changes as they pass from infancy through puberty, in contrast to adult-onset disease, where more autoimmune features predominate. This may be somewhat reminiscent of the effects of Th1/Th2 in allergy phenotypes in young children generally under 5 years of age, in which their initial reactivity may diminish with age for some allergens, such as atopic asthma, but not for others, such as peanuts (22). An age-related switch occurs from a Th1 to a Th2 cytokine phenotype. The differences between autoimmune and autoinflammatory disease may also be part of a developmental concept.

Potential for improved treatment decisions and outcomes in JIA

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES

Molecular characterization of JIA subsets holds promise for guiding treatment and predicting outcomes in patients with JIA. For example, subsets of patients with polyarticular JIA defined by gene expression profiling were found to have differences in their prescribed treatments as well as in their 2-year outcomes (additional analyses of data from the Griffin et al study [18]). Not surprisingly, a subset of patients that appeared to have more severe disease was more likely than the other subsets to be treated with biologic response modifiers that affect tumor necrosis factor. Conversely, a subset of patients with less severe disease at baseline had worse outcomes at 2 years, suggesting that they were not treated aggressively enough. Other investigators have recently reported that in oligoarticular JIA, a strong interferon-γ–induced signature in synovial mononuclear cells at disease onset might predict the extension of the disease to polyarthritis (23). It is expected that early introduction of more-aggressive treatment in these patients might improve the long-term outcome of the disease. Another interesting conclusion drawn from the gene expression studies in JIA is that in most patients, disease remission is associated with a better balance between proinflammatory and antiinflammatory networks rather than simply a return to normal immune function (24). This intriguing concept also has the potential to influence treatment decisions made by clinicians. Thus, in the future, gene expression profiles could be used to guide treatment and attain optimal outcome. Taken together, these examples suggest that gene expression profiling (or its surrogates) in combination with genetic analyses hold promise for providing a wealth of information about pathogenic mechanisms, disease activity levels, and likely outcomes in JIA.

Gene expression studies have also provided additional evidence that the role of the adaptive immune system is rather limited in systemic JIA as compared with the other JIA subtypes, whereas the contribution of the innate immunity may be much more prominent (19, 25, 26). Several recent gene expression (microarray) studies, for example, have shown that systemic JIA can be distinguished from other subtypes of JIA by the presence of an up-regulation of the innate immune pathways, including interleukin-6, Toll-like receptor/interleukin-1 receptor, as well as a down-regulation of the gene networks involving T cell–related and MHC-related biologic processes, including antigen presentation (19). This pattern is strikingly similar to that seen in patients with autoinflammatory syndromes such as neonatal-onset multisystem inflammatory disease (NOMID). Although some elements of the “innate signature” that is seen in systemic JIA are also present in patients with seronegative polyarticular JIA, these studies further support the idea that systemic JIA is markedly different from the other subtypes of JIA and should be viewed as an autoinflammatory syndrome, rather than a “classic” autoimmune disease.

Conclusion

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES

Genome-wide technologies for both DNA and RNA measurements, the increasing availability of validated disease-specific biomarkers, and the development of novel analytic strategies afford opportunities for correlating biology with disease outcomes and for providing a biologic basis for disease classification. Most important to these efforts are the existing cohorts and ongoing collections which, together, will comprise sufficient numbers of patient samples with uniform clinical data to allow replication of findings in a manner consistent with genome-wide statistical testing. In the near future, a molecular toolbox will be available to the clinician that will aid in diagnosis and in selection of therapies, thus providing a better outcome for patients with JIA.

REFERENCES

  1. Top of page
  2. Introduction
  3. Differences in phenotype between children and adults with arthritis
  4. HLA associations with JIA
  5. Genome-wide association studies identify additional risk factors for JIA
  6. Genome-wide gene expression analyses in JIA
  7. Childhood development and autoimmunity
  8. Potential for improved treatment decisions and outcomes in JIA
  9. Conclusion
  10. AUTHOR CONTRIBUTIONS
  11. REFERENCES