NIH Campus, Building 31, Room 4C32
Susana Serrate-Sztein, MD, NIAMS
Stephen I. Katz, MD, PhD, NIAMS
John Reveille, MD, University of Texas Health Science Center at Houston
The objective of this roundtable discussion was to provide NIAMS leadership with information about scientific opportunities, needs, and roadblocks in the biology and treatment of psoriasis, psoriatic arthritis, and rheumatoid arthritis. Consultation with outside experts working within this field pointed to topics, concerns, challenges, and innovations viewed by the scientific community. Emerging themes that arose from the discussion will help NIAMS frame its long-term scientific planning and priority-setting.
Key questions addressed in the roundtable included:
- What are the most promising areas of science in the field?
- What are the most pressing scientific needs?
- What can NIAMS/NIH contribute uniquely to fill gaps or address needs?
Inflammation represents an early and key event in the development of rheumatoid arthritis (RA), psoriasis (Pso), and psoriatic arthritis (PsA). Compelling evidence indicates that the production of tumor necrosis factor alpha (TNF-α) plays a central role in these diseases by sustaining the inflammatory process in the skin, as well as in the joints. The diseases also share targets for intervention involving leukocyte-endothelial cell interactions, T-cell activation, and antigen-presenting cells.
Genetic studies suggest that RA, Pso, and PsA share some susceptibility genes. Human leukocyte antigen (HLA) genes have been the subject of extensive investigation. HLA-C is associated with Pso risk in specific human populations: Caucasians, but not Asians or Sardinians. The mechanisms involving HLA molecules in pathways of disease remain elusive. Major histocompatibility (MHC) genes have also been associated with RA (DRB-1), but the mechanism is unknown. There may be gene duplications and gene fusions, but the studies have not been conclusive. PSORS1, widely recognized as a major genetic determinant of psoriasis, has been associated with 40-50% of all Pso cases. Recent research exploring non-HLA genes has revealed new molecules and pathways involved in immune and inflammatory processes, such as the protein tyrosine phosphatase non-receptor type 22 (PTPN22) gene. Polymorphisms in interleukin (IL)-12B and IL-23R genes are associated with Pso.
The variation in clinical phenotype of RA, Pso, and PsA creates formidable challenges in understanding the underlying causes, progression, and treatment of these diseases. Several recent research studies report that autoantibody expression can be detected before the appearance of clinical symptoms, and the presence of autoantibodies hints at the stage of a disease. Macrophages may mirror the different stages of diseases; they may be the trigger that begins the cascade of activation of specific T-cells, resulting in the release of specific cytokines. Technical challenges (e.g., poor yield from synovial fluid, difficulty in extraction from skin, instability of stage-specific phenotype) could hamper the development of macrophages as biomarkers, however methods such as laser capture microdissection to obtain RNA profiles are promising. MicroRNAs have been employed as early markers in cancer, so there is significant potential for their use in inflammation. There are many "AU rich elements" ("AREs") in mRNA, associated with inflammation, to which microRNAs can bind.
Common biomarkers should be identified and characterized, to assess inflammation "load." There is speculation that microRNAs could be involved in the coordination of inflammatory mediators. A systems biology approach would provide a broader picture of inflammation: common, and divergent, pathways. Reducing inflammation many not necessarily limit damage, such as fibrosis, which is an indication of different pathways. Inflammation and organ damage must be differentiated. Differences between spinal and joint arthropathies also suggest different pathways, so it is important to understand systemic effects of inflammation.
Advances in lupus research may be applicable to Pso, PsA, and RA, including the different reactions to cytokine inhibitors in these diseases; RA is associated with TNF and lupus is associated with interferon (IFN), raising questions about potential crossover in their mechanistic pathways and whether tissue damage may be caused by one or the other pathway. The adaptive immune system, via immune complexes, could stimulate the innate immune system in these diseases. There could be negative regulation of immune system activation through toll-like receptors (TLRs): reactions could be dampened as well as stimulated. Recent attention to the skin microbiome suggests that there are not immune responses to all of these microbes under normal conditions, but in certain genetic backgrounds, the presence of these microbes may trigger autoimmune diseases.
Other analyses, such as positron emission tomography (PET) scans-which have not been validated for joint or skin inflammation--should be employed, in order to get an aggregate inflammation picture from co-morbidities and other organ involvement. In these cases, biomarkers may provide insights into disease mechanisms, but they may not be useful as predictors, due to the complexity of these diseases. For example, magnetic resonance imaging (MRI) has shown that the initial site of involvement may be only a transient observation, which may not yield useful information of a sustained condition.
