Article Text

Download PDFPDF

Prevention of autoimmune rheumatic disease: state of the art and future perspectives
  1. Lars Klareskog1,
  2. Peter K Gregersen2,
  3. Tom W J Huizinga3
  1. 1Rheumatology Unit, Department of Medicine, Karolinska Institutet and Karolinska University Hospital, Solna, Stockholm, Sweden
  2. 2Robert S Boas Center for Genomics and Human Genetics, The Feinstein Institute for Medical Research, North Shore LIJ Health System, Manhasset, New York, USA
  3. 3Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands
  1. Correspondence to Lars Klareskog, Rheumatology Unit, Department of Medicine, Karolinska Institutet and Karolinska University Hospital, Solna, 171 76 Stockholm, Sweden; lars.klareskog{at}


Prevention of disease can in principle be accomplished by identification of environmental and/or lifestyle risk and protective factors followed by public health measures (such as for smoking and lung cancer), or by modification of the individual's reactions to disease-inducing factors (such as in vaccinations against microbes). This review discusses both options based on emerging understanding of aetiologies in inflammatory rheumatic diseases such as rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). The major current opportunity for public health-based prevention lies in avoiding smoking. In RA, recent studies have calculated that, in Sweden (a country characterised by a low frequency of smoking), 20% of all RA cases and 33% of all cases of ACPA-positive RA would not have occurred in a smoke-free society. Smoking is also a major risk factor for SLE but no population attribution is yet available. New avenues for individualised and biology-based prevention are provided by the demonstration that several autoimmune rheumatic diseases are preceded by emergence of subclinical autoimmunity followed by laboratory-based signs of inflammation and finally overt disease. Examples of this process are provided from studies of autoimmunity to citrullinated proteins (in RA), to dsDNA (in SLE in general) and to Ro52 epitopes (in the case of neonatal heart block). The recognition of this sequence of events provides opportunities to intervene specifically and potentially curatively before onset of full-blown disease. Such prevention can be accomplished by modification of inciting antigens (environment), by modification of immunity (more or less specific immunomodulation) or by modification of specific gene functions. In all cases, prevention will be different in different subsets of disease and differ at different time points of disease development. Thus, the road map towards prevention of autoimmune rheumatic diseases includes increased understanding of how genes, environment and immunity interact.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Autoimmune diseases can be seen as a consequence of the imminent problem facing the immune system—that is, a system which needs to be able to attack and eliminate microbes which themselves have a large capacity to adapt their exterior to be similar to molecules present in the host. The evolution of the immune system has dealt with this problem at many different levels. Thus, in humans as well as in other mammals, there are many mechanisms for generation of diversity of the immunological repertoire and means of immune attack, ensuring that all microorganisms can be attacked one way or the other. A consequence of this feature of the immune system is that autoimmunity is a normal physiological state, with a consequent need for control mechanisms that prevent autoimmunity from progressing to overt pathology. The fact that these control mechanisms sometimes fail means that the immune system in most individuals retains the capacity for re-regulation. Accordingly, experimental studies, mainly in rodent models of autoimmune disease, have revealed several different ways whereby autoimmunity can be manipulated to prevent or cure autoimmune disease.1,,3

In order to accomplish this in humans, we need to know enough about what aspects of immunity contribute to disease and what environmental and genetic factors determine the evolution of autoimmunity and autoimmune disease. As there are many ways whereby autoimmunity can lead to a clinical phenotype, careful delineation of the heterogeneity of disease mechanisms will likely be required in order to direct specific preventive strategies to different individuals and different situations.

Autoimmunity and autoimmune rheumatic diseases

The term ‘autoimmune disease’ is most often used for conditions where autoimmunity is present, although in most cases it is not known whether the autoimmunity is causatively involved in disease pathogenesis or merely serves as a diagnostic biomarker. The potential for immunological intervention is much larger if an autoimmune reaction actually contributes to the disease. Although such a causative linkage formally requires that modulation of a specific immune reaction leads to amelioration of disease, there are cases where substantial evidence exists for causation and we will devote most of our further discussion to a few of these cases.

For rheumatoid arthritis (RA) (figure 1), almost all the evidence for an autoimmune condition is limited to the rheumatoid factor (RF) and/or antibodies to citrullinated protein antigen (ACPA) positive subset of disease (here also named ‘seropositive RA’), whereas it is still unclear how much autoimmunity is associated with ‘seronegative RA’.4,,6 The genetic risk factors, including HLA-DR alleles that indicate contributions from adaptive immunity, are mainly present for the ACPA-positive subset of disease.7,,10 In this subset of RA, the antibodies are (1) very specifically associated with one subset of the disease5 6; (2) some antibody reactivates occur before onset of disease11 12; and (3) immunity to citrullinated proteins/peptides may contribute to arthritis, as has been shown in rodents.13,,16

Figure 1

Outline of the longitudinal course of disease emergence and disease development in human autoimmune diseases, including the major subset of rheumatoid arthritis (RA), systemic lupus erythematosus (SLE) and cardiac heart block associated with SLE and Sjögren's disease. CV, cardiovascular.

