The nexus between atopic disease and autoimmunity: a review of the epidemiological and mechanistic literature


  • The views expressed in this report are the personal opinions of the authors and are not the official opinion of the US Food and Drug Administration or the Department of Health and Human Services.

R. L. Rabin, Office of Vaccines Regulation and Research, Center for Biologics Evaluation and Research, US Food and Drug Administration, 29 Lincoln Drive, Building 29, Room 203A, Bethesda, MD 20892-4555, USA.


There has been considerable interest in defining the relationship between the expression of allergic and autoimmune diseases in populations of patients. Are patients with autoimmune disease ‘protected’ from developing allergic (immunoglobulin E-mediated) diseases? Does the establishment of an atopic phenotype reduce the risk of the subsequent development of autoimmune diseases? Although there are clinical studies addressing this question, methodological problems, particularly in identification of atopic subjects, limits their usefulness. Moreover, an immune-based explanation of the observed epidemiological findings has relied on a paradigm that is currently undergoing increased scrutiny and modification to include newly defined effector cell subsets and the interaction between genetic and environmental factors, such as early endotoxin or mycobacterial exposure. To address this question, we reviewed a series of clinical reports that addressed coincidence or co-prevalence of atopy with four autoimmune diseases: psoriasis, rheumatoid arthritis, multiple sclerosis and type I diabetes mellitus. We present a model whereby active T helper type 1 (Th1) inflammation may suppress the development of atopy, and atopy may suppress the severity but not necessarily the onset of autoimmunity, and then discuss our model in the context of mechanisms of adaptive immunity with particular reference to the Th1/Th2 paradigms. Because the ultimate goal is to ameliorate or cure these diseases, our discussion may help to predict or interpret unexpected consequences of novel therapeutic agents used to target autoimmune or atopic diseases.


There has been considerable interest in defining the relationship between the expression of allergic and autoimmune diseases in populations of patients. Are patients with autoimmune disease ‘protected’ from developing allergic [immunoglobulin E (IgE)-mediated] diseases? Does the establishment of an atopic phenotype reduce the risk of the subsequent development of autoimmune diseases? Although there are clinical studies addressing this question, methodological problems, particularly in identification of atopic subjects, limits their usefulness. Moreover, an immune-based explanation of the observed epidemiological findings has relied on a paradigm that is currently undergoing increased scrutiny and modification to include genetic predisposition and its interaction with environmental factors, such as early endotoxin or mycobacterial exposure [1,2].

Until recently, the adaptive cellular immune response has been characterized broadly as being polarized in one of two directions: type 1 or type 2 [3]. Type 1 responses, directed by T helper type 1 (Th1) CD4+ T cells and identified by the signature cytokine interferon (IFN)-γ, are considered to protect against infections by intracellular pathogens [4], and have been incriminated in the pathogenesis of autoimmune diseases such as rheumatoid arthritis (RA), multiple sclerosis (MS) and type 1 diabetes mellitus [5,6]. By contrast, type 2 responses, directed by Th2 CD4+ T cells and identified by the signature cytokines interleukin (IL-4), IL-5 and IL-13, are considered to protect against helminthic infections [7] and to play major pathogenic roles in allergic diseases and asthma [8]. Reciprocal counter-regulation of Th1 and Th2 cells [9] predicted that Th1-type autoimmune diseases and Th2-mediated allergic diseases would occur in mutually exclusive populations of patients. However, recent observations have challenged the validity of the long-standing Th1/Th2 paradigm [10], and a far more complex story explaining the immunological basis of cellular immune-mediated host defence and the pathogenesis of autoimmune and allergic diseases is emerging. The new paradigm identifies additional lymphocyte subsets, such as Th17 T cells, regulatory T cells (Treg) and novel soluble factors. These help to explain experimental observations not predicted by the Th1/Th2 paradigm, and provide a new prism through which to examine the intersection of autoimmune and allergic disease [10]. In this paper, we first take a critical look at the epidemiological literature bearing on the relationship between allergic and autoimmune diseases, and then examine the findings in the context of current principles that underlie immune mediated tissue damage.