Unaffected skin in Pso patients is often abnormal, and there may be similar examples in other rheumatic diseases (such as the synovium of an unaffected knee in a psoriatic arthritis patient). There are ethical issues in conducting such proof-of-concept studies, because biopsy of unaffected tissues could stimulate disease. Hence, a concerted effort towards advancing non-invasive imaging deserves attention. For example, "smart" contrast agents could be developed to monitor activation of specific T-cell populations.
Numerous co-morbidities must be considered in association with these diseases, which can also contribute to the broader inflammation picture. Vascular, joint, and skin inflammation should be addressed together. In RA, there is clear evidence of bone loss—an example of an end-organ system affected by local inflammation. The joint may be acting like an endocrine organ, releasing cytokines and inflammatory factors abnormally. Recent publications cite the impact of inflammation on bone and cartilage remodeling and resorption, via wnt and bone morphogenic protein (BMP) pathways. Matrix and cartilage degradation products stimulate inflammation; for example, fibronectin fragments bind to TLRs.
Cardiovascular risks increase in these diseases, so it is important to include cardiologists in these research projects. There may also be a mechanistic link between these rheumatic diseases and atherosclerotic plaques, which produce inflammatory cytokines. Adipose tissue also produces inflammatory cytokines, such as TNF-α and C-reactive protein, accompanied by macrophage infiltration. Type II diabetes has also been tied to generalized inflammation patterns.
Smoking has been viewed as a low-grade inflammatory stimulant; the mechanism has not been established. Enhancement of antigen presentation has been proposed. Many macrophages and antibodies are associated with chronic obstructive pulmonary disorder (COPD). Citrullination of type II collagen, which is associated with smoking, has been observed in rheumatoid arthritis; it does not trigger the disease, but it enhances it. As well, vasoconstriction from smoking may reduce severity of psoriatic plaques.
The common and divergent disease pathways invite therapeutic targets. Industry is very interested in developing TNF inhibitors for treatment of heart disease, which could be concurrently applicable to these rheumatic diseases. Improvement has been seen with anti-TNF treatment in RA patients with reduction of some cardiovascular complications. However, responses are variable, depending on the individual and the TNF inhibitor.
IL-23 and its stimulation of T-helper (Th)17 cells have emerged as being directly associated with Pso (Th1 may be a bystander and anti-inflammatory; the disease gets worse when Th1 is knocked out). Anti-p40 therapy, which targets an IL-23 subunit, is very effective in Pso treatment. The IL-23/Th17 pathway has a lot of highly-specific, potential targets. Other targets include TGF-β, IL-6, IL-22, and pro-apoptotic signals (TNF-receptor associated factors: TRAFs), as well as TLRs and JAK/STAT pathways.
All of these topics shed light on unmet needs and opportunities for NIH to facilitate solutions. For example, there are tremendous software needs for the utilization of proteomics data and linking gene expression information to other data for interdisciplinary projects. Most of this expertise lies with industry or in large groups. Rather than "reinventing the wheel," particularly for smaller research groups, perhaps NIH could facilitate interactions with industry for broader access to these tools and create a national core for NIH-funded researchers to obtain bioinformatics support. Availability of central core technology facilities, such as microarray processing, could create an economy of scale. The consortium of NIH Roadmap-funded Clinical and Translational Science Awards (CTSAs) is discussing acquisition of core facilities for technologies such as genomics or imaging.
Sufficient clinical study populations are needed, particularly to study disease subsets, as noted in Pso. There is a higher prevalence of rheumatic diseases in Native American populations. There is great interest in African/African American populations, because of the longer evolutionary lineage, in comparison to European/North American populations. As a result, there is greater genetic heterogeneity and fewer linked (in other words, many more independently-distributed) single nucleotide polymorphisms (SNPs) in African/African American populations. Hence, smaller blocks of DNA from these populations can be compared in genetic studies. Unfortunately, a lot of clinical research data are skewed by over-representation of Caucasians. There is a recognized difficulty in getting the participation of African Americans in clinical trials (including lupus and RA studies); cultural barriers are often cited. Patient advocacy organizations could be very effective conduits for providing education to encourage participation.
Researchers in several countries with less genetic heterogeneity have created banks of samples and DNA information from large populations (Iceland, France, Canada, Sweden). Some have NIH collaborations already, but perhaps NIH could facilitate more of these collaborations; however, there are concerns about differing standards of data collection among collaborators.