Likewise, in systemic lupus erythematosus (SLE), autoimmunity to dsDNA is (1) very specific for one (the major) subset of the disease17; (2) present before disease onset18; and (3) certain anti-dsDNA antibodies can contribute to SLE-like symptoms in mouse transfer experiments.19 20 An even more compelling situation in SLE (and Sjögren's syndrome) is one where antibodies towards the immunodominant epitope (p200) of Ro52 are associated with neonatal heart block21 and where antibodies to this epitope can cause neonatal heart block in rodents in vivo22 and cause functional deficits in cardiac heart muscle cells in vitro.23

These cases may thus serve as examples of current targets for prevention and provide a focus for future research.

Genetic and environmental determinants of autoimmunity and autoimmune disease in subsets of RA and SLE

The case of RA

The first known step in the evolution towards ACPA-positive RA is when tolerance to certain citrullinated proteins/peptides is broken and autoimmunity to these proteins/peptides occurs.11 12 This immunity is preferentially triggered in individuals carrying certain HLA class II alleles, mainly HLA-DR04.5 6 8 9 24 Smoking and other airway stimuli such as silica dust act as environmental factors. A possible mechanism is induction of local expression of peptidylarginine deiminases which enhance formation of citrullinated proteins in the lungs.5 25 In some individuals the presence of ACPAs may be accompanied by increased presence of proinflammatory cytokines in the blood (as markers of systemic inflammation), still without clinical signs of disease.26 This stage may in turn be followed by non-specific signs of arthritis (without fulfilment of the 2010 ACR criteria of RA) and eventually develop into the kind of polyarthritis that even fulfils the 1987 criteria.27 28 Although it is possible that not all cases of ACPA-positive RA follow this gradual evolution scheme, there seems to be a lag time of a couple of years in the majority of cases in which pre-disease samples are available. This provides a potential ‘window of opportunity’ for disease prevention.

Also in SLE, antibodies towards dsDNA and Ro52 are present several years before onset of disease.18 Less is known about environmental factors and genes that determine the different steps towards disease in SLE than in RA. Similar to RA, most studies have focused on risk factors for full-blown disease. Smoking is the best known environmental risk factor,29 30 although smoking in SLE appears to have its main effects in the context of HLA-DR3 alleles.31 Sun exposure also appears to trigger disease and its flares,32 and there are a few classical examples of how drugs enhance the development of SLE-like symptoms.33 Intriguingly, the SLE induced by interferon α administration for the treatment of various malignancies34 35 fitted nicely with later observations that gene expression profiles in patients with SLE were compatible with genes upregulated by interferon α (for further discussions see below).36

Prospects for prevention: addressing environment, immunity and genes


The most obvious target for prevention for all diseases is environment and lifestyle (figure 2). Smoking is by far the most recognised risk factor for RA as well as for SLE. As mentioned above, it has recently been shown that smoking is a risk factor only for ACPA- and/or RF-positive RA, and not for seronegative RA.5 Also, it has been shown that smoking interacts heavily with HLA-DR shared epitope (SE) genes, conferring a very high risk for ACPA-positive RA in individuals carrying these SE genes.5 8 9 Interestingly, smoking also exerts its effects several years before onset, as former smokers are also at an increased risk for RA up to 20 years after smoking cessation with a gradually decreasing risk during this time.37

Figure 2

Schematic illustration of possible principles for prevention in autoimmune rheumatic diseases. CV, cardiovascular; RA, rheumatoid arthritis.

These data illustrate one of the main points of this review, namely that a preventive measure—elimination of smoking—will have distinct effects on the societal and on the individual level. An example of this is provided by a recent study in Sweden, a country with a relatively low prevalence of smoking. Based on statistical associations, one can calculate that smoking seems to be responsible for 22% of all RA, 33% of all ACPA-positive RA and for 55% of all cases of RA in individuals carrying double HLA-DRB1 SE alleles.38 It is possible that these numbers are even higher in countries with a higher frequency of smoking and where HLA-DR SE genes are prevalent. Other airway irritants such as silica dust have also been shown to be environmental risk factors for ACPA-positive RA and to be particularly dangerous in combination with smoking.39

Concerning protective lifestyle and environmental factors, even less is known. The only well reproduced protective factor is alcohol consumption. In addition to epidemiological studies in humans,40 41 alcohol consumption has been shown to ameliorate arthritis in rodent models of disease.42 There are now four independent cohort or case–control studies showing similar results with up to 50% risk reduction of RA associated with moderate alcohol intake compared with no intake.40 41 This information may be taken as a start for identification of mechanisms that cause the observed risk reduction associated with alcohol consumption. An exposure that has been speculated to infer an increased risk of RA is vaccination. An indication that this risk may be non-existent or small was recently provided in the EIRA case–control study where common vaccinations 5 years before onset of RA did not confer any increased risk for RA.43 Additional studies are, however, needed to definitely exclude an eventual influence of vaccinations on RA whenever in life the vaccinations are performed.

