Toward Precise Forecasting of Autoimmune Endocrinopathy

Simon H. S. Pearce and Nicola J. Leech

Institute of Human Genetics (S.H.S.P.) and School of Clinical Medical Sciences (N.J.L.), University of Newcastle, Newcastle upon Tyne, NE1 3BZ United Kingdom

Address all correspondence and requests for reprints to: Dr. Simon H. S. Pearce, Institute of Human Genetics, Centre for Life, Central Parkway, Newcastle upon Tyne, NE1 3BZ, United Kingdom. E-mail: s.h.s.pearce{at}ncl.ac.uk.

The practice of medicine is moving from an era of treating disease to disease prevention. Thus, it is not surprising that the scientific literature is currently filled with studies of organ-specific autoantibodies as markers of autoimmune endocrine disease. The association of autoantibodies with autoimmune thyroid disease (AITD) was established more than 40 yr ago, and yet their role in routine clinical practice remains ill-defined. Individuals can be identified as being at high risk of certain autoimmune diseases by virtue of personal or family history. This, combined with the appearance of autoantibodies long before the development of organ failure, means in principle that the autoimmune endocrinopathies have many characteristics suited to prediction and prevention. The current practice remains, however, to intervene with hormone replacement at the point of almost complete organ failure, leaving little or no scope for disease modification. This is in contrast to the approach in many other autoimmune disorders (e.g. rheumatoid arthritis).

To be worthwhile, a screening program must meet several criteria: there needs to be an effective intervention that is more acceptable than the impact of the disease itself. The population screening approach must have high sensitivity and specificity. The screening methodology must also be cost-effective and practical to apply. This article explores how far we have traveled in forecasting autoimmune endocrinopathy in both common endocrine disease and in the context of rare but clustering disorders. We refer specifically to two papers that are being published in this issue of JCEM by Söderbergh et al. (1) and Kifor et al. (2). We review the strategies for prediction and perhaps targeted prevention by endocrine physicians in the future. Given that physicians who predict the future are often proven wrong, it is also useful to briefly review the scientific basis upon which to advise individuals or families about their future risk of the common, genetically complex, autoimmune diseases.

Prediction of Common Endocrine Disorders

Type 1 diabetes (T1D)

By virtue of its incidence and the major limitations of insulin therapy, it is not surprising that prediction and, more recently, prevention strategies have been pursued aggressively in T1D. Pancreatic autoimmunity is frequently initiated in the neonatal period (3), with autoantibodies apparent by early childhood. In family members (i.e. those with a high background risk), screening with serum islet cell antibodies (ICAs) has high sensitivity for predicting diabetes, but only those individuals with multiple ß-cell autoantibodies appear to progress (4). Several large prospective studies have measured antibodies to the pancreatic ß-cell antigens insulin, glutamic acid decarboxylase, and tyrosine phosphatase-related IA-2 protein in at risk individuals. The presence of serum antibodies to any one of these ß-cell antigens is associated with progression to overt diabetes in 2–15% of siblings of a proband with T1D (5, 6, 7). However, if antibodies to all three of these ß-cell antigens are present, the 10-yr risk of overt T1D increases markedly to 70–100%, with a sensitivity of 80–90%.

Recently, predictive strategies were validated in two landmark clinical intervention trials. The Diabetes Prevention Trial 1 (DPT-1) screened 84,228 first-degree relatives for ICA (8). Subjects identified as having a 5-yr risk of diabetes of 50% by virtue of ICA and insulin antibody positivity, and with decreased first phase insulin response were randomized. Although the prevention strategy, prophylactic parenteral insulin, was ineffective, the screening strategy was accurate with a rate of development of diabetes of 15% per year (8). The European Nicotinamide Diabetes Intervention Trial (ENDIT) used a similar screening strategy and again confirmed the feasibility and accuracy of large-scale prediction strategies (9).

Can prediction strategies be applied to the low-risk general population?