Epidemiological reports

To review the epidemiological literature addressing this issue, multiple strings were used to search through the National Library of Medicine database (Pubmed). The term ‘autoimmunity’ as well as each autoimmune disease was searched in combination with ‘allergy’, ‘atopy’ and ‘asthma’ (e.g. MS and allergy, MS and atopy, MS and asthma) with the qualifier ‘human’. For purposes of this study, we used ‘atopy’ to mean a history of atopic dermatitis (AD), allergic rhinitis and/or asthma. While not all asthma is truly atopic − i.e. IgE-mediated − Th2-mediated adaptive responses are associated generally with asthma regardless of the levels of allergen-specific IgE.

The reports used for this review, along with the method of the study, can be found in Table 1. As expected for two sets of diseases that arise as a consequence of environmental and polygenic factors, none are perfect. Studies that were derived from national registry databases and/or questionnaires allow for a large number of study subjects, but rely on historical reports from clinicians or the subjects rather than objective diagnostic criteria. They also risk recall bias that would predispose towards a type II error − failure to infer exclusivity between autoimmunity and atopy when indeed it exists. Case–control studies include objective data [e.g. laboratory tests, skin prick tests (SPT), spirometry] but risk selection bias in the choice of control subjects. Most studies were not powered for sample size. While studies with very large numbers of subjects (> 10 000) would have survived power analysis, some of the smaller studies are clearly underpowered. If the statistical evidence did not support reported differences in the experimental and control groups [i.e. P-values > 0·05, or confidence intervals (CI) that overlap 1·0], we rejected those conclusions. Not withstanding these hurdles, the reports segregated largely into comparisons of the incidence of atopy with three autoimmune diseases that are considered historically to be Th1-mediated: MS, RA and type 1 diabetes mellitus.

Table 1.  Clinical reports comparing the incidence or prevalence of autoimmunity versus atopy and their method of study.
1st author, reference no.YearIndex diseaseControl populationNo. study subjectsNo. control subjectsMethodsConclusionComments
  1. AD, atopic dermatitis; ACD, allergic contact dermatitis; AR, allergic rhinitis; BA, bronchial asthma; CC, case–control; CD, Crohn's disease; CR, chart review; DM, diabetes mellitus (unspecified); DTH, delayed type hypersensitivity; FH, family history; Hx, history; IDI, intra-dermal skin test for type I hypersensitivity; IDIV, intra-dermal skin test for type IV hypersensitivity; IgEs, specific serum IgE measurements; IgEt, total serum IgE measurements; n.d., no differences between study and control populations; OR, odds ratio; Q, questionairre; RA, rheumatoid arthritis; RDz, rheumatic diseases; SPT, skin prick testing; T1D, Type I diabetes mellitus; T2D, Type 2 diabetes mellitus; Urt, urticaria; ↑, Increased; ↓, Decreased; Ig, immunoglobulin.

