Humoral Manifestations of Malignancy
Dolores Shoback MD
Janet Funk MD
ECTOPIC HORMONE & RECEPTOR SYNDROMES
Some of the most challenging endocrine problems occur in patients with malignancies of diverse cell-types. This is because both endocrine and nonendocrine tumors secrete polypeptide hormones. As it became recognized that a polypeptide hormone could be produced by tumor cells derived from a tissue that normally did not secrete the hormone, the notion of ectopic hormone production developed. Most tumors associated with ectopic hormone syndromes are derived from cells that are normally capable of producing peptide hormones. Initially, it was thought that ectopic hormone production by tumor cells was a rare event. Interestingly, both the frequency and the original conception of this syndrome have been redefined over the last few decades. It has come to be appreciated—through the use of modern biochemical and molecular biologic techniques—that the synthesis of peptide hormones and the transcription of their genes by tumor cells are in fact quite common occurrences. Tumor cells may differ from normal cells in their ability or inability to process precursor molecules, which may account for the presence or absence of hormone excess states and for the profile of peptide hormone forms and fragments present in the circulation and in tumor cell extracts. However, tumor production of hormone fragments or precursors is much more common than the clinical syndromes of hormone excess.
The classic criteria used to confirm that a tumor is the source of a hormone excess state include the following: (1) evidence of an endocrinopathy in a patient with a tumor; (2) remission of the endocrinopathy after tumor resection; (3) detection of an arteriovenous gradient across the tumor; and (4) documentation of hormone protein and messenger RNA production in the tumor tissue.
In addition to classic hormone excess states resulting from the ectopic or inappropriate secretion of a hormone by an endocrine or nonendocrine tumor, endocrinopathies can result from the ectopic expression of a hormone's receptor. This is well illustrated, for example, by the occurrence of Cushing's syndrome in pregnancy or in relation to meals, due to the ectopic expression of luteinizing hormone or gastric inhibitory polypeptide receptors in adrenal tissue, respectively. Several other examples of ectopic receptor syndromes have been documented. Some of these will be discussed below, particularly as a cause for unusual forms of ACTH-independent Cushing's syndrome.
A variety of peptides are produced by both benign and malignant tumors as listed in Table 21-1. The biochemical pathways and machinery leading to the synthesis, processing, and secretion into the circulation of a peptide hormone are present in all cells. In contrast, the multiple enzymatic steps that lead to the production of a highly active steroid hormone (eg, cortisol or 1,25-dihydroxyvitamin D) are restricted, with rare exceptions, to steroid-producing cells or their precursor cells. Hence, the occurrence of 1,25-dihydroxyvitamin D excess caused by a tumor is distinctly unusual, being observed infrequently in hematologic malignancies and only in those with the ability to 1-hydroxylate 25-hydroxyvitamin D, the immediate precursor of 1,25-dihydroxyvitamin D. Ectopic hormone syndromes—the
most common of the paraneoplastic syndromes—thus predominantly reflect peptide hormone excess states. The most common ectopic peptide hormone syndromes are described in greater detail in subsequent sections of this chapter.
Table 21-1. Polypeptide hormones produced ectopically by benign and malignant tumors and their associated endocrinopathies.
APUD Concept of Neuroendocrine Cell Tumors
Over the years since the initial recognition that nonendocrine tumors were the source of the ectopic hormones produced in these endocrine syndromes, the notion developed that the hormones originated from highly specialized neuroendocrine cells in tumors. These cells were thought to derive from the neural crest and were postulated to be able to synthesize and store biogenic amines and were thus designated amine precursor uptake and decarboxylation (APUD) cells. Neuroendocrine cells, like calcitonin-secreting C cells and adrenal chromaffin cells, clearly had these properties, and tissues giving rise to ectopic hormone syndromes (eg, lung and gastrointestinal tract) also had APUD cells scattered throughout them. It was originally thought that the tumor cells producing excessive amounts of polypeptide hormones were derived exclusively from APUD cells in the tissue of origin (eg, ACTH-producing cells of the lung).
Newer insights into tumor cell biology have led to a better understanding of the pathogenesis and etiology of the ectopic hormone production states. Studies have shown that not all APUD cells are derived from the neural crest. Some of these cells are of endodermal origin. Furthermore, with careful examination of tumors, it has become clear that peptide hormones are often produced by non-APUD cells. It has, however, been appreciated that the clinically evident hormone hypersecretion states are typically caused by tumors that are in fact derived from APUD cells. These cells are the ones with the full capacity to produce and store peptide hormones efficiently in dense secretory granules and then release biologically significant quantities of active hormones in circulating plasma.
HYPERCALCEMIA OF MALIGNANCY
Hypercalcemia, the most common paraneoplastic endocrine syndrome, occurs in 10–15% of malignancies. In the majority of patients (98%), the identity of the tumor is apparent at the time of presentation, and the prognosis is very poor as most patients with hypercalcemia of malignancy do not survive beyond 6 months.