Less than one in 10 of subjects with T1D has an affected family member. It must, therefore, be established whether autoantibody screening strategies can be applied to a low-risk unselected population. The prevalence of a single pancreatic autoantibody in nonrelatives far exceeds those who progress to clinical diabetes (3), but as with family studies, combinations of autoantibodies give a high positive predictive value for the development of diabetes. In a large prospective study screening 4500 healthy schoolchildren, antibodies to two or three ß-cell antigens were present in 12 subjects (0.3%), six of whom developed T1D over an 8-yr observation period (10). Thus, the presence of antibodies to more than one ß-cell antigen still predicts T1D at least 50% of the time even in the unrelated population.

Does genetic screening aid prediction?

The use of human leukocyte antigen (HLA) genotying, in addition to assay of antibodies to ß-cell antigens alone, has been shown to lead to a small gain in predictive value over antibodies alone. Siblings of a T1D proband who have identical HLA or who carry the most diabetes-susceptible DQB*0201/DQB*0302 heterozygous genotype, along with multiple antibodies have the highest risk of progression to overt diabetes (70–100% over 10 yr) (11). Although genotyping appears to add little predictive value to autoantibody screening, it may provide a once-only method of identifying those within the general population who would benefit from second-round screening with autoantibodies. Those individuals with the dominantly protective allele DQB1*0602 or who carry neither the DQB*0201 or DQB*0302 haplotypes (about 50% of the Caucasian population) need no further screening. Such an approach has been applied for inclusion in the neonatal cow’s milk exclusion intervention trial (12).

Prediction of the other common endocrine autoimmune diseases

AITD has been subject to less systematic study because the perceived benefits of early identification of at-risk individuals are less clearly defined. In the population-based Whickham study (13), serum thyroid peroxidase (TPO; microsomal) antibody levels measured together with circulating TSH were shown to predict thyroid failure in healthy unrelated subjects. Women with positive microsomal antibodies (but with normal TSH) had a 2.1% per year risk of developing overt hypothyroidism. Overall, 27% of these euthyroid women who initially had serum thyroid antibodies were hypothyroid 20 yr later. Of women with both positive microsomal antibodies and an elevated serum TSH, 55% were hypothyroid at 20 yr, progressing at a rate 4.3% per year (13). This study provides firm evidence on which to advise the 10–15% of female subjects in the general population with positive TPO antibodies about their low future risk of hypothyroidism. The data are less robust for men, although it appears that both microsomal antibodies and elevated TSH are, if anything, more predictive of overt hypothyroidism in men compared with women.

Screening for postpartum thyroiditis (PPT) using TPO autoantibodies during pregnancy has high sensitivity, correctly identifying 90% of women who will develop the problem (14). Specificity is low, however, with only about 50% of antibody-positive individuals manifesting clinical thyroid dysfunction. PPT is a self-limiting condition in the majority of cases, and the benefits of early identification are, therefore, not obvious. It has been argued that TPO antibody screening for PPT should be pursued, because PPT is associated with postnatal depression (14). Furthermore, it identifies women at high risk of progressing to chronic autoimmune hypothyroidism (about 25% over 5 yr). However, TPO antibody positivity during pregnancy may have no more long-term significance than that found at other times of life (15).

Predicting a second endocrinopathy

The discussion so far has focused on endocrine diseases that are common in the general population in whom prediction programs would be applicable to large numbers of low-risk individuals. For subjects with a first endocrinopathy seeking guidance about their personal risk of a second disorder, it seems their risk is, to a substantial extent, determined by the nature and prevalence of their first disease. For instance, the rare subjects with Addison’s disease as a first endocrinopathy have about 30% risk of AITD, whereas subjects with a more common disorder, childhood T1D, have only a 15–20% risk of developing an additional autoimmune endocrine disorder over a lifetime. Similarly, subjects with AITD, the most common endocrine disease, have the lowest probability of developing a second disorder (~5%). Is there a role for the application of antibody screening to forecast further endocrinopathy?