Type 1 diabetes mellitus (T1D)
Swern [11]1932T1D, T2DNone4000 CohortInverse associationInteresting historical perspective
Siegel [12]1954T1D, T2D,Non-DM40227Hx, IDI; Cohortn.d.Mixed T1D + T2D
Helander [13]1958T1D, T2DBA32363151Cohortn.d. in older ↓ subjects; coincidence in younger subjectsMultiple refs of previous studies
Hermansson [14]1971T1DSiblings and population controls128310Medical records, physician exams; cohortInverse association of probands versus population controls, and of siblings of probands versus population controls 
Stromberg [15]1995T1DSchoolmates6172CR, Q, SPT, IDIV, IgEs; CCn.d.Atopy rates of 80–90%; better DTH in T1D subjects
Douek [16]1999T1DSiblings157173Q; CCHigher wheeze control versus DM 
EURODIAB Substudy 2 [17]2000T1DPopulation-based controls12043606T1D and schools, clinics registers; Q; CCInverse relationship T1D and atopySee text
Kero [18]2001CD, T1D, RANational registry94, 181, 7759 867CohortPositive association between CD and asthma, and RA and asthmaLow overall prevalence of asthma 3·3%
Olesen [19]2001T1DCase–controls9289732Q; CCLess AD in T1D if AD before T1D; after T1D, no difference 
Douek [20]2002T1D, AD, ARSiblings206209Q; CCNo association with AR, inverse association between AD and T1D 
Mattila [21]2002T1DSiblings and population controls306506, 406Q; CCWeak inverse association to ‘animal dust’ 
Meerwaldt [22]2002T1DPopulation survey555777Q; CCn.d.Authors state difference, but not statistically significant
Cardwell [23]2003T1Dn.a.n.a.n.a.Meta-analysisSlight inverse association between asthma and T1D 
Multiple sclerosis (MS)
Frovig [24]1967MSHospital-based practice, Norway6140CR; CCHigher incidence of allergy among subjects with MS 
Alter [25]1968MSHospital-based practice (MN, USA)3672Q, CCNo difference in incidence of allergy in the MS versus control 
Cendrowski [26]1969MSPolish epidemiological survey300300CCNo difference between the two groups in the frequency of allergic diseases 
Khurshed [27]1976MSHospital-based practice, MN, USA3640CCNo difference serum Total IgE between MS and control 
Warren [28]1981MSHospital-based practice100100CCHigher incidence of DM in MS population 
Casetta [29]1994MSItalian epidemiological survey104150CCSlightly higher incidence of allergy in MS versus control 
Oro [30]1996MSNon-inflammatory neuroconvulsive disorders3518Q, IgEt, IgEs; CCDecreased atopy (IgEs, MAST score, symptom scores) 
Neukirch [31]1997MSFrench epidemiological survey6106926CCLower incidence of allergy in MS subjects 
Solaro [32]2001MSItalian epidemiological survey312312CCNo difference in incidence of allergy in the MS versus control 
Tremlett [33]2002MSDatabase controls, age- and sex-matched320320CohortDecreased asthma in MS 
O'Driscoll [34]1985RACase–controls266; 4040Q, SPT, IgEt, IgEs; CCRA in atopics = RA in general pop. Atopy in RA = atopy in control 
Verhoef [35]1998RARA and non-RA rheumatic diseases304339Consecutive clinic patients; Q, IgEs, IDI; Cohort↓ AR in RA pts ↓ RA severity in RA alone versus RA + ARSee text
Hilliquin [36]2000RACase–control173173Q; CCInverse correlation between RA and atopyDid not define atopy or criteria for atopy
Redwaleit [37]2002RA, ASHospital staff487, 248536Q; CCInverse RA and eczema, AR; those with atopy before RA, less severe RA 
Olsson [38]2003RARandom selection263541Q; CCNo statistically significant differences 
Simpson [39]2002All atopic Dz‘Any Th1 disease’18 3077881Scottish registry, total 252 538; cohortTh1 and atopy correlatePsoriasis and eczema concurrence account for the correlation
Sheikh [40]2003Lumped Th1 diseases; allergic diseasesUK Nat'l Survey1938; 42418 071, 6869Q, SPT; CohortNo differences 
Tirosh [41]2006AsthmaAll subjects drawn from military recruits450 000 ∼∼40 000Medical record review; cohortAsthma protects against autoimmune diseaseWell-documented study

Rheumatoid arthritis and atopy

Because IL-12 and IL-18 are expressed in the synovial lining in inflamed joints, RA is considered a Th1 disease [42–45]. While RA is associated with human leucocyte antigen (HLA) class II alleles [46,47] and polymorphisms that encode tumour necrosis factor (TNF) and IL-10 [48,49] genes, there are no known genetic associations with Th1 cytokines or their associated signalling molecules. RA was linked recently to a single nucleotide polymorphism (SNP) in SLC22A4, an organic cation transporter that is highly expressed in inflamed joints of mice with collagen-induced arthritis [50]. Polymorphisms in SLC22A4 have also been linked to Crohn's disease, also associated with Th1 mechanisms [51].

O'Driscoll and colleagues compared 40 subjects with RA to age-matched controls and found a similar incidence of atopy as determined by history or SPT. They also found that two of 266 atopic subjects followed in an allergy clinic had RA, a prevalence rate similar to that for RA in the general population [34]. Olsson and colleagues [38,52] found no clear associations between RA and atopy on a retrospective analysis of 281 Swedish patients treated for RA and 507 population-based controls. Although the prevalence of autoimmune and AD in the respective groups of subjects and control populations did not differ, the studies were small and probably underpowered.

