Crooke's Changes In Cushing's Syndrome Depends on Degree of Hypercortisolism and Individual Susceptibility

The Journal of Clinical Endocrinology & Metabolism, Aug 2015

Although Crooke's changes in the pituitary corticotrophs were initially described in 1935, the prevalence in which the changes occur in patients with Cushing's syndrome (CS) has not been established.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://academic.oup.com/jcem/article-pdf/100/8/3165/17636486/jcem3165.pdf

Crooke's Changes In Cushing's Syndrome Depends on Degree of Hypercortisolism and Individual Susceptibility

Endocrine Society Received June Crooke's Changes In Cushing's Syndrome Depends on Degree of Hypercortisolism and Individual Susceptibility Edward H. Oldfield 0 Mary Lee Vance 0 Robert G. Louis 0 Carrie L. Pledger 0 John A. Jane Jr. 0 Maria-Beatriz S. Lopes 0 0 Department of Neurological Surgery (E.H. O., R.G.L., C.L.P., J.A.J) , Department of Medicine (M.L.V.), and Department of Pathology (M.-B.S.L.), University of Virginia Health System , Charlottesville, VA 22903 , USA Context: Although Crooke's changes in the pituitary corticotrophs were initially described in 1935, the prevalence in which the changes occur in patients with Cushing's syndrome (CS) has not been established. Objective: This study aimed to determine the prevalence and assess clinical features associated with the presence or absence of Crooke's changes in a large set of patients with CS. Design: Information from a prospective computer database and retrospective chart review was analyzed. Setting: The setting was an academic medical center. Patients: Consecutive patients (N Cushing's disease are included. - Abbreviations: CS, Cushing’s syndrome; EACTHS, ectopic ACTH syndrome; H&E, hematoxylin and eosin; MRI, magnetic resonance imaging; UFC, urinary free cortisol; ULN, upper limit of normal. doi: 10.1210/JC.2015-2493 press.endocrine.org/journal/jcem I the Institute of Pathology in London, first described n 1935 Dr Arthur Crooke, Grocers’ Research Scholar at replacement of the cytoplasmic granules of the basophilic cells of the anterior pituitary with homogeneous hyaline material in patients with Cushing’s syndrome (CS), now known as Crooke’s changes ( 1 ). Although it has been nearly 80 years since that original description, there is limited and conflicting information on the frequency in which Crooke’s changes occur in Cushing’s disease. Furthermore, other than that it is associated with excess glucocorticoid exposure, whether endogenous or iatrogenic, there is no information on the basis of its presence or absence in patients with hypercortisolism. The goals of this study were to document the prevalence of Crooke’s changes in the pituitary gland in a large consecutive series of patients who had CS and who underwent pituitary surgery and to examine clinical and laboratory features that may underlie the presence or absence of Crooke’s changes in the pituitary corticotrophs. Materials and Methods Some of the data used for this study were kept in a prospective computerized database. Others were obtained by retrospective review of patient files and from a computerized hospital database after approval by our local institutional internal review board. The study includes a consecutive series of 224 patients who underwent pituitary surgery with a preoperative diagnosis of Cushing’s disease between January 2008 and March 2014. The diagnosis was based on the results of conventional endocrine testing for the presence of CS and for its differential diagnosis. Patients with a negative pituitary magnetic resonance imaging (MRI) and patients with conflicting results of endocrine testing for the differential diagnosis of CS underwent bilateral inferior petrosal sinus sampling. Pituitary imaging with high resolution MRI with and without contrast was obtained in all patients. Not all data was available on all patients. In many patients the preoperative 24-hour urinary free cortisol (UFC) was performed at the referring center’s laboratory. Thus, values for UFC are presented as the ratio of the measured value divided by the upper limit of normal of the assay used (ULN). Many patients were from distant referrals; the normal range of the assay for UFC was not available in 18 patients, 13 of whom had ACTH-staining tumors. Specimen handling and