Addison’s disease

Addison’s disease, although rare, still carries a substantial mortality when unrecognized. This provides a clear rationale for early detection in at-risk individuals. A large survey of nearly 9000 adult Italian patients with a first autoimmune endocrinopathy (predominantly AITD or T1D), assessed the risk of subsequently developing autoimmune adrenocortical failure (16). Overall, 0.8% of subjects were positive for adrenal cortex antibodies (ACAs). However, 9% of the subjects with autoimmune premature ovarian failure were ACA-positive. Of the 36 subjects who had positive ACA but normal adrenal function at baseline, one third progressed to symptomatic Addison’s disease or impaired adrenal function on serial ACTH testing (16). The mean time to overt Addison’s disease from a state of normal adrenal function was 5.2 yr, with a range of 23 to 71 months. Antibodies against 21 hydroxylase led to a small improvement in predictive value over ACA, but 17 hydroxylase and side-chain cleavage p450 antibodies did not provide additional information over and above the 21 hydroxylase antibody assay (16). HLA genotyping added little predictive information.

From the complex to a simple monogenic disorder

It is against this background that the study of the monogenic, autoimmune polyendocrinopathy syndrome type 1 (APS1) or autoimmune polyendocrinopathy, candidiasis, and ectodermal dystrophy syndrome is set. This is a rare disorder with onset in childhood and adolescence, but with a relatively predictable course, such that it provides a unique model in which to study the relationship between autoantibodies and hormonal failure (17, 18).

APS1 is due to recessive mutations in the autoimmune regulator (AIRE) gene, which is located on chromosome 21q22. The most frequent initial manifestation is chronic mucocutaneous candidiasis with an average onset occurring at 5 yr, which is typically followed by hypoparathyroidism (8 yr) and then Addison’s disease (12 yr) (17, 18). Of note, autoimmune hypoparathyroidism is very rarely seen outside the context of APS1. Other less common autoimmune features are diabetes, primary gonadal failure, pernicious anemia, intestinal malabsorption, hepatitis, hypothyroidism, alopecia, and vitiligo. In comparison to patients with the common complex polyendocrinopathy syndrome (type 2 or 3), in whom the average number of disease components is two, subjects with APS1 have an average of four components. Interestingly, APS1 appears to be associated with a unique spectrum of autoantibodies, some of which are rarely, if ever, found in other disorders.

In this issue of JCEM, Söderbergh et al. (1) present an extensive evaluation of 10 different autoantibodies in a large cohort of European APS1 subjects. They make three findings of note. First, their results suggest that there is redundancy in testing for antibodies to multiple steroidogenic enzymes and that 21 hydroxylase and side-chain cleavage enzyme are sufficient for the prediction of adrenocortical and gonadal failure, respectively. This extends a similar observation made in adults at risk of Addison’s disease (16).

Second, antibodies against tryptophan hydroxylase, initially identified as being associated with intestinal dysfunction in APS1, are shown in this study to be the strongest predictor of autoimmune hepatitis. This is an important finding, because in one large series up to 5% of APS1 subjects died of this complication (17). Screening and early detection may allow timely intervention.

Finally, of 73 APS1 subjects with hypoparathyroidism, none had antibodies against parathyroid calcium-sensing receptor (CaR). These findings are in contrast to the previous study of Li et al. (19) in which about 50% of sera from patients with autoimmune hypoparathyroidism selectively recognized the extracellular domain of the CaR on immunoblotting studies. Kifor et al. (2), also in this issue, clearly demonstrate that parathyroid antibodies directed against the CaR are present and functionally important in two subjects with hypoparathyroidism, and they have previously demonstrated a comparable biological role for anti-CaR antibodies in subjects with autoimmune hypercalcemia (20). These discrepant findings are likely to relate to differences in antibody assay technique, highlighting the need for standardization of sensitive antibody measurements. It is possible that naturally occurring autoantibodies to the CaR only recognize a conformational epitope on the mature glycosylated receptor, but additional work is needed to clarify this currently controversial area.

The study of organ-specific autoantibodies in APS1 has provided both a prediction model and a valuable insight into the pathogenesis of autoimmune endocrinopathy. There is currently a flurry of investigation into the mechanism by which the defective AIRE gene causes an aberrant immune response. AIRE is expressed in the thymic epithelial antigen-presenting cells, which have a role in educating T cells (negative selection), i.e. eliminating self-reactive T cells. This may explain why APS1 subjects have a unique spectrum of autoantibodies and a high penetrance of certain endocrinopathies. As we noted before, hypoparathyroidism is rare in situations other than APS1, suggesting that negative selection of T cells that recognize the CaR and other unknown parathyroid antigens is critical for immune tolerance. Novel targeted treatments for endocrine autoimmunity may stem from a better understanding of this well-characterized disease in the next few years.