In contrast, two questionnaire-based studies suggest an inverse correlation of the prevalence of self-reported atopy among cohorts who meet the American Rheumatism Association criteria for RA. Hilliquin and colleagues [36] compared 173 consecutive RA patients from their clinic to age- and sex-matched case controls and found, independent of treatment for RA, a decreased incidence and point prevalence of atopy. Rudwaleit et al. [37] used questions extracted from the European Community Respiratory Health Survey and the International Study of Asthma and Allergies in Children (ISAAC) protocols to analyse 728 subjects with RA from a database of university out-patient clinics and 900 controls comprised of a combination of hospital staff members and elderly patients with osteoporosis. In this study, prevalence of ‘hay fever’ and AD in general was decreased in the RA subjects by about 45% and 35% respectively. Most interesting was a subanalysis that found lower severity of RA among those subjects who presented with AD prior to the onset of RA.

These two studies are supported by a Dutch retrospective cohort study of 643 consecutive rheumatology patients, of whom 304 met the American College of Rheumatology criteria for RA. Allergic rhinitis was confirmed by SPT, and was less prevalent in the RA subjects (4% versus 8%) with a significant decrease in relative risk (0·48, 95% CI 0·25, 0·92, P < 0·05). The subjects with both RA and allergic rhinitis had lower RA clinical scores, erythrocyte sedimentation rates (ESR) and C-reactive protein levels (CRP). Furthermore, during the pollen season, peripheral blood mononuclear cells (PBMC) from subjects with both RA and allergic rhinitis secreted less IFN-γin vitro compared with PBMC from subjects with RA alone. Once the pollen season ended, IFN-γ expressed by PBMC from the dually affected subjects rose to levels similar to those from subjects with RA alone [35].

Although not without flaws, these studies taken together suggest that first, atopy is decreased in patients with RA compared with controls, and second that among those patients with both diseases, the severity of RA is decreased.

Multiple sclerosis and atopy

Multiple sclerosis is a chronic autoimmune disease in which the antigenic target is central nervous system myelin. Both CD4 and CD8 T cells are found in MS lesions, as is expression of IFN-γ and TNF. However, CD4+ Th1 cells have been incriminated traditionally as the major orchestrators of the acute inflammatory process. The strongest genetic linkages are with HLA class II haplotypes [53]. While several questionnaire-based studies showed no association between MS and atopy [24–29,32], two recent studies demonstrate an inverse relationship between them. Oro and colleagues [30] studied 35 MS patients and 18 control patients who were followed for non-inflammatory neuroconvulsive disorders. All patients answered questionnaires, and all but 11 MS patients had blood drawn for total and allergen-specific IgE. Allergy symptom scores and the number of positive allergen-specific IgE tests were lower among the patients with MS. Tremlett et al. [33] evaluated the hospital records of 320 Welsh patients and matched controls and found a lower prevalence of patients who received the diagnosis and a prescription for treatment of atopic asthma, but not AD among patients with MS.

Whether atopy may affect the clinical course of MS similar to its suggested effect on RA is difficult to determine because of the widespread use of sedating anti-histamines until 1985, when a non-sedating anti-histamine was licensed. While both sedating and non-sedating of anti-histamines are inhibitors of histamine receptor 1 (HR1), only the sedating anti-histamines cross the blood–brain barrier easily [54]. Multiple reports demonstrate a role for HR1 in generating Th1 responses [55–57] and HR1 antagonists decrease the severity of experimental autoimmune encephalomyelitis, the animal model for MS [58]. Recently, Alonso and colleagues reported a case–control study that showed an inverse correlation between MS risk and treatment of allergy with sedating, but not non-sedating anti-histamines [59]. Thus, while we may conclude an inverse relationship between MS and atopy, it must be confirmed by studies that take the use of sedating HR1 anti-histamines for atopy into account.

Type I diabetes mellitus and atopy

Type I diabetes mellitus is an autoimmune disease caused by cell-mediated immunity targeting pancreatic islet beta cells. In animal models, Th1 cytokines such as IFN-γ and IL-12 are associated with islet cell destruction [60,61], and Th2 cytokine-secreting T cells are protective [62]. Among many known autoantigens, only pro-insulin and insulin are islet-cell specific [63]. Specific T cells stimulated with pro-insulin peptides secrete IFN-γ, while cells from many controls secrete the regulatory cytokine IL-10 [64].

Type I diabetes mellitus is linked strongly with HLA class II haplotypes, some of which predispose to disease and others that protect [65]. While the HLA linkages are by far the strongest immunological risk factor, type I diabetes mellitus is also associated with polymorphisms in non-coding regions of two cytokine genes, IFN-γ[66] and IL-12 p40 [67], and an amino acid substitution in cytotoxic T lymphocyte-associated molecule 4, a co-stimulatory protein essential for attenuating cell-mediated immunity [68–70].