The HapMap Project is a powerful resource. Although SNPs may be useful markers, they may not be correlated directly with a disease. It would be necessary to sequence large stretches of DNA, following identification with SNP information, to find useful sequence information associated with a disease; MHC polymorphisms are relevant examples to rheumatic diseases.
Involvement of healthcare systems may enhance clinical trial recruitment. The Association of American Medical Colleges (AAMC) can lobby the Centers for Medicare and Medicaid Services (CMS) to include financial incentives to create local cohorts. Industry-sponsored clinical trials generate a localized source of well-characterized patients. Utilizing these would be more efficient than creating a cohort de novo. The Food and Drug Administration (FDA) is also interested in creating such cohorts, as it tries to collect information on adverse events. Patients at VA facilities may develop RA and may be a good population for longitudinal studies.
Researchers have had difficulty obtaining R01 support for longitudinal studies, which take more than five years to bear results (for example, cardiovascular complications with Pso or RA). This is especially daunting for international collaborations, which need more than one source of support.
Many researchers entreat their colleagues to obtain and "bank" as much sample as possible. Cost and logistical issues, such as storage, must be addressed when developing a large cohort and associated samples. There are also concerns about obtaining samples for unspecified research purposes. There needs to be standardization in sampling (particularly with skin) so that results from different studies can be compared rigorously. NIH could facilitate efforts among research groups to reach consensus on these standards.
Lack of accurate clinical phenotypes hampers genetic and epigenetic studies. Standardization is also needed for classification by stage of the disease when the sample is obtained (presenting or evolving phenotype). NIH could lead a biologically-based (e.g., genetics, mechanisms) classification of disease subsets.
Consent form language could allow unrestricted use of anonymized data, but experience shows that local institutional review boards (IRBs) are variable and lack consistency in interpreting consent form language. Uniform IRB policies for obtaining appropriate samples are essential, especially when participants from multiple sites are needed for sufficient power to analyze MHC trends and shared epitopes. NIH and the FDA are working on consent form harmonization.
Industry is generating interesting therapeutic molecules which can provide important insights on disease mechanisms. Mechanistic studies are peripheral to most companies' interests, so perhaps NIH could fund these projects, in association with industry-sponsored clinical trials. There have been some successful precedents of this model, but proprietary concerns of companies have impaired others.
Some in the extramural community sense a disparity between the stated NIH priorities on translational research and study sections' evaluation of applications. Many of the criticisms are related to logistical weaknesses in the planned projects. There is concern that the members of standing integrated review groups (IRGs) do not understand all aspects of translational research proposals when the applications cross disciplines. Perhaps there should be separate study sections for clinical/translational research.
ABECASIS, Gonçalo, Ph.D.
Associate Professor, Center for Statistical Genetics
Department of Biostatistics
University of Michigan
ANDERSON, Paul J., M.D., Ph.D.
K. Frank Austen Professor of Medicine
Division of Rheumatology, Immunology and Allergy
Brigham and Women's Hospital
BLAUVELT, Andrew, M.D.
Chief, Dermatology Service
Portland VA Medical Center
Professor, Departments of Dermatology and Molecular Microbiology and Immunology
Oregon Health and Science University
BATHON, Joan M., M.D.
Professor, Department of Medicine
Director, The Johns Hopkins Arthritis Center
Johns Hopkins University
CALLIS, Kristina, M.D.
Assistant Professor, Department of Dermatology
University of Utah Health Sciences Center
COOPER, Kevin D., M.D.
Professor and Chair, Department of Dermatology
Case Western Reserve University and
University Hospitals of Cleveland
CROW, Mary K., M.D.
Associate Chief, Division of Rheumatology
Director of Rheumatology Research
Hospital for Special Surgery
de MUMMEY, Carmen C.
Thousand Oaks, California
GOLDRING, Steven R., M.D.
St. Giles Chair and Chief Scientific Officer
Hospital for Special Surgery
MATHIS, Diane J., Ph.D.
Professor, Department of Medicine
Section Head, Immunology and Immunogenetics
Joslin Diabetes Center
Harvard Medical School
REVEILLE, John, M.D.
Director, Division of Rheumatology and Clinical Immunogenetics
Department of Internal Medicine
University of Texas Health Science Center
WINCHESTER, Robert, M.D.
Professor of Medicine, Pediatrics, and Pathology
Columbia University College of Physicians and Surgeons