What can we hope to achieve from the measurement of autoantibodies?

Over the last 40 yr, autoantibody assays have been developed to the point of precise disease prediction for certain common endocrinopathies such as T1D. This has now set the stage for prevention programs in T1D, as and when appropriate prevention strategies exist. In individuals with a single autoimmune endocrinopathy, autoantibody screening can aid the prediction of a life-threatening second manifestation of autoimmunity. The study of unique autoimmune syndromes such as APS-1 will provide many additional insights into the pathogenesis of these disorders and will advance the pursuit of targeted immunotherapy for autoimmune endocrinopathy.

Acknowledgments

We are grateful to Drs. Kate Owen and Tim Cheetham for critical reading of the manuscript.

Footnotes

The work in Dr. Pearce’s laboratory was funded by the Wellcome Trust and Medical Research Council (London, UK).

Abbreviations: ACA, Adrenal cortex antibody; AIRE, autoimmune regulator; AITD, autoimmune thyroid disease; APS1, autoimmune polyendocrinopathy syndrome type 1; CaR, calcium-sensing receptor; HLA, human leukocyte antigen; ICA, islet cell antibody; PPT, postpartum thyroiditis; T1D, type 1 diabetes; TPO, thyroid peroxidase.

Received December 11, 2003.

Accepted December 11, 2003.