Smoking is also a major environmental risk factor in SLE although it has not been studied as extensively as in RA.29 30 Sunlight is otherwise the most well-known exposure that enhances disease flares,32 and most probably disease development too. Unfortunately, no quantification of the relative contributions of different lifestyles or different environments to disease development and to disease flares is available.

In the specific case of immunity to the p200 of Ro52—that is, the immunity specifically associated with the risk for neonatal cardiac heart block—nothing is known about environmental factors that may cause this autoimmunity, and this absence of information is typical of many autoimmune and potentially disease-inducing reactions.


The recognition that anticitrulline immunity may be causatively related to the onset of ACPA-positive RA provides a major opportunity to identify individuals at high risk for developing RA, and then to modify either the citrulline immunity itself or other factors that may be needed for an ACPA-positive individual to develop full-blown RA. Only three major studies have so far addressed this specific issue.

In the PROMPT study, individuals with undifferentiated arthritis (UA) not fulfilling the 1987 ACR criteria for RA were treated with methotrexate for 1 year, with the aim of evaluating whether this treatment could prevent development into a disease state as defined by fulfilling the ACR87 criteria for RA. Interestingly, methotrexate was significantly more effective than placebo in preventing progression to this disease state in ACPA-positive patients with UA whereas no difference between methotrexate and placebo was seen in the ACPA-negative group of patients with UA. On the other hand, fewer patients overall developed RA in the ACPA-negative UA subset.28 Although it is a relatively small study and has not been reproduced, PROMPT represents a very interesting principle for early, selective and preventive intervention against the progression towards development of RA. Thus, given the fact that no clear influences of methotrexate on the development of the ACPA response are known, one may consider the possibility that methotrexate modulated inflammation rather than specific immunity, but had its largest impact on an autoimmunity-driven variant of arthritis. This example indicates that it may be possible to interfere with proinflammatory events rather non-specifically at some critical and early stage of autoimmune disease development, and to achieve long-lasting results and maybe even a cure without specifically interfering with the autoimmunity.

In a recent pilot study in ACPA-positive patients with arthralgia, it was tried unsuccessfully to alter the disease course by two steroid injections.44

A few randomised short-term intervention studies have tried to alter the course of early UA (thus arthritis instead of arthralgia). In the STIVEA trial, three weekly intramuscular methylprednisolone injections in inflammatory polyarthritis of 4–10 weeks' duration postponed the need for disease-modifying antirheumatic drugs and resulted in a small increase in the proportion of patients in whom the disease had resolved after 12 months.45 Clearly, longer follow-up is needed. In the larger SAVE study in a similar patient population, a single 120 mg intramuscular injection of methylprednisolone was not effective in inducing remission or delaying development of RA.46 A tapered high oral dose of prednisone has not yet been evaluated in UA.

Two other powerful antirheumatic therapies—infliximab (a tumour necrosis factor (TNF) blocker) and abatacept (a T cell costimulation inhibitor)—have been tested47 48 in two small studies in patients with UA. Anti-TNF did not prevent progression to RA in the small pilot trial.47 The small trial with 6 months of abatacept was performed in ACPA-positive patients with UA.48 After 1 year, 12 (46%) in the abatacept group and 16 (67%) in the placebo group fulfilled the ACR criteria for RA. Although not statistically significant, this result argues for larger studies to investigate whether very early interventions directed towards the adaptive immune system may influence specific immunity and subsequent disease development.