Enhanced bone resorption is the primary cause of hypercalcemia of malignancy. Tumor-derived factors induce this increase in osteoclast-mediated resorption via two distinct mechanisms: (1) humoral effects of systemically elevated tumor-derived factors and (2) local autocrine or paracrine effects of factors produced by tumor cells that have metastasized to bone and induce localized osteolysis (Figure 21-1). While the latter mechanism was thought to be the primary cause of hypercalcemia of malignancy when it was initially described in the 1920s, work within the last 2 decades has in fact identified a humoral basis as the most frequent (80%) cause even in settings, such as breast cancer, where lytic bone metastases are present. Decreased renal calcium excretion may also contribute to the pathogenesis, either because of the hypocalciuric effects of certain humoral mediators of hypercalcemia, such as PTH-related protein (discussed below), or because of the decreased glomerular flow that occurs with hypercalcemia-induced nephrogenic diabetes insipidus.
Figure 21-1. Bone histology in hypercalcemia of malignancy versus primary hyperparathyroidism. A: Local osteolytic hypercalcemia due to leukemia. Note the presence of tumor cells in marrow spaces and numerous bone-resorbing osteoclasts lining the trabecular surface.B: Humoral hypercalcemia due to squamous cell carcinoma. Note the absence of tumor cells in marrow spaces but the presence of numerous active osteoclasts on the trabecular surface. Note also the absence of bone-forming osteoblasts, consistent with uncoupling of bone formation and resorption. C: Hyperparathyroidism. Note abundant osteoblasts (small arrows), osteoclasts (large arrow), and osteoid. (Panel A is reproduced, with permission, from Stewart AF, Insogna KL, and Broadus AE: Malignancy-associated hypercalcemia. In: Endocrinology, 3rd ed. DeGroot LJ [editor]. Saunders, 1995. Panels B and C are reprinted, with permission, from Stewart AF, Vignery A, Silverglate A, Ravin ND, LiVolsi V, Broadus AE: Quantitative bone histomorphometry in humoral hypercalcemia of malignancy: uncoupling of bone cell activity. J Clin Endocrinol Metab 1982;55:219. Copyright Š 1982 by The Endocrine Society.)
In the 1940s, Fuller Albright, in describing a case of hypercalcemia of malignancy occurring in the absence of significant bone metastases, was the first to propose the existence of a humoral cause of this syndrome. It was not until the late 1980s, however, that this humoral factor was identified. Unlike most other paraneoplastic endocrine syndromes which are due to the ectopic production of well-described hormones with known physiologic functions, the vast majority of cases of humoral hypercalcemia of malignancy are due to the overexpression of PTH-related protein (PTHrP) (Figure 21-2). PTHrP is a peptide that had not previously been identified until it was isolated simultaneously by several independent groups in 1987 from tumors commonly associated with humoral hypercalcemia of malignancy—squamous cell carcinoma of the lung, breast carcinoma, and renal carcinoma.
The amino terminal portion of PTHrP bears strong homology to PTH and binds with equal affinity to PTH receptors (now known as the PTH/PTHrP-1 receptor subtype) in bone and kidney. Therefore, the biochemical markers of PTHrP-mediated hypercalcemia in
vivo are similar to those of hyperparathyroidism and include a decrease in serum phosphate and an increase in nephrogenous cAMP. Some unexplained differences, however, are seen in PTHrP-mediated hypercalcemia of malignancy (versus primary hyperparathyroidism), including normal or suppressed 1,25-dihydroxyvitamin D levels and an uncoupling of bone resorption and formation that results in severe bone loss (Figure 21-1). The reasons for these differences are not yet known but may include the ability of chronic PTHrP (versus intermittent PTH) stimulation or profound hypercalcemia per se to suppress 1,25-dihydroxyvitamin D levels and, the contributions of additional tumor-derived cytokines, such as interleukin-1α or interleukin-6, to the process of bone resorption.
Figure 21-2. Serum levels of amino terminal PTHrP (iPLP[1-34]) in normal subjects, patients with hyperparathyroidism, normocalcemic patients with malignancy, and hypercalcemic patients with malignancy. In this series, 55% of the patients with malignancy-associated hypercalcemia had serum PTHrP levels that exceeded the upper limits of normal, including patients with solid tumors, breast carcinoma, and hematologic malignancies. HPT = hyperparathyroidism; open squares = undetectable PTHrP plotted at detection limit. (Reprinted, with permission, from Budayr AA et al: Increased serum levels of a parathyroid hormone-like protein in malignancy-associated hypercalcemia. Ann Intern Med 1989:111:807.)
Normal physiologic functions of PTHrP are still being investigated. PTH-related peptide and PTH are ancestrally related genes that have evolved separately. Consistent with this, while PTH is produced primarily at one site, the parathyroid gland, and acts as a calciotropic hormone, PTHrP is produced by a wide variety of cell types and exhibits diverse functions, most of which are unrelated to calcium homeostasis. The normal actions of PTHrP are autocrine or paracrine, rather than humoral, as both PTHrP and the PTH-PTHrP receptor are now known to be expressed in numerous organs where PTHrP is thought to act locally. Known functions of PTHrP include (1) regulation of endochondral bone formation during development; (2) growth and differentiation of mammary gland, skin, and pancreatic islets; (3) relaxation of vascular and nonvascular smooth muscle; (4) cytokine-like effects during the inflammatory response; (5) neuroprotective effects during aging; and (6) transepithelial calcium transport in the placenta.