The earliest observations of an inverse association between diabetes mellitus and allergic diseases prompted many reports throughout the mid-20th century. While two of these are included in Table 1 for historical reference [11], only those addressing subjects with type I diabetes mellitus are considered here. A recent meta-analysis concluded a ‘small but significant’ decrease in asthma prevalence in children with type I diabetes mellitus [23]. Studies of interest are discussed below.

Douek and colleagues used the ISAAC questionnaire to compare 157 probands with type I diabetes mellitus to 173 unaffected siblings, and found that fewer type I diabetes mellitus subjects than controls had wheezed at all within 12 months of the study, or had multiple or speech-limiting episodes of wheezing [16]. The EURODIAB ACE Substudy 2 study group was comprised of eight centres in eastern and western Europe and reported data collected by interviews from five of the centres and by questionnaire from the other three. Probands with type I diabetes mellitus were compared with population-based controls. AD, and asthma in particular, were decreased in children with type I diabetes mellitus. Two sites in the United Kingdom, where the incidence of atopy was the highest, accounted for 40% of the diabetics and may have contributed disproportionately to the study as a whole. Furthermore, only the western European centres demonstrated the inverse relationship between type I diabetes mellitus and atopy, and the incidence of atopy was higher in the type I diabetes mellitus than controls in the Bulgarian cohort. Because atopy preceded type I diabetes mellitus, the authors' inference that atopy may protect children from type I diabetes mellitus [17] may reflect the greater prevalence of atopy in western populations.

Stromberg and colleagues [15] compared 61 Swedish children with type I diabetes mellitus to age- and sex-matched controls, and found no differences in history, SPT and total and specific IgE. Similarly, data from a 1987 Finnish registry [18], a British health and nutrition survey [40] and a Dutch cross-sectional survey [22] found no differences in the cumulative incidence of asthma between type I diabetes mellitus patients and controls.

Perhaps the most provocative report is a retrospective case–controlled comparison of 928 Danish children with type I diabetes mellitus to a random sample of 10 000 population-based controls that found a lower cumulative incidence of AD in the diabetics [19]. Uniquely, this report showed the inverse correlation only among those diabetics who had AD prior to the onset of type I diabetes mellitus. Diabetics in whom the onset of AD followed type I diabetes mellitus were no different from controls. This study highlights a unique feature of type I diabetes mellitus relative to autoimmune diseases such as MS and RA: the pancreatic islet beta cells are diminished when type I diabetes mellitus presents [71], such that the inflammation, while not completely resolved, has diminished to a level insufficient to affect other responses. The concept of waned type 1 inflammation after clinical presentation of type I diabetes mellitus is supported by high serum levels of IL-18, IFN-γ and CXCL9 (an IFN-γ inducible chemokine) in newly diagnosed diabetics compared with those with long-standing disease [72] and low-risk controls [73]. Alternatively, defective regulatory mechanisms may make those with active type I diabetes mellitus equally prone to subsequent atopic or autoimmune disease.

Consistent with the interpretation that active type 1 inflammation protects against clinical presentation of AD is a recent analysis of almost 500 000 Israeli adults at the time of their enrolment into military service between 1980 and 2003. The diagnosis of asthma was confirmed by spirometry. Asthma prevalence and incidence were correlated inversely with a number of autoimmune diseases that were also diagnosed at enrolment [41], suggesting that those with newly diagnosed autoimmunity were less prone to have asthma.

Taken together, the studies that generalize the least and support the diagnoses with objective clinical and/or laboratory data suggest a model whereby Th1 inflammation suppresses the development of atopy [19], while Th2 inflammation suppresses the severity and perhaps the onset of some autoimmune diseases [35].

Mechanisms of type 1 and type 2 counter-regulation

As noted previously, the reciprocal counter-regulation of type 1 and type 2 pathogenic mechanisms led to the hypothesis that diseases mediated by these respective processes might not occur or would be less severe in the same patients. Evidence for the Th1/Th2 paradigm was provided originally by the observation that signature cytokines and transcription factors are both necessary and sufficient for polarization in one direction or the other. IL-12 is central to developing Th1 responses [74] and is secreted primarily by dendritic cells (DC) in response to infection with intracellular pathogens, or stimulation of cell surface proteins such as those in the Toll-like receptor family [75]. The central role for IL-12 in the paradigm is emphasized by the biological activity of IL-18, which enhances Th1 development in the presence of IL-12, and Th2 development in its absence [76]. The finding that IFN-γ in the absence of IL-12 may similarly enhance Th2 development was surprising [77], as IL-12 shares no signalling pathways with IL-18.