Finally, it will also be possible to specifically modify specific autoimmune reactions. This is the most challenging possibility where we need more information on what specific arthritis-causing autoimmunity is present in single individuals. Thereafter, methods to modify this immunity can be evaluated in investigations that include observations of the clinical course of disease. The hope is obviously that the autoimmunity should be possible to alter and disease development ultimately prevented. Importantly, major studies are underway directed towards a detailed longitudinal analysis of the autoimmune and inflammatory events that precede the development of RA.49

In SLE and Sjögren's disease, neonatal cardiac heart block represents a particularly interesting case for prevention in the context of a pre-existing and disease-inducing immunity. Here, women with high antibody titres against the p200 epitope of Ro52 are those who almost exclusively carry the risk that their fetuses will develop heart block. As the damage to the fetal heart occurs in a rather narrow time window (approximately weeks 20–24 of gestation), it is possible to direct a very intensive monitoring of fetal heart function to these fetuses during the high-risk weeks.50 At early signs of developing lesions, corticosteroid treatment may stop further development of these lesions and thus function as a preventive in utero intervention against either fetal death or life-long pacemaker treatment that may otherwise be the result of uninterrupted action of the anti-Ro52 antibodies.51 The fact that the development of the lesions of the autoimmune attack occur over this limited time, and that the affected individual (the fetus) is ‘cured’ if the inflammation can be ameliorated during a limited time, illustrates how a well-timed intervention can be efficient if an antigen is only transiently exposed (which is the case for the antigens of fetal cardiac muscle cells). It may well be that this is the case in many other autoimmune situations as well.


Genetic polymorphisms represent the most precise information about factors that determine the risk for arthritis and, more specifically, the risk for different subsets of RA. Here, MHC class II genes are by far the dominating risk genes for ACPA-positive RA (with about 25% of the total risk), whereas other genes including PTPN22 provide less than 5% of the total risk.52 Knowledge of these genetic risk factors may be used as a basis for the development of protective interventions, as the genetic information may provide information about molecular pathways involved in specific subsets of disease. In RA there is one major example (mentioned above) which indicates how such information can be used in a preventive setting—that is, that smoking mediates the largest risk for ACPA-positive RA in individuals carrying HLA-DR SE genes. There are several examples from animal studies that suggest that it will be possible to influence specific molecular pathways to prevent arthritis in individuals who carry specific risk genes. One such case is represented by the Ncf1 gene in mice, where a genetic variant associated with a low oxidative burst is much more prone to develop arthritis than mice with another Ncf1 variant and higher oxidative burst.53 Administration of agents that enhance oxidative burst have been used successfully to treat and prevent arthritis in these mice54 and, interestingly, this treatment appears to be able to change the function of a genetically-determined mechanism from an arthritis-susceptible to an arthritis-resistant phenotype.

In SLE, many recent studies have focused on the effects of polymorphism in genes that regulate α-interferons and their functions.55 56 These studies have contributed to the interest in α-interferon antagonists as therapy for some cases of SLE (particularly those with a high production of α-interferons). It can be anticipated that individuals with a prominent interferon signature, where a combination of antibodies and genes infer a very high risk for SLE, might be treated with drugs interfering with this interferon pathway even before fulfilling the clinical criteria for SLE. Interestingly, the presence of a high interferon state or interferon ‘signature’ in peripheral blood may itself be a heritable trait found in subjects without disease.57 It is currently unclear how stable and predictive this is for the future development of clinical SLE. A major effort is underway to study this and related issues in a large cohort of unaffected sisters of patients with SLE in North America (see, and these data should provide more specific information on factors that predispose to the development of SLE, some of which may be targets for preventive therapies.

Concluding remarks and future perspectives

This mini review illustrates that the prospects for prevention against autoimmune rheumatic disease are developing from the classical population-based perspective aimed at broadly influencing known environmental and lifestyle factors into a situation where pre-disease events will be identified specifically in selected individuals. Obviously, both the population-based approach and the individual approach have to be used in our efforts to diminish the suffering from autoimmune diseases. However, given the multitude of different molecular pathways towards autoimmune diseases, the multitude of different genetic and environmental risk factors and the emerging potentials to modify specific immune reactions, we need to use the potentials for prevention in different ways in different subpopulations of RA, SLE and other autoimmune diseases. This thinking parallels the experience in animal models of autoimmune diseases where it is possible to prevent disease development in many of these models once the knowledge of the disease mechanisms is good enough. Furthermore, many interventions that are made in the pre-disease phase in such animal models are able to completely abrogate further development of disease, whereas the same intervention may have much more limited effects when given when disease is already present.

The emerging opportunities for disease prevention inevitably raise questions of large practical and ethical importance. Which populations should be subject to this study and, in particular, when, how and in which populations is it acceptable to perform pharmaceutical interventions? These are issues that will be increasingly important with increasing abilities to predict future disease development based on genetics and predictive biomarkers. At present we believe that pharmaceutical pre-disease interventions should be reserved for individuals who both have some symptomatology and are at high risk for further disease development. In the future this may obviously change with increasing precision in predicting disease and with access to interventions that may provide a higher chance of prevention compared with later chances of cure.



  • Funding The research programmes that have provided data to this review from the authors' laboratories were supported by the EU-funded programs AutoCure and Masterswitch.

  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer-reviewed.