While PTHrP is by far the most common mediator of hypercalcemia of malignancy, other calciotropic hormones can also cause the syndrome. To date, ectopic production of PTH has been reported as a cause of hypercalcemia of malignancy only in seven isolated cases, including several neuroendocrine tumors and small cell carcinomas. Additionally, elevated circulating levels of 1,25-dihydroxyvitamin D levels, thought to be due to increased 1α-hydroxylase activity in lymphoproliferative cells, can cause humoral hypercalcemia of malignancy in various types of lymphoma, including Hodgkin's disease, B cell lymphomas, T cell lymphomas, and Burkitt's lymphoma. Other tumor-derived bone-resorbing
factors, such as prostaglandins, may also contribute to hypercalcemia of malignancy in certain cases.
Tumors Associated with Hypercalcemia of Malignancy
Hypercalcemia of malignancy occurs frequently with certain common tumors, including squamous cell carcinoma of the lung and breast carcinoma. In contrast, hypercalcemia is rarely seen in other commonly occurring cancers (eg, colon, gastric, thyroid, and oat cell carcinomas), including tumors such as prostate carcinoma that are frequently metastatic to bone.
- SQUAMOUS CELL CARCINOMAS
Squamous cell carcinomas account for over one-third of all cases of hypercalcemia of malignancy. Humoral effects of tumor-derived PTHrP account for most cases of hypercalcemia in this setting. Twenty-five percent of patients with squamous cell lung carcinomas develop PTHrP-mediated hypercalcemia, while other sites of squamous cell carcinoma (head, neck, esophagus, cervix, vulva, and skin) are similarly associated with a high incidence of hypercalcemia.
- RENAL CELL CARCINOMA
Renal cell carcinoma is another solid tumor that is frequently associated with hypercalcemia. As with squamous cell carcinomas, humoral effects of tumor-derived PTHrP are causative. While 100% of renal cell tumors in some series have been reported to stain positively for immunoreactive PTHrP, hypercalcemia is only seen in 8% of cases.
- BREAST CARCINOMA
Hypercalcemia occurs in 20% of patients with advanced-stage breast carcinoma and is only rarely seen in the absence of bone metastases. However, the majority of patients with bone metastases (ie, 70% of those with advanced disease) do not have hypercalcemia. Despite the frequent presence of bone metastases in patients with hypercalcemia, their hypercalcemia is in fact (and paradoxically) mostly humorally mediated. Humoral effects of PTHrP have been shown to be responsible for up to 65% of cases of breast carcinoma-associated hypercalcemia, as evidenced by elevated circulating levels of PTHrP occurring in association with increased levels of nephrogenous cAMP. In the remainder of cases, localized osteolysis, induced by factors produced by tumor cells metastatic to bone, is the cause of hypercalcemia. Even in this setting, however, PTHrP contributes to the etiology of hypercalcemia, as 97% of breast carcinoma bone metastases have been shown to be PTHrP-positive. Evidence obtained from an elegant series of studies using an animal model of metastatic breast cancer now suggests that (1) local bone factors, such as TGF-β, may enhance PTHrP expression in tumor cells metastatic to bone, even when the primary tumor is PTHrP-negative; and (2) this local increase in PTHrP production enhances osteoclast-mediated destruction of adjacent bone. Additional, locally produced bone cytokines such as interleukin-1α, interleukin-6, and tumor necrosis factor-α probably act in concert with PTHrP at sites of bone metastases to enhance osteoclast-mediated lytic bone destruction and thus contribute to hypercalcemia. While the hypercalcemia seen in late stages of cancer is usually unremitting and associated with a survival time of weeks to months, hypercalcemia in breast cancer can be episodic. For example, hypercalcemia often occurs transiently in response to treatment with estrogens or antiestrogens and is associated with a beneficial response to therapy. Recent demonstrations of enhanced PTHrP expression in breast carcinoma cells in response to estrogen or tamoxifen suggests that these episodes of transient hypercalcemia may also be PTHrP-mediated.
- OTHER SOLID TUMORS
Hypercalcemia is also associated with other solid tumors, though less frequently. In most of these cases, humoral effects of PTHrP are causative. PTHrP-mediated hypercalcemia has been reported for ovarian carcinomas, a tumor type for which one case of tumor-derived, PTH-mediated hypercalcemia has also been reported; bladder carcinoma; large cell and adenocarcinoma of the lung; and endocrine tumors, including islet cell tumors, pheochromocytoma, and carcinoid tumors.
- MULTIPLE MYELOMA
Most patients with multiple myeloma have extensive bone destruction. However, only 30% of patients develop hypercalcemia, which can be remitting. Although systemic increases in circulating PTHrP have been reported in some cases, locally produced osteolytic factors—rather than humoral mediators—are thought to be responsible for most cases of hypercalcemia in this hematologic malignancy. While the factor or factors stimulating osteoclast activity in this disease have not yet been identified, studies suggest that myeloma cells in the marrow express cytokine-like factors such as TNF-α, TNF-β, IL-1α, IL-1β, IL-6, and PTHrP, which act locally to stimulate the bone-resorbing activity of adjacent osteoclasts. Because renal disease occurs frequently in myeloma due to the filtration of Bence Jones proteins (light chain fragments of IgG), it is hypothesized that patients with renal impairment may be predisposed to development of hypercalcemia in this setting of increased bone resorption.