The transcription factor T-bet (T-box, expressed in T cells) is expressed in Th1, but not Th2 cells [78]. T-bet transactivates the IFN-γ gene promoter [74]. T-bet itself is expressed after a sequence of events that consists of IL-12 secretion, activation in the T cell of the transcription factor signal transducer and activator of transcription-4 (STAT-4), followed by the secretion of IFN-γ by the responding T cell, and then activation of STAT-1. STAT-1 stimulates transcription of T-bet which initiates an autostimulatory positive feedback loop of subsequent IFN-γ secretion and STAT-1 activation [74]. The importance of T-bet and Th1-polarized CD4+ T cells in protecting against AD was demonstrated by studies conducted in T-bet knock-out mice. These animals developed spontaneously a pulmonary disease that recapitulated the clinical and pathological features of asthma [79]. A recent report associates a SNP in the 3′-untranslated region of the T-bet gene with asthma [80].

Interleukin-4 is the principal cytokine leading to the development of Th2 cells. An initial source of IL-4 may be the natural killer T (NK T) cell, which is essential for the development of asthma in a murine model [81]. GATA binding protein 3 (GATA-3) is a major Th2 regulatory factor [74] that is necessary for polarization towards the Th2 phenotype, in which cells express IL-4 and IL-5 and fails to express the β chain of the IL-12 receptor (IL-12Rβ) [82,83]. GATA-3 expression is repressed by the combination of IL-12 and IFN-γ and augmented by IL-4. Activated GATA-3 enhances transcription of its own gene, and represses IFN-γ and STAT-4 expression [84,85]. In the context of the differentiation of naive T cells, GATA-3 dominates T-bet independently of IL-4 and thus Th2 cells emerge. Once the T cell response has developed, however, T-bet dominates GATA-3 independent of IFN-γ[86,87].

Thymic stromal lymphopoeitin (TSLP) is the latest addition to the list of Th2 polarizing molecules [88]. Appreciated originally as a molecule produced under homeostatic conditions in the thymus, it is now clear that mucosal and skin epithelial cells also produce TSLP. TSLP conditions DC to express OX40L [89], which when presenting antigens to naive CD4+ T cells, promotes their differentiation to Th2 cells [90]. High concentrations of TSLP can be found in the bronchial epithelium and submucosa of allergic asthmatics and in the skin of patients with AD [91]. It has been proposed that TSLP-DC conditioned Th2 cells in these environments express an inflammatory phenotype characterized by the production of large amounts of the classical Th2 cytokines and the proinflammatory cytokine TNF, but not the anti-inflammatory cytokine IL-10 [89]. Recent studies conducted in TSLP-deficient and TSLP-transgenic mouse models support strongly a major role for TSLP in the development and maintenance of AD [92,93].

On the other hand, TSLP-DCs also may lead to the generation of a population of a subset of anti-inflammatory Th2 cells that are distinguished by their production of the classical Th2 cytokines and IL-10, but not TNF [89,91]. For example, it appears that intestinal epithelium secretion of TSLP provides a homeostatic mechanism for counteracting the Th1 properties of commensal gut organisms through the generation of such ‘tolerogenic’ Th2 cells [94]. Of note, Crohn's disease has been reported to be associated with a loss of such TSLP-mediated homeostatic control [94].

The above discussion highlights current understanding of the cellular and molecular basis of the Th1 and Th2 paradigm. From the mechanistic standpoint, this discussion helps to explain why atopic and certain autoimmune disorders have been categorized as Th2- or Th1-mediated disorders.