histological analysis All pituitary resected tissues were submitted fresh within minutes of completing the surgical resection. Specimens were then fixed in 10% formalin and embedded in paraffin for light microscopy. All specimens were initially evaluated using a 5- mthick hematoxylin and eosin (H&E)-stained section. In cases in which an adenoma was readily identified, a panel consisting of Wilder’s reticulin and immunohistochemical stains for ACTH ACTH, prolactin, GH, and FSH were subsequently performed. When a pituitary adenoma was not identified on the initial H&E section, a “Cushing’s panel” was ordered, consisting of 10 sequential 5- m-thick sections through the paraffin block with every third slide stained with H&E. In sections suspicious for an adenoma, the intervening unstained sections were stained for reticulin (Wilder’s reticulin), an ACTH and either a prolactin or a GH immunostain. If an adenomatous area was not identified, further “Cushing’s panels” were ordered until the tissue block was exhausted. Selective cases were also immunostained with a cytokeratin monoclonal antibody (CAM 5.2). Crooke’s hyaline change in corticotroph cells was identified as a concentric accumulation of glassy, hyaline material in the cytoplasm displacing the secretory granules to the periphery of the cells by either the H&E, negative immunostaining for ACTH or CAM 5.2 immunoreactivity (Figure 1). The partial lung lobectomy specimens were fixed in 10% formalin and routinely embedded in paraffin for light microscopy. After review of the H&E-stained slides and identification of the neuroendocrine tumor, immunohistochemical stains for ACTH were performed for confirmation of the diagnosis. Statistical analysis was completed with GraphPad Prism (GraphPad Software). For comparisons of medians (UFC, tumor size, etc.) between groups, the Mann-Whitney U test was used. For comparisons of the number of subjects or conditions within groups, Fisher’s exact test was used. The specific tests used are described in the Table and Figure legends. Results Histopathology The pituitary surgical specimens were evaluated for the presence of both a pituitary adenoma and normal anterior pituitary gland. If normal anterior pituitary gland was present, the presence or absence of Crooke’s changes in corticotroph cells was noted. No quantitative measurement of either the number of cells containing Crooke’s changes or the extension of changes in the normal gland was performed due to the uneven size of the para-adenomatous anterior gland. In 11 of the 224 patients there was insufficient normal gland in the specimen to evaluate for Crooke’s changes. Thus, the results of 213 patients are analyzed. The presence of an ACTH-immunoreactive pituitary adenoma was pathologically confirmed in 174 patients (Table 1). In 39 patients there was no ACTH pituitary adenoma in the tissue submitted from pituitary surgery despite extensive histopathological analysis (Methods). Patients Among the 39 patients in whom there was no ACTHstaining specimen in the tissue submitted from pituitary surgery four had the ectopic ACTH syndrome (EACTHS). Three had pituitary surgery despite results from inferior petrosal sinus sampling that suggested the EACTHS, one of whom had small nonenhancing lesions in the pituitary on high-resolution pituitary MRI. Three of these patients were later diagnosed with ectopic ACTH-producing neuroendocrine tumors; in two of them it was a well-differentiated bronchial neuroendocrine carcinoid (typical carcinoid) removed surgically; one patient had remission of CS associated with profound hypocortisolism after radiofrequency ablation of a small lung lesion; the diagnosis has not been established in the fourth patient. The remaining 35 patients had preoperative endocrine assessment that included testing for the differential diagnosis of CS that was consistent with Cushing’s disease, but there was no ACTH-pituitary adenoma tumor in the tissue from surgery. No patient had had prior adrenalectomy or recent medical therapy to block cortisol production or action. Patients with ACTH-positive neuroendocrine tumors The initial analysis is for the 177 patients in whom an ACTH-positive neuroendocrine tumor (174 patients with ACTH pituitary adenoma and three patients with ACTH pulmonary neuroendocrine carcinoma) was in the tissue removed at