Hypercalcemia occurs in association with 1-2% of lymphomas and leukemias, is seen primarily in patients with bone involvement, and can occur with a variety of cell types. With the exception of human T cell leukemia virus (HTLV-1)-induced adult T cell leukemia or lymphoma, which will be discussed below, approximately half of the cases of lymphoma-associated hypercalcemia are thought to be due to the local lytic effects of tumor-derived factors, as was discussed for myeloma. The remainder appear to be mediated by a mechanism unique to lymphoma—ie, humoral effects of tumor-derived 1,25-dihydroxyvitamin D. The mechanism leading to increased circulating 1,25-dihydroxyvitamin D in lymphoma is believed to be the same as that seen in hypercalcemic granulomatous disorders, ie, increased production of 1,25-dihydroxyvitamin D by the involved hematopoietic cells due to 1α-hydroxylation of circulating epidermal- and diet-derived 25-hydroxyvitamin D. Both increased intestinal calcium absorption and increased bone resorption are thought to contribute to hypercalcemia in this setting. HTLV-1-induced adult T cell leukemia or lymphoma must be considered separately when evaluating possible causes of hypercalcemia. In this subset of patients, hypercalcemia occurs in two-thirds of cases, responds poorly to treatment, and is due to the humoral effects of tumor-derived PTHrP which is induced by a viral transactivating factor (Tax).
The diagnosis of hypercalcemia is discussed in detail in Chapter 8. Primary hyperparathyroidism and hypercalcemia of malignancy account for over 90% of all causes of hypercalcemia. Because the incidence of primary hyperparathyroidism is twice that of hypercalcemia of malignancy, primary hyperparathyroidism must also be considered as a potential cause of hypercalcemia in patients with malignancy and can be simply evaluated with current methods by determination of intact PTH levels using a standard two-site immunoradiometric assay. In the setting of hypercalcemia of malignancy and normal renal function, PTH will be suppressed. Further evaluation can be guided, in part, by the tumor type. Elevated PTHrP will be detected in a majority of cases of hypercalcemia of malignancy associated with solid tumors, including breast carcinoma, and in HTLV-1-induced T cell lymphomas. Measurement of 1,25-dihydroxyvitamin D should be considered in all cases of lymphoma-associated hypercalcemia. Lytic bone lesions are usually identified during tumor staging.
The treatment of hypercalcemia is discussed in detail in Chapter 8. Because bone resorption is central to all causes of hypercalcemia of malignancy, bisphosphonates are a mainstay of treatment. Moreover, recent evidence suggests that these agents, in addition to reversing hypercalcemia due to both humoral and local lytic factors, may also prevent the progression of bone metastases, particularly in multiple myeloma. Glucocorticoids can be used with some success in the treatment of all causes of hypercalcemia of malignancy but may be particularly efficacious when hypercalcemia is due either to the local lytic effects of neoplastic plasma cells in multiple myeloma or to the increased production of 1,25-dihydroxyvitamin D in lymphoma.
ECTOPIC CUSHING'S SYNDROME
Many tumors produce the ACTH precursor proopiomelanocortin (POMC), and only a fraction of such tumors release sufficient ACTH to cause Cushing's syndrome. Initially, the tumors recognized to cause this syndrome were of nonpituitary origin but were endocrine tumors, such as islet cell carcinomas and pheochromocytomas. Subsequently, a wide variety of different tumor cell types, both endocrine and nonendocrine, have been associated with the “ectopic” ACTH syndrome.
The classic description of the ectopic ACTH syndrome was made by Grant Liddle and coworkers in the early 1960s and was based on a series of patients who mostly had highly malignant tumors (eg, oat cell or small cell carcinoma of the lung). More recently, the ectopic ACTH syndrome has been recognized with increasing frequency with benign tumors, specifically carcinoids. Benign (versus malignant) lesions typically present in a more subtle clinical manner, often over months to years before the tumor is identified. The more gradual development of the clinical syndrome plus the more subtle biochemistry have led to a considerable challenge in distinguishing this form of the ectopic ACTH syndrome from pituitary tumors causing Cushing's disease. This subtle variant of tumor-induced ACTH excess has been dubbed the “occult” ectopic ACTH syndrome. In addition, it is now recognized that tumors can cause an ectopic ACTH-like syndrome through production of corticotropin-releasing hormone (CRH). Indeed, some of the tumors that make the latter cosecrete ACTH as well. Ectopic CRH production has been seen in bronchial carcinoids, medullary thyroid carcinoma, and metastatic prostatic cancer.
Cushing's syndrome—signs and symptoms resulting from unregulated production of glucocorticoids—is caused by a number of underlying disturbances. These must be differentiated to ensure successful treatment. Causes include pituitary ACTH-dependent Cushing's disease, adrenal tumors or ACTH-independent Cushing's syndrome, and the ectopic ACTH syndrome. In several large series, it has been reported that in 50–80% of patients with Cushing's syndrome there is a pituitary cause. Adrenal adenomas (and very rarely carcinomas) account for 5–30% of cases of Cushing's syndrome. The ectopic ACTH syndrome comprises approximately 10–20% of cases of Cushing's syndrome from referral center populations.
A wide variety of tumors cause ectopic ACTH syndrome (Table 21-2). In the classic and initial descriptions of this syndrome, there was a preponderance of malignant tumors, particularly small cell carcinomas of the lung. It is now clear that most cases of ectopic ACTH syndrome are due to benign tumors. Most recently, microscopic carcinoid “tumorlets,” particularly in the lung, have been recognized to cause occult ectopic ACTH syndrome. These tumors may be exceptionally difficult to diagnose by standard techniques.