Challenges to the Th1/Th2 paradigm

T helper type 1 cells in the pathogenesis of asthma and AD

Several studies challenge that the Th1/Th2 paradigm has served as a useful basis for thinking about the pathogenesis autoimmune and AD. For example, Th1 cells can transfer airway hypersensitivity in an animal model of allergic asthma [95,96] and mediate inflammation during the chronic stages of AD (reviewed in [97]). In addition, it has been reported recently that Th1 cytokine secretion was associated with the size of immediate hypersensitivity skin test to allergen and bronchial hyperresponsiveness in a large cohort of affected children [98]. Furthermore, while clinical trials support a role for IFN-γ in the pathogenesis of MS [99] and TNF in RA and Crohn's disease [100], others indicate that TNF inhibitors might exacerbate MS and other demyelinating diseases [101,102]. Also, experimental models of MS and RA demonstrated that TNF and IFN-γ might have protective effects if administered after the clinical appearance of tissue destruction [103,104]. Thus, it appears that Th1 cytokines may either be pro- or anti-inflammatory in the same autoimmune disease, with the outcome dependent upon when, in the course of pathogenic events, they are introduced.

Contributions of IL-17 to immune-mediated inflammation

The emergence importance the Th17 lineage represents the latest challenge to the inviolability of the Th1/Th2 paradigm [105,106]. Th17 cells secrete IL-17A and IL-17F, two of six members of the IL-17 cytokine family (reviewed in [107]). These cells differentiate along a pathway that is totally independent of Th1 and Th2 cells, expressing neither the Th1 transcription factors, T-bet and STAT-4, nor the Th2 transcription factors, GATA-3 and STAT-6 [97]. Th17 cells different from naive CD4+ T cells in response to IL-6 secreted by stimulated DC and transforming growth factor-β secreted by Treg[108,109], and are maintained by IL-23. Through expression of their signature cytokines IFN-γ and IL-4, Th1 and Th2 cells antagonize the differentiation of Th17 cells [110]. Th17 cells play a pivotal role in the pathogenesis of several autoimmune diseases, largely by induction of cytokines and chemokines that promote chemoattraction of inflammatory cells. Moreover, it is becoming clear that Th17 cells contribute to the pathogenesis of the chronic inflammation seen in several autoimmune diseases considered traditionally to be mediated by CD4 Th1 cells [111,112].

Blurring distinctions further between mechanisms contributing to autoimmunity and AD, IL-17 may contribute to the pathogenesis of asthma. Expression of IL-17 is increased in the airways of asthmatic patients and it can induce bronchial fibroblasts to produce IL-6 and IL-11 [113]. The concentration of IL-17 in sputum from asthmatics is increased, and the levels correlate with the degree of bronchial hyperresponsiveness [114]. IL-17 is also over-expressed in nasal polyps in association with increased collagen deposition [115]. These results highlight a possible contribution of IL-17 to the structural changes in the airway seen in chronic asthma. The production of IL-8 by human airway smooth muscle cells is stimulated by IL-17, suggesting a role for this cytokine in neutrophilic forms of asthma [116]. While the responsible IL-17 family members were not specified in these reports, the reagents used to identify responses induced by IL-17 suggest that Th17-derived IL-17 cytokine family members were very probably the responsible factors.

To make matters even more confusing, IL-17E is emerging as a factor in the development of allergic inflammation. IL-17E, which is now called IL-25, is expressed by murine Th2 cells and targets a novel innate immune cell to produce IL-4, IL-5 and IL-13 and the CC chemokines, CCL5 and CCL11 (eotaxin). IL-25 induces many of the features of allergic inflammation in mice, and suppresses both Th1- and Th17-mediated inflammation in some experimental models of autoimmune disease [107]. The differentiation of and interrelationships between these various CD4 T cell subsets and the potential for adverse consequences to therapeutic intervention are depicted in Fig. 1.

Figure 1.

Differentiation of and nexus between CD4+ T cell subsets. The figure depicts the diversity of CD4+ T cell subsets and their multiple modes of interaction. Green arrows signify positive influences and red lines inhibitory influences. All lines represent the activity of individual cytokines except in the case of T helper type 17 (Th17) cell inhibition by interferon-γ and interleukin-4 where both cytokines must act in concert. The crossed green arrows with the question mark below indicate the potential for a Th1-specific immunomodulatory agent to reveal or exacerbate a Th2 disease or vice versa. IL, interleukin; TGF, transforming growth factor; IFN, interferon.