surgery. The patient with EACTHS successfully treated with radiofrequency ablation of a pulmonary lesion is included. One hundred forty-four patients had Crooke’s changes (81%) and 33 did not (19%). All patients with a diagnosis of EACTHS had Crooke’s changes in the pituitary. The only feature demonstrating a substantial and significant difference between patients with or without Crooke’s changes was the maximum 24-hour UFC before surgery (Figure 2, Table 1). In 16 of the 177 patients the ULN of the assay used for UFC was unavailable. Among the remaining 161 patients only six of the 31 patients without Crooke’s changes (19%) had a UFC of at least 4-fold the ULN, compared with 59 of the 130 patients with Crooke’s changes (45%; P .008, Fisher’s exact test). In patients with Crooke’s changes the preoperative UFC (mean, 6.3-fold ULN) was substantially greater than in patients without Crooke’s changes (2.9-fold ULN) (Table 1). There was no difference between patients with or without Crooke’s changes in the random ACTH or cortisol Comparison of Patients With and Without Crooke’s Changes Patients with ACTH-staining pituitary adenomas Maximum preoperative UFCa With Crooke’s changes Without Crooke’s changes Minimum plasma cortisol 2–3 d after surgery, ug/dL With Crooke’s changes Without Crooke’s changes Maximum diameter adenoma at surgery, mm With Crooke’s changes Without Crooke’s changes All patients with tissue evaluable for Crooke’s changes Maximum preoperative UFC, ug/d With Crooke’s changes Without Crooke’s changes UFC ACTH tumor present or nota ACTH tumor No ACTH tumor Minimum plasma cortisol 2–3 d after surgery, ug/dL With Crooke’s changes Without Crooke’s changes Mean values obtained before surgery, the diameter of the tumor as measured at surgery, the nadir postoperative cortisol, or ACTH nadir over the first 48 –72 hours after surgery, or the change in weight or change in body mass index at 1 year after surgery (data not shown). All patients, with or without histopathological documentation of an ACTH-positive neuroendocrine tumor We also analyzed the same features in the entire set of 213 patients, patients who received pituitary surgery and in whom Crooke’s changes could be assessed, including patients with or without histological evidence of an ACTH-positive tumor (Table 1). Among the 213 patients 158 had Crooke’s changes (74%) and 55 did not (26%). Among the 55 patients with no Crooke’s changes, in 21 (38%) no ACTH-positive tumor was found and removed at surgery, whereas no ACTH-positive tumor was found in only 15 (10%) of the 158 patients with Crooke’s changes (P .0001, Fisher’s exact test; Figure 2, left). Among the patients with no Crooke’s changes there was no difference in the preoperative 24-hour UFC production between patients who had or did not have an ACTHpositive adenoma found and removed at surgery. The incidence of Crooke’s changes was higher in patients in whom an ACTH-staining adenoma was in the surgical specimen. In the 36 patients with CS and no ACTH-positive adenoma in the specimen, 15 had Crooke’s changes (42%) compared with 143 of the 177 patients with CS and an ACTH-positive tumor (81%) (P .0001, Fisher’s exact test; Figure 2, right). Again, there were substantial and significant differences in preoperative cortisol production between patients with and without Crooke’s changes (Figure 3). Note in Figure 3 that the likelihood of Crooke’s changes was higher in patients with maximum 24-hour UFC of 4-fold ULN or greater; only 13% of patients had no Crooke’s changes, compared with 35% of patients whose maximum UFC was less than 4-fold the ULN (P .0008, Fisher’s exact test). In 18 patients the ULN of the UFC assay was not available. Discussion Perinuclear bundles of cytokeratin filaments are a characteristic feature of corticotroph cells. Under conditions of glucocorticoid excess, either exogenous or endogenous, corticotroph cells accumulate cytokeratin filaments in the cytoplasm in an annular or concentric fashion with displacement of the secretory granules to either the periphery of the cell or around the nucleus ( 2 ). This transformed corticotroph cell is recognized as Crooke’s cell. The accumulation of cytokeratin filaments is considered as the morphologic manifestation of functional suppression ( 3 ). Crooke’s hyaline change is a phenomenon observed only