The diagnosis of Cushing's syndrome requires a rigorous approach. Cushing's syndrome should be suspected first on solid clinical grounds and then established biochemically. This is accomplished by demonstrating the presence of hypercortisolism—a frankly elevated 24-hour urinary free cortisol level or the lack of suppression of plasma cortisol levels after a 1 mg overnight dexamethasone suppression test (seeChapter 9). In Cushing's syndrome due to any cause and often in ectopic ACTH syndrome, random cortisol levels are elevated. Once hypercortisolism is established, plasma ACTH levels are measured. These levels are markedly elevated in classic forms of the ectopic ACTH syndrome, typically secondary to malignant lung neoplasms. There is, however, considerable overlap between the milder cases of the ectopic ACTH syndrome, caused by benign and slowly growing tumors, and Cushing's disease due to a pituitary tumor. In the former case, the tumors are often small and clinically silent—hence the descriptor “occult” ectopic ACTH syndrome. For these reasons, rigorous biochemical criteria must be applied in appropriate clinical situations to make certain that the correct diagnosis is made. Plasma ACTH levels in patients with clinically evident tumors are often strikingly elevated (390–2300 pg/mL [87–511 pmol/L] by radioimmunoassay). Individuals with ectopic ACTH syndrome due to occult tumors have ACTH levels that overlap with pituitary-dependent Cushing's disease (42–428 pg/mL [9.3–95 pmol/L]). It is said that patients with plasma ACTH levels greater than 200 pg/mL [44.4 pmol/L] typically have the ectopic ACTH syndrome, though further testing must be done to prove this and to localize the tumor.
Table 21-2. Tumors responsible for the ectopic ACTH syndrome.1
After hypercortisolism and ACTH excess are established, the degree of suppressibility of ACTH with exogenous glucocorticoid is determined. In classic Cushing's disease due to a pituitary tumor, supraphysiologic doses of dexamethasone usually suppress the elevated plasma ACTH and cortisol levels. Tumors responsible for the ectopic ACTH syndrome are, however, classically unresponsive to these doses of dexamethasone. High-dose dexamethasone suppression testing, as this diagnostic maneuver is called, is accomplished (1) by administering 2 mg of dexamethasone every 6 hours for 2 days and measuring urinary free cortisol or plasma cortisol on the second day, or (2) by administering 8 mg of dexamethasone the night before obtaining an 8 AM plasma cortisol level. In both tests, the expected suppression of baseline urinary free cortisol and plasma cortisol should be 50% or greater if the Cushing's syndrome is due to a pituitary adenoma (ie, Cushing's disease). However, between 15% and 33% of patients with ectopic ACTH syndrome will also meet these suppression criteria (false positives), mimicking Cushing's disease. In addition, 10–25% of patients with Cushing's disease fail to suppress with high-dose dexamethasone (false negatives). The overnight test probably has greater sensitivity and accuracy than the classic 2-day test and is preferred.
Two additional tests have been developed to improve the diagnostic discrimination between Cushing's disease and ectopic ACTH syndrome. The first is CRH testing. Pituitary corticotrophs are normally responsive to CRH in Cushing's disease and unresponsive when ectopic ACTH production or an adrenal lesion is responsible for the cortisol excess. A positive response to CRH is defined as a 50% or greater increase in plasma ACTH and a 20% or greater increase in the plasma cortisol concentrations. An increase in ACTH of 100% and in cortisol of over 50% greatly reduces the likelihood of ectopic ACTH syndrome; however, false-positive and false-negative tests (up to 10%) have been reported. Moreover, in the rare instance of ectopic production of CRH (without concomitant ACTH) by a tumor, a false-positive result may be seen, leading to the erroneous diagnosis of pituitary-dependent Cushing's disease. For these reasons, many clinicians sample the inferior petrosal sinuses for plasma ACTH levels both before and after the injection of CRH to assist with the differential diagnosis. These sinuses drain the pituitary gland. Concomitant peripheral and petrosal sinus samples are obtained, and the central:peripheral ACTH ratio is calculated. In Cushing's disease, the ratio should be ą 2.0 in the basal state. After CRH administration, this ratio should be ą 3.0 in pituitary-dependent Cushing's disease. In the ectopic ACTH syndrome, this ratio should not rise after the CRH. In rare instances of ectopic CRH syndrome, the basal ratio may be 2.0. The stimulation by CRH gives close to 100% discrimination between ectopic ACTH production and a pituitary tumor secreting ACTH. Generally, a combination of tests is performed to reach a biochemical diagnosis before extensive radiologic studies are undertaken.
The majority of patients (70% or more) with ectopic ACTH syndrome will also cosecrete other hormones or tumor marker peptides, among them carcinoembryonic antigen, somatostatin, calcitonin, gastrin, glucagon, vasoactive intestinal peptide, bombesin, pancreatic polypeptide, alpha-fetoprotein, and many others. The presence and secretion of these other hormones (in addition to ACTH) suggests that the source of ACTH is nonpituitary in these patients. Given the variety of peptides and the expense inherent in any screening paradigm, measuring a panel of these hormones in patients suspected of ectopic ACTH syndrome is not recommended.