Emerging role of NK T cells in the pathogenesis of asthma

Whereas Th2 cytokines are still considered to be principal effector molecules in the pathogenesis of asthma, it is no longer clear that CD4+ T cells are the major source of these proinflammatory molecules. As noted above, studies in mouse models of allergic asthma had indicated that NK T cells were required for the development of allergen-induced airway hyperreactivity [81]. NK T cells express features of both traditional CD4+ T cells and NK cells and express an invariant type of T cell receptor that recognizes glycolipid antigens in the context of CD1, a set of non-polymorphic major histocompatibility complex class I proteins. While NK T cells may express CD4, CD8, or neither CD4 nor CD8, it is primarily the CD4 expressers that secrete both Th1 and Th2 cytokines rapidly upon activation [117]. Recently, Akbari and colleagues made the striking observation that CD4+ NK T cells, and not classical CD4+ T cells, represented the dominant IL-4- and IL-13-secreting cell population in the lungs of patients with moderate to severe persistent asthma [81,118]. However, two subsequently published studies showed that while the percentage of NK T cells found in bronchoalveolar lavage fluid may be elevated, the magnitude of that increase is much lower than reported by Akbari et al. [119–121]. Thus, while NK T cells appear to contribute to the pathogenesis of mouse models in asthma, their importance in human asthma remains to be established.

The consensus is that NK T cells are protective in experimental animal models of autoimmune disease as well as in several types of human autoimmune disease [122]. It is considered that the NK T cells, which secrete a Th2 profile of cytokines, are those that ameliorate autoimmunity [117]. Thus, the data at hand raise the possibility that the Th2-like NK T cells may contribute to the pathogenesis of asthma, and have a salutary immunoregulatory function in autoimmune disease.

In summary, whereas the Th1/Th2 paradigm served as a useful early framework for considering the relationship between autoimmune and AD, the emergence of newly characterized proinflammatory and Treg cell subsets help to explain why it can no longer be generally applied to models of the pathogenesis of autoimmune or AD. Thus, it is not surprising that the autoimmune and AD reviewed here do not adhere consistently to the yin–yang relationship predicted by this once-hallowed construct.


From a mechanistic standpoint, the Th1/Th2 paradigm suggested that diseases mediated by these two effector CD4+ T cell populations would segregate in distinct patient populations. Despite the flaws of the epidemiological studies referenced, the complex genetic and environmental factors that underlie atopy and autoimmune diseases, and modifications of the Th1/Th2 paradigm to account for newly discovered subsets of T cells such as Treg and Th17 cells [114], our review supports Th1/Th2 segregation in some pairs of Th2-mediated AD and putative Th1-mediated autoimmune disorders. Because of this complexity, however, it is not surprising that this inverse relationship is neither consistent nor straightforward.

Some of the epidemiological reports suggest a model in which ongoing Th1 inflammation inhibits atopy. For example, as insulitis diminishes, type I diabetes mellitus becomes clinically manifest and protection by Th1 inflammation against development of AD abates. Two reports of AD following anti-TNF therapy for RA [123,124] suggest that pharmacological attenuation may have the same effect. Whether or not atopy protects against autoimmunity, there is objective evidence (ESR, CRP) that, at least in RA, it can attenuate the clinical expression of autoimmunity. Because mature Th1 responses cannot be switched, transient attenuation of RA in atopic subjects during the pollen season may be due to suppressed Th1 inflammation, rather than an actual change in phenotype. For example, IL-4 secreted by mast cells may not only trigger their apoptotic death, but also suppress secretion of TNF and IL-1β by synovial macrophages [125–127].

Even before novel selective therapeutic agents enter clinical trials, the complex relationship between Th1 and Th2 inflammation outlined here must be considered, particularly in light of recent studies that suggest that increasing prevalence of both autoimmune and AD are due to dysfunctional regulatory mechanisms such as Treg cells [128] that may control either type of pathology. In this context, a therapy that targets Th1- or Th2-specific responses may reveal underlying atopy or autoimmunity respectively. While proof of concept studies for novel therapeutic agents are performed generally in rodent strains that are susceptible for the disease under investigation, it may be worthwhile to determine, for example, whether an anti-Th2 immunomodulatory agent that treats asthma lowers the threshold for clinical presentation of DM in the non-obese diabetic mouse. Conversely, how might a potential anti-Th1 immunomodulatory therapeutic for MS affect the threshold for allergic sensitization? Finally, as clinical testing of selective therapeutics for autoimmune and AD progress towards larger pivotal trials, it would be prudent to monitor for manifestations of the onset of atopy or autoimmunity respectively, and continue post-marketing monitoring after licensure of the new drug.