Advances in Understanding Autoimmune Pituitary Disease: Standardized Methods for Autoantibody Detection

An increasing focus on autoimmune disease has centered on the most commonly affected organ systems. Most of these diseases are associated with a disrupted architecture and function. They result from the infiltration of immunocompetent cells such as T lymphocytes, activation of residential cells, and tissue remodeling (1). In practice, assessing cellular reactivity requires the invasive interrogation of frequently inaccessible tissues. Many diseases in the spectrum of autoimmunity are also associated with production of circulating autoantibodies. In no disease has their importance in clinical decision making been more apparent than in type I diabetes mellitus (2). Despite their frequent lack of disease specificity, several autoantibodies have become invaluable diagnostic markers, and their levels are now routinely assessed by commercial laboratories. They have also been used to judge whether an individual might be at increased risk for developing disease in the future. Therefore, their presence can allow stratification in those who might benefit from increased clinical surveillance. Monitoring many autoantibodies has become a mainstay of managing autoimmune diseases such as those affecting the thyroid, pancreas, and adrenal glands. Furthermore, they have the potential for use as biomarkers for assessing the responses to therapy.

Within the spectrum of human endocrine autoimmunity, the processes targeting the pituitary have received considerably less attention, in large part because of their perceived rarity (3). Lymphocytic hypophysitis, also known as autoimmune hypophysitis, is a poorly understood process where lymphoid cells infiltrate the pituitary (4). The condition is associated with disruption of tissue organization and frequently results in glandular malfunction.

Scattered reports describing pituitary autoimmunity began to appear in the early 1960s when Goudie and Pinkerton (5) first observed lymphocytic infiltration of the adrenal, thyroid, and pituitary in a woman who had died following childbirth. Engelberth and Ježková (6) described antipituitary antibodies shortly thereafter, but others have failed to detect these antibodies in idiopathic panhypopituitarism (7). An association between thyroid autoimmunity and detection of pituitary antibodies was discovered several years ago (8, 9). This relationship is implied by finding that immunoglobulins directed at as yet unidentified pituitary antigens can be detected in individuals with Hashimoto's thyroiditis and Graves' disease. Some disagreement exists in the frequency of the concomitance (8–11). This constellation of findings has been classified as autoimmune polyglandular syndrome 3A.

The sine qua non for definitively establishing the diagnosis of pituitary autoimmunity must include a biopsy of the gland. Because of its inaccessibility, the vast majority of cases in which hypophysitis is suspected must be assessed through indirect means. These include endocrine testing, imaging (12), and assaying sera from affected patients for the presence of antipituitary antibodies. A particular barrier to reliably detecting and identifying these antibodies has been the uncertain identity of their cognate autoantigens (3). Thus, the techniques for quantifying these antibodies and characterizing their interactions with target epitopes have not been developed into clinically useful and accessible assays. These deficits undoubtedly limit the clinical evaluation and care of patients with pituitary disease.

A major hurdle to better defining the pathogenesis of pituitary autoimmunity results from the imprecision with which pituitary antibodies are detected and classified. Little is known about any pathogenic role that they might play in the development of pituitary dysfunction. Most studies examining these antipituitary antibodies have relied on indirect immunofluorescence (IIF) where pituitary tissue is stained with sera or partially purified immunoglobulins from patients and is then judged to be either positive or negative for reactivity. Inconsistent results can be found in existing literature. In one of the largest and most comprehensive reports, Bottazzo et al (11) interrogated sera from 287 patients with one or more autoimmune endocrine diseases and found that 19 of these reacted with prolactin-producing cells obtained and analyzed by IIF from a hypophysectomized woman. A more recent study by Manetti et al (13) analyzed sera from 961 patients with autoimmune thyroid disease and 135 healthy controls for antipituitary antibodies using IIF with baboon anterior and posterior pituitary tissues. They found that the antibodies were more prevalent in sera collected from individuals with thyroid disease (13). These findings are congruent with those of other groups who also find that pituitary antibodies are more frequently detected in cohorts with autoimmune thyroid disease.

Despite its perceived rarity, autoimmune hypophysitis has emerged as a more frequently encountered process in patients undergoing immune modulatory therapies for the treatment of certain cancers (14). Agents that alter immune function have been found to provoke autoimmune reactions against several tissues, including the pituitary. In some cases, this has resulted in clinical disease such as pan-hypopituitarism. Anti-cytotoxic T lymphocyte antigen-4 (CTLA-4) antibodies, such as ipilimumab and tremelimumab, have been evaluated as therapy for advanced melanoma (15), and the former has received Food and Drug Administration approval as a first-line therapy. These agents can lead to autoimmune hepatitis, uveitis, enterocolitis, dermatitis, as well as hypophysitis (16, 17). Similar monoclonal antibodies are being assessed for their effectiveness against additional malignancies and thus may come into even wider use. A very recent report from Iwama et al (18) has shed light on the pathogenic mechanisms underlying this form of secondary (iatrogenic) autoimmune hypophysitis. Ipilimumab apparently decorates CTLA-4 for binding to complement C3d and C4d and sets the stage for an inflammatory response. These initial experiences with anti-CTLA-4 antibodies suggest that as immunomodulation becomes even more common in cancer therapy, cases of pituitary disease are likely to become more frequent. Improved means of detecting antipituitary antibodies should aid in the care of these affected patients.

Despite measured progress in identifying individuals with pituitary disease, substantial deficits remain in detecting pituitary autoimmunity. Detection of antipituitary antibodies has yet to become widely established as a clinically useful diagnostic tool. This has resulted in large part from a lack of standardization and an absence of commercially available assays. The variable methods employed in conducting IIF and the lack of consensus regarding the correct interpretation of assay results have confounded attempts at reconciling the literature and impeded the development of laboratory-based tools for detecting pituitary autoimmunity.

This issue of the JCEM contains a report from Ricciuti et al (19) who set out to comprehensively review earlier studies utilizing IIF for detecting antipituitary antibodies. They identify specific shortcomings in the antibody assays and suggest technical improvements that should allow more reproducible means of diagnosing pituitary disease. The paper contains a thorough analysis of the relevant literature by identifying 122 articles in which IIF was utilized to detect patient-derived antibodies against pituitary tissue. What emerged from their review was a great deal of variability in the results reported. This was attributed to the use of pituitaries from several different species in the IIF-based assays, different methods of tissue fixation, and inconsistent definitions of disease-associated staining patterns. This review is followed by descriptions of the authors' efforts to develop optimized methods for detecting pituitary antibodies by IIF. The array of human, cynomolgus monkey, dog, and mouse pituitary tissues was evaluated for suitability as the antigenic target. The authors adhered strictly to a protocol of tissue fixation, antigen blocking, staining, and independent interpretation conducted by two investigators. Their studies revealed that the source of pituitary tissue was an important determinant of IIF quality. Human pituitary was one of the lowest emitters of autofluorescence, and Sudan black B was an excellent attenuator of these emissions. Having defined optimal conditions for sample preparation, two of the masked investigators independently interpreted tissue staining, which proved highly concordant (3% discrepant). The assay protocol was then used to determine the prevalence and biological importance of antipituitary antibodies in human diseases. The original cross-sectional case-control study included in the report of Ricciuti et al (19) analyzed sera from 413 cases and 60 healthy controls. Of these, 218 emanated from patients with benign pituitary tumors, 24 donors with developmental abnormalities of the pituitary, and 31 biopsy-proven and 27 suspected cases of hypophysitis. In addition, 23 donors carried the diagnosis of Hashimoto's thyroiditis. This cohort of patients represents the largest ever evaluated for the prevalence of antipituitary antibodies. From the results emerging from their use of the standardized protocol, the authors concluded that a binary interpretation of pituitary staining by IIF (ie, either positive or negative) allowed them to determine that antibodies were more common in sera from individuals with pituitary disease (95 of 390 cases, or 24%) compared to those from healthy donors (3 of 60 cases, or 5%). But it did not allow them to distinguish between pituitary diseases. Furthermore, antipituitary antibodies were detected as frequently in patients with nonautoimmune pituitary processes as in individuals with autoimmune hypophysitis. However, when IIF was scored for staining intensity, area of stain distribution, and staining pattern, only the cytoplasmic staining patterns (perinuclear, diffuse, and granular) varied among different pituitary diseases. Notably, a granular cytosolic pattern was highly predictive of pituitary autoimmunity. A diffuse pattern was observed both in cases of autoimmunity and in pituitary tumors. The perinuclear pattern was found both in healthy controls and in cases from all disease categories, rendering it useless as a diagnostic marker. Their study further disclosed improved results when purified immunoglobulins were used instead of whole patient serum. They also found that blocking Fc receptors was useful in cases yielding ambiguous staining results, such as those associated with perinuclear staining. In contrast, the granular staining pattern, indicative of pituitary autoimmunity, was unaffected by Fc blockade, attesting to its disease specificity. FSH-secreting cells were the most frequently identified cell type that stained with specificity, followed by thyrotrophs, and cells secreting LH and ACTH. In contrast, cells producing prolactin and GH were recognized by antipituitary antibodies less frequently.

Increasing awareness of pituitary autoimmunity has uncovered an unmet need for improved detection of antipituitary antibodies and for a more complete understanding of their role in disease pathogenesis. Introduction of commercially available assays for clinical use would allow further assessment of their potential diagnostic utility. Ricciuti et al (19) provide many of the essential details for optimizing antibody detection. Recognizing the advantages of human pituitary tissue use as the staining substrate in IIF appears to be among their most important findings. Human pituitary has rarely been used in previous studies, probably because of the relative difficulty in acquiring suitably fresh and well-preserved tissue. In that regard, the current study reinforces earlier findings of Glück and Scherbaum (20) who also concluded that human pituitary was superior. Beyond the technical details it provides, the paper of Ricciuti et al (19) calls attention to a group of autoimmune processes that have eluded our understanding and thus may fail to be recognized during routine clinical care. The paper also provides insight into potentially important directions for further inquiry and offers the practical tools for these future studies. Among the most pressing issues to be addressed is the identification of the pituitary antigen(s) that are targeted by these antibodies (21). Determination of whether any of these antibodies are pathogenic should advance our understanding of pituitary autoimmunity. Alternatively, these antibodies might serve as useful biomarkers of disease activity and severity as well as outcome measures, should prevention and medical treatment of pituitary disease become possible. Improved means of detecting antipituitary antibodies should make the identification of individuals at risk for developing these diseases easier and allow earlier diagnosis, perhaps in advance of gland dysfunction.