The path to localization of the tumor responsible for ectopic ACTH production generally starts with a chest radiograph. Most tumors are in the chest or abdomen. Small cell carcinomas of the lung are usually visible on chest x-ray, while bronchial carcinoids are often difficult to detect by plain radiographs. In some situations, these tumors may require a long period (as many as 4 or 5 years) of close follow-up before the tumors are detected. Chest CT scanning should be employed in all subjects with ectopic ACTH to rule out a chest or mediastinal lesion (such as a thymic carcinoid). Abdominal CT scanning is also performed in these patients to confirm the presence of bilateral adrenal enlargement, a sine qua non of the ectopic ACTH syndrome, and to screen for other possible abdominal tumors responsible for the syndrome (pheochromocytoma, islet cell tumor, etc.). In the radiologic evaluation of Cushing's syndrome, it is always important to bear in mind that the presence of a pituitary microadenoma on MRI does very little to support the diagnosis of pituitary-dependent Cushing's disease—as opposed to an ectopic tumor producing ACTH—because of the great numbers (10–20%) of normal individuals with incidental pituitary microadenomas (see Chapter 5).
Octreotide scanning, another important diagnostic technique, can successfully localize tumors responsible for ectopic ACTH production. It relies on the expression of somatostatin or octreotide receptors in neuroendocrine cells. Iodinated or, more recently, indium-111-labeled octreotide scanning has demonstrated medullary carcinomas of the thyroid, small cell lung cancers, islet cell tumors, pheochromocytomas, and other tumors. There remains controversy among experts about whether this form of scanning exceeds the sensitivity and specificity of thin-cut CT-MR scans of the chest and abdomen reviewed by expert radiologists. Octreotide therapy has been used to lower ACTH levels and relieve symptoms of cortisol excess in many of these tumors.
Cushing's syndrome causes truncal obesity, violaceous striae, hypertension, fatigue, glucose intolerance, osteopenia, muscle weakness, moon facies, easy bruisability, buffalo hump, depression, hirsutism, and edema. Patients with ectopic ACTH syndrome may show some, all, or none of these features depending on the underlying tumor. It has been appreciated from the initial descriptions of this syndrome that these patients typically present with myopathy, weight loss, and electrolyte and metabolic disturbances more commonly than with the classic features of slowly developing Cushing's disease. Hyperpigmentation is also recognized as more common in the ectopic ACTH syndrome than in Cushing's disease. Cortisol excess in older men, especially those at risk for lung tumors, is most commonly due to ectopic ACTH production, whereas ACTH-producing pituitary tumors predominate in young and middle-aged women. Glucose intolerance or frank diabetes and hypokalemic alkalosis are typical metabolic disturbances of the ectopic ACTH syndrome.
Because of the extreme elevation in plasma cortisol levels in many of these patients, they are at considerable risk for and often succumb to overwhelming opportunistic infections, often with fungal pathogens.
A critical caveat to remember in the clinical evaluation of patients with ACTH-dependent Cushing's disease is that slowly growing and occult tumors producing ACTH may present in exactly the same way as classic Cushing's disease due to a pituitary tumor. Therefore, both the clinical findings and the laboratory studies summarized above show considerable overlap and may engender confusion in distinguishing these occult tumors from a pituitary lesion.
Increasing numbers of patients who have classic features of Cushing's syndrome have been shown to have adrenal expression of ectopic receptors as the cause of their hypercortisolism. The pathophysiology of this form of Cushing's syndrome is ACTH-independent since other hormones are driving the glucocorticoid hypersecretion. Ectopic expression of receptors for gastric inhibitory peptide, vasopressin, serotonin,β-adrenergic agonists, luteinizing hormone (LH), and interleukin-1 have been reported. In the case of gastric inhibitory peptide, food-stimulated cortisol hypersecretion has been described. In a case report of ectopic LH receptor expression in the adrenals associated with macronodular adrenal hyperplasia, the patient had mild cushingoid features with pregnancy and the gradual development of full-blown Cushing's syndrome with menopause. Thus, it has become increasingly evident that ectopic receptors as well as hormones can be responsible for hypercortisolemic states.
SYNDROME OF INAPPROPRIATE ADH (SIADH) SECRETION
This syndrome is characterized by inappropriate retention of water such that hypertonic or inappropriately concentrated urine is excreted in the presence of hyponatremia. There are many well-recognized causes of this syndrome (Table 21-3).
Etiology & Pathogenesis
Tumors are a common cause of SIADH (see Table 21-3). Bronchogenic carcinoma, particularly small cell carcinoma, has been associated with this syndrome since its initial description in 1957. Small cell carcinoma accounts for 80% of cases of SIADH, but only 3-15% of patients with this tumor have SIADH. Most of these tumors, even from patients without the clinical syndrome, contain ADH by immunostaining. Other tumors that cause the syndrome include breast, pancreatic, and thymic carcinomas in addition to those listed in Table 21-3. These tumors typically produce AVP, oxytocin, and the carrier protein neurophysin.
Table 21-3. Causes of SIADH.1
Excessive production of vasopressin by tumors leads to an inability to excrete free water. This is due to the fact that tumors release vasopressin independent of serum osmolality. In addition to the production of vasopressin, many tumors also contain the mRNA for the hormone atrial natriuretic peptide (ANP). This has suggested that hyponatremia in these patients may be due to the natriuretic effects of ANP rather than the antidiuretic effects of vasopressin.