in the human pituitary gland under influence of high cortisol levels with no successful reproduction in experimental models. Crooke, in his original paper, studied the pituitary of 350 other patients in addition to the 12 patients with Cushing’s disease. He himself noted that these changes do not occur in the absence of exposure to glucocorticoid excess; “A hyaline change in the basophil cells of the pituitary gland is described which does not seem to be an expression of cell degeneration in the ordinary sense but is probably an expression of altered physiological activity. It was found in slight amount in a few basophil cells of the anterior lobe in nine only of some 350 pituitary glands from various conditions in which the syndrome attributed to basophil adenoma was absent.” ( 1 ) The prevalence of Crooke’s changes associated with CS in prior reports is inconsistent and is not fully established, ( 4 ) to a certain extent due to the uneven presence of normal, para-adenomatous pituitary gland in surgical specimens. In a series of 111 pituitary surgical specimens of ACTH-positive tumors from the German Pituitary Registry, 100% of cases of active ACTH cell adenomas with satisfactory para-adenomatous pituitary tissue sampling were reported to show Crooke’s changes ( 4 ). However, this is in conflict with a different conclusion in a report coauthored by two of the same authors in which just 20 of 55 patients with surgery for Cushing’s disease had Crooke’s changes (36%) ( 5 ). The results of both of those studies contrast with the current study, in which Crooke’s changes occurred in 81% of patients with histologically demonstrated ACTH-positive neuroendocrine tumors (pituitary and extrapituitary tumors) and in 74% of all patients who had pituitary surgery. The only prior report of a link of Crooke’s changes with abnormal clinical physiology are the reports of Saeger et al in 1988 ( 6 ) and Flitsch, et al in 1999 ( 5 ), who describe a correlation between the presence of Crooke’s changes and the interval to recovery of the hypothalamic-pituitary-adrenal axis after treatment of Cushing’s disease. Flitsch et al ( 5 ) found Crooke’s changes in 10 of 15 patients (67%) who required cortisol replacement therapy for at least 36 months after remission on Cushing’s disease, but in only five of 29 patients who required cortisol replacement therapy for less than 36 months (17%), thus statistically correlating Crooke’s changes with longer postoperative corticotroph suppression and hypocortisolism ( 5 ). They suggest that the degree of Crooke’s changes may be an indicator for the degree of the ACTH suppression level of the normal corticotrophs. Our results clearly demonstrate a correlation between the degree of cortisol production and the presence of Crooke’s changes. Patients with cortisol production exceeding 4-fold the ULN almost all had Crooke’s changes. In contrast, some patients with cortisol production less than 4-fold the ULN had Crooke’s changes, others did not (Figure 3). This could have been the result of differences in the duration of CS, variation in the constancy of excess cortisol exposure, or individual variation in the sensitivity to the effects of excess cortisol from patient to patient. The presence of Crooke’s changes also probably is a reflection of the duration of exposure to excess glucocorticoids. In studies of pituitary tissues from autopsy of patients who received sustained treatment with glucocorticoids, Uei et al ( 7 ) found a significant correlation with the presence of Crooke’s changes and 1) the total dose of glucocorticoids and 2) the duration of treatment. However, others did not find such a correlation ( 8 ). We attempted to examine the duration of CS as reflected in the patient history, but the timing of the onset of the syndrome is often not clear, many patients cannot recall when their initial symptoms first seemed, and that information was not available for many patients. In addition, the degree of hypercortisolism was only examined at one point in time, during the preoperative evaluation for detection of CS, and is, thus, not a reliable measure of the degree of cortisol exposure over an interval of time. In some patients with a diagnosis of Cushing’s disease no tumor can be found at surgery. This usually occurs in patients with a negative pituitary MRI. In