References

  1. Söderbergh A, Myhre AG, Ekwall O, Gebre-Medhin G, Hedstrand H, Landgren E, Miettinen A, Eskelin P, Halonen M, Tuomi T, Gustafsson J, Husebye ES, Perheentupa J, Gylling M, Manns MP, Rorsman F, Kämpe O, Nilsson T 2004 Prevalence and clinical associations of 10 defined autoantibodies in autoimmune polyendocrine syndrome type I. J Clin Endocrinol Metab 89:557–562[Abstract/Free Full Text]
  2. Kifor O, McElduff A, LeBoff MS, Moore Jr FD, Butters R, Gao P, Cantor TL, Kifor I, Brown EM 2004 Activating antibodies to the calcium-sensing receptor in two patients with autoimmune hypoparathyroidism. J Clin Endocrinol Metab 89:548–556[Abstract/Free Full Text]
  3. Ziegler AG, Hummel M, Schenker M, Bonifacio E 1999 Autoantibody appearance and risk for development of childhood diabetes in offspring of parents with type 1 diabetes: the 2-year analysis of the German BABYDIAB Study. Diabetes 48:460–468[Abstract]
  4. Gardner SG, Gale EA, Williams AJ, Gillespie KM, Lawrence KE, Bottazzo GF, Bingley PJ 1999 Progression to diabetes in relatives with islet autoantibodies. Is it inevitable? Diabetes Care 22:2049–2053[Abstract]
  5. Verge CF, Gianani R, Kawasaki E, Yu L, Pietropaolo M, Jackson RA, Chase HP, Eisenbarth GS 1996 Prediction of type I diabetes in first-degree relatives using a combination of insulin, GAD, and ICA512bdc/IA-2 autoantibodies. Diabetes 45:926–933[Abstract]
  6. Kulmala P, Savola K, Petersen JS, Vahasalo P, Karjalainen J, Lopponen T, Dyrberg T, Akerblom HK, Knip M 1998 Prediction of insulin-dependent diabetes mellitus in siblings of children with diabetes. A population-based study. The Childhood Diabetes in Finland Study Group. J Clin Invest 101:327–336[Abstract/Free Full Text]
  7. Bingley PJ, Christie MR, Bonifacio E, Bonfanti R, Shattock M, Fonte MT, Bottazzo GF, Gale EA 1994 Combined analysis of autoantibodies improves prediction of IDDM in islet cell antibody-positive relatives. Diabetes 43:1304–1310[Abstract]
  8. Diabetes Prevention Trial–Type 1 Diabetes Study Group 2002 Effects of insulin in relatives of patients with type 1 diabetes mellitus. N Engl J Med 346:1685–1691[Abstract/Free Full Text]
  9. Gale EA; European Nicotinamide Diabetes Intervention Trial Group 2003 Intervening before the onset of type 1 diabetes: baseline data from the European Nicotinamide Diabetes Intervention Trial (ENDIT). Diabetologia 46:339–346[CrossRef][Medline]
  10. LaGasse JM, Brantley MS, Leech NJ, Rowe RE, Monks S, Palmer JP, Nepom GT, McCulloch DK, Hagopian WA; Washington State Diabetes Prediction Study 2002 Successful prospective prediction of type 1 diabetes in schoolchildren through multiple defined autoantibodies: an 8-year follow-up of the Washington State Diabetes Prediction Study. Diabetes Care 25:505–511[Abstract/Free Full Text]
  11. Kulmala P, Savola K, Reijonen H, Veijola R, Vahasalo P, Karjalainen J, Tuomilehto-Wolf E, Ilonen J, Tuomilehto J, Akerblom HK, Knip M 2000 Genetic markers, humoral autoimmunity, and prediction of type 1 diabetes in siblings of affected children. Childhood Diabetes in Finland Study Group. Diabetes 49:48–58[Abstract]
  12. Sadeharju K, Hamalainen AM, Knip M, Lonnrot M, Koskela P, Virtanen SM, Ilonen J, Akerblom HK, Hyoty H; Finnish TRIGR Study Group 2003 Enterovirus infections as a risk factor for type I diabetes: virus analyses in a dietary intervention trial. Clin Exp Immunol 132:271–277[Medline]
  13. Vanderpump MP, Tunbridge WM, French JM, Appleton D, Bates D, Clark F, Grimley Evans J, Hasan DM, Rodgers H, Tunbridge F, Young ET 1995 The incidence of thyroid disorders in the community: a twenty-year follow-up of the Whickham Survey. Clin Endocrinol (Oxf) 43:55–68[Medline]
  14. Harris B, Othman S, Davies JA, Weppner GJ, Richards CJ, Newcombe RG, Lazarus JH, Parkes AB, Hall R, Phillips DI 1992 Association between postpartum thyroid dysfunction and thyroid antibodies and depression. BMJ 305:152–156[Medline]
  15. Amino N, Tada H, Hidaka Y, Crapo LM, Stagnaro-Green A 1999 Therapeutic controversy: screening for postpartum thyroiditis. J Clin Endocrinol Metab 84:1813–1821[Free Full Text]
  16. Betterle C, Volpato M, Rees Smith B, Furmaniak J, Chen S, Greggio NA, Sanzari M, Tedesco F, Pedini B, Boscaro M, Presotto F 1997 I. Adrenal cortex and steroid 21-hydroxylase autoantibodies in adult patients with organ-specific autoimmune diseases: markers of low progression to clinical Addison’s disease. J Clin Endocrinol Metab 82:932–938[Abstract/Free Full Text]
  17. Perheentupa J 2002 APS1/APECED: the clinical disease and therapy. Endocrinol Metab Clin North Am 31:295–320[Medline]
  18. Pearce SH, Cheetham TD 2001 Autoimmune polyendocrinopathy syndrome type 1: treat with kid gloves. Clin Endocrinol (Oxf) 54:433–435[CrossRef][Medline]
  19. Li Y, Song YH, Rais N, Connor E, Schatz D, Muir A, Maclaren N 1996 Autoantibodies to the extracellular domain of the calcium sensing receptor in patients with acquired hypoparathyroidism. J Clin Invest 97:910–914[Abstract/Free Full Text]
  20. Kifor O, Moore Jr FD, Delaney M, Garber J, Hendy GN, Butters R, Gao P, Cantor TL, Kifor I, Brown EM, Wysolmerski J 2003 A syndrome of hypocalciuric hypercalcemia caused by autoantibodies directed at the calcium-sensing receptor. J Clin Endocrinol Metab 88:60–72[Abstract/Free Full Text]