Clinical & Laboratory Features
SIADH is the most common cause of hyponatremia in hospitalized patients. It may present with symptoms due to water intoxication and hyponatremia. Although many patients are asymptomatic, depending on the
magnitude and chronicity of their hyponatremia, symptomatic individuals usually have fatigue, headache, nausea, and anorexia initially which can progress to altered mental status, seizures, coma, and even death. Most patients will experience weight gain due to water retention but will not have edema. Significant clinical symptoms usually do not develop unless the serum sodium is 125 mEq/mL or less, and there is usually a correlation between the level of symptomatology in these patients and their serum sodium values.
Patients with SIADH exhibit hyponatremia, serum hypoosmolality, a less than maximally dilute urine, and the presence of sodium in the urine. Clinically, the diagnosis of SIADH cannot be made unless there is euvolemia with intact renal, adrenal, and thyroid function. Cirrhosis, nephrosis, and congestive heart failure must be excluded. Generally, the diagnosis is made by the presence of hyponatremia, low serum osmolality, and urine osmolality that is less than maximally dilute. Urinary sodium levels are usually high, and urea nitrogen levels are typically low, as are serum uric acid levels. Rarely is it necessary to measure vasopressin levels to make this diagnosis, although these determinations are now widely available. It is rarely if ever necessary to perform a water-loading test, which can be dangerous in these patients because of their impaired ability to excrete a free water load and their propensity to become water-intoxicated. Once the diagnosis of SIADH is made, all possible causes should be considered (see Table 21-3). Water restriction and demeclocycline are the mainstays of treatment. In an acutely symptomatic patient or when the serum sodium is reduced to dangerously low levels, infusion of hypertonic saline and administration of loop diuretics (eg, furosemide) are the treatments of choice. Many patients with SIADH due to neoplasms will improve and even remit with effective therapy for their underlying cancer.
NON-ISLET CELL TUMORS & HYPOGLYCEMIA
Tumors that cause hypoglycemia are quite rare—especially those derived from tissues other than the pancreatic islets. These tumors are usually large mesenchymal lesions in the chest or abdomen and can be benign or malignant. They include fibromas, fibrosarcomas, mesotheliomas, hemangiopericytomas, lymphomas, hepatomas, hypernephromas, and adrenocortical carcinomas. The hallmark of this clinical syndrome is fasting hypoglycemia. Its differential diagnosis and clinical presentation are discussed more fully in Chapter 18.
The pathogenesis of the hypoglycemia in these cases may involve a variety of mechanisms, including excessive consumption of glucose by what is typically a large tumor; ectopic or abnormal secretion of insulin or insulin-like growth factor-II (IGF-II) and IGF-binding proteins; or inadequate production of counterregulatory hormones such as growth hormone. Examples of all of the above mechanisms in relation to specific tumors have been described. The typical clinical presentation of these patients is the occurrence of increased levels of IGF-II; suppressed levels of insulin, C-peptide, proinsulin, and IGF-I; and lower than expected levels of growth hormone. It is very uncommon to find a non-islet cell tumor that can produce authentic insulin—only rarely has this been well-documented. Patients with tumor-induced hypoglycemia can manifest all the signs and symptoms of fasting hypoglycemia: sweating, intense hunger, anxiety, altered consciousness, and visual and behavioral changes. The presenting symptoms may be subtle, and rarely is the underlying diagnosis suspected at initial presentation.
Several cases of non-islet cell tumors causing hypoglycemia have been carefully studied and reported in the literature. Daughaday and coworkers reported in 1981 that the causative hormone in these cases was probably IGF-II. Subsequently, further work in his laboratory and in others has shown that the IGF-II produced is generally of a higher molecular weight than normal IGF-II—so-called “big IGF-II” with a molecular weight of 11-18 kDa (normal MW of IGF-II is 7.5 kDa). It has been hypothesized that “big” IGF-II is present in excess in these patients because the tumor fails to process IGF-II appropriately. Normally, IGF-II is produced by the liver and circulates bound mainly to IGF binding protein-3 and an acid-labile subunit in a heterotrimeric complex. Fully processed IGF-II from hepatic sources does not cause hypoglycemia because it is sequestered in this complex and is unavailable to interact with receptors. “Big” IGF-II causes hypoglycemia by one of two mechanisms: (1) It does not readily associate into this ternary complex and has greater access to insulin receptors and hence greater biologic activity, or (2) it is produced in excess and can readily bind up all available IGF-BP3 but there is sufficient unbound (“free”) IGF-II to interact with insulin receptors and cause hypoglycemia. In tumors causing hypoglycemia, it is estimated that as much as 80% of the circulating IGF-II is free. Increased free IGF-II may also alter the levels of binding proteins, among them IGF binding protein-3 and the acid-labile subunit. Thus, there are a number of possible explanations for this clinical syndrome.
Treatment of this paraneoplastic syndrome usually involves surgery to debulk the tumor. If the lesion is benign, this usually brings relief of the hypoglycemia or even definitive cure. Radiotherapy may also be employed adjunctively. These patients often require continuous glucose infusions to control their symptoms
prior to surgery, and glucagon can be used acutely to raise blood glucose levels. In occasional patients, diazoxide therapy has been useful. In one study of a small number of patients, glucocorticoids reduced IGF-II levels, thereby restoring a more normal IGF and IGF-binding protein profile. Owing to the size and advanced stage of clinical progression of these tumors at the time of diagnosis, the outcome is often poor.