fact, our series represents one in which there is a referral bias toward patients with suspected Cushing’s disease who have a negative MRI, as 48 of the 213 patients (23%) had a negative preoperative pituitary MRI. The presence of Crooke’s changes can be used to suggest that a patient who had surgical exploration of their pituitary, but in whom no tumor can be found, has CS, as the presence of Crooke’s is a reliable indication of excess exposure to glucocorticoids, including from ectopic ACTH syndrome, as occurred in all three of the patients with proven ectopic ACTH secretion in this series. However, the absence of Crooke’s changes does not reliably suggest the absence of CS, as 19% of patients with a proven ACTH-producing tumor did not have Crooke’s changes. The likelihood of not finding a tumor in patients with no Crooke’s changes (38%) was substantially higher than in patients with Crooke’s changes (10%). Furthermore, in patients in whom no ACTH pituitary adenoma was in the tissue specimen removed at surgery the incidence of Crooke’s changes was 58%. These observations suggest that some patients in whom a tumor was not found may not have had CS. If this was so, the distribution of maximum cortisol production (UFC) among patients with no Crooke’s changes should be different between patients with and without an ACTH-pituitary adenoma from the surgery, but that was not the case. Thus, the higher likelihood of absent Crooke’s changes in patients without an ACTH-pituitary adenoma from surgery may be explained by those patients having smaller tumors, less severe hypercortisolism, or shorter exposure to high cortisol levels. The analysis also suggests that there is variability among individuals in their susceptibility to develop Crooke’s changes in the face of chronic hypercortisolism, as has previously been observed in patients receiving high dose glucocorticoid therapy ( 7 ). To summarize, Crooke’s changes occur in 75– 80% of patients with CS. The presence or absence of Crooke’s changes depend on the degree of hypercortisolism and individual variability. They are present in almost all patients with UFC greater than 4-fold the ULN, whereas with less severe hypercortisolism the expression of Crooke’s changes varies from person to person. The presence of Crooke’s changes is a clear indication of the presence of CS, although the absence of Crooke’s changes does not exclude it. Acknowledgments Address all correspondence and requests for reprints to: Edward H. Oldfield, MD, FACS, Department of Neurological Surgery, Uni 1. Crooke A . Change in the basophil cells of the pituitary gland common to conditions which exhibit the syndrome attributed to basophil adenoma . J Pathol Bacteriol . 1935 ; 41 : 339 - 349 . 2. Neumann PE , Horoupian DS , Goldman JE , Hess MA . Cytoplasmic filaments of Crooke's hyaline change belong to the cytokeratin class. An immunocytochemical and ultrastructural study . Am J Pathol . 1984 ; 116 : 214 - 222 . 3. Kovacs K. The pathology of Cushing's disease . J Steroid Biochem Mol Biol . 1993 ; 45 : 179 - 182 . 4. Saeger W , Lüdecke DK , Buchfelder M , Fahlbusch R , Quabbe HJ , Petersenn S. Pathohistological classification of pituitary tumors: 10 years of experience with the German Pituitary Tumor Registry . Eur J Endocrinol . 2007 ; 156 : 203 - 216 . 5. Flitsch J , Lüdecke DK , Knappe UJ , Saeger W. Correlates of longterm hypocortisolism after transsphenoidal microsurgery for Cushing's disease . Exp Clin Endocrinol Diabetes . 1999 ; 107 : 183 - 189 . 6. Saeger W , Geisler F , Lüdecke DK . Pituitary pathology in Cushing's disease . Pathol Res Pract . 1988 ; 183 : 592 - 595 . 7. Uei Y , Kanzaki M , Yabana T. Further immunohistochemical study of Crooke's hyalin . Pathol Res Pract . 1991 ; 187 : 539 - 540 . 8. DeCicco FA , Dekker A , Yunis EJ . Fine structure of Crooke's hyaline change in the human pituitary gland . Arch Pathol . 1972 ; 94 : 65 - 70 .


This is a preview of a remote PDF: https://academic.oup.com/jcem/article-pdf/100/8/3165/17636486/jcem3165.pdf

Oldfield, Edward H., Vance, Mary Lee, Louis, Robert G., Pledger, Carrie L., Jane, John A., Lopes, Maria-Beatriz S.. Crooke's Changes In Cushing's Syndrome Depends on Degree of Hypercortisolism and Individual Susceptibility, The Journal of Clinical Endocrinology & Metabolism, 2015, 3165-3171, DOI: 10.1210/JC.2015-2493