OTHER HORMONES SECRETED BY TUMORS
- Growth Hormone-Releasing Hormone & Growth Hormone
It was recognized in the 1960s that carcinoid tumors are associated with acromegaly. This led to the idea that these tumors could secrete a growth hormone-releasing factor. In 1982, two laboratories reported the purification of the hypothalamic peptide growth hormone-releasing hormone (GHRH). Several biologically active forms of the 44-amino-acid peptide GHRH are typically present in extracts from tumors responsible for ectopic GHRH production. Patients whose tumors release excessive quantities of this peptide develop acromegaly. This classic disorder is characterized by acral enlargement, coarsened facies, soft tissue overgrowth, excessive sweating, arthropathy, and other manifestations (see Chapter 5). Most cases of acromegaly, however—over 90%—are due to pituitary tumors secreting excess GH. Patients with acromegaly due to the ectopic production of GHRH will manifest increases in serum growth hormone (GH) and insulin-like growth factor-I (IGF-I) levels. Circulating levels of GHRH are also elevated, and this is an absolutely critical measurement to make. Evaluation of the pituitary gland typically shows no tumor. The pituitary pathology in patients with ectopic GHRH secretion is usually somatotroph hyperplasia.
In rare instances, hypothalamic tumors such as hamartomas, ganglioneuromas, and gangliocytomas produce excessive GHRH and acromegaly. Since the site of GHRH production is within the hypothalamus, this is not considered ectopic GHRH secretion. The syndrome of ectopic GHRH secretion has been recognized in patients with tumors outside the hypothalamus, most frequently carcinoid tumors of the lung and gastrointestinal tract and islet cell tumors. Other tumors associated with this syndrome include pheochromocytoma, paraganglioma and adenocarcinoma of the lung, hepatic neuroendocrine tumors, and pituitary adenoma. In the latter case, adenoma cells stained positively for both GH and GHRH by immunocytochemistry, and plasma GHRH levels were elevated. Thus, in this situation, there was both autocrine and paracrine regulation of GH secretion and possibly tumor progression.
Because of the nonmalignant nature of most of the tumors responsible for the ectopic GHRH syndrome, the presence of a tumor outside the pituitary or hypothalamus is often not suspected for many years. Symptoms due to the presence of a tumor outside the pituitary gland (such as gastrointestinal or pulmonary complaints) may be the first clue that the acromegaly is due to a nonpituitary tumor. Dynamic testing can also provide a clue that classic pituitary-dependent acromegaly is not the cause of the GH excess. Tumors releasing excessive GHRH may exhibit a GH increase with thyrotropin-releasing hormone (TRH) or with glucose administration and are more likely to have elevated prolactin levels—compared with classic pituitary tumors secreting GH. None of these features allow definitive diagnosis of the ectopic GHRH syndrome, since they are also observed with GH-secreting pituitary tumors. As a group, however, these features can increase the suspicion that one is confronted with an ectopic source of GHRH.
Tumors responsible for the ectopic GHRH syndrome in the chest or mediastinum can often be localized by chest x-ray and CT scanning. In the case of abdominal tumors, endoscopic ultrasound or abdominal CT scanning or MRI may be necessary. Since somatostatin receptors are often present in these tumors, octreotide scanning may also be helpful. Given the low prevalence of the ectopic GHRH syndrome among patients with acromegaly, it is not recommended to screen all acromegalic patients for this possible cause. Rather, it is advised that GHRH levels be measured in patients with any atypical features of acromegaly. If elevated GHRH is established, careful investigation to determine its source is indicated.
There are rare reports of ectopic production of GH by malignant tumors outside the pituitary-hypothalamic region. Acromegaly has been reported due to GH overproduction by a pancreatic islet cell tumor or by a non-Hodgkin lymphoma.
Calcitonin is one of the hormones produced most commonly by tumors. Estimates indicate that 10–30% of malignancies actually make calcitonin or its precursor form procalcitonin. Since calcitonin excess does not produce a clinically evident ectopic hormone syndrome, these patients rarely come to medical attention for their excessive calcitonin secretion. Calcitonin levels have been used as a tumor marker in following responses to treatment.
In addition to medullary carcinoma of the thyroid (discussed in Chapter 7 and 24), in which calcitonin
hypersecretion is eutopic (not ectopic), the commonest tumors in which excessive calcitonin is produced include small cell cancers of the lung and pulmonary carcinoids. In studies of the lung malignancies that synthesize calcitonin, it has become clear that pulmonary neuroendocrine cells contain large amounts of calcitonin. These cells are thought to be the cells of origin for lung carcinoid tumors and small cell cancers. In other primary lung cancers (non-small cell), it has been observed that there is often accompanying pulmonary neuroendocrine cell hyperplasia (perhaps secondary to chronic smoking) that may account for the calcitonin hypersecretion. In other tumors not derived from the lung, the cell which is the source of the calcitonin has not been fully elucidated. In addition, in many tumors, the larger forms of calcitonin—a hormone that has complex transcriptional and posttranscriptional mechanisms governing its expression and processing—are made preferentially by these tumors. Such cancers are thought to lack the specialized enzymes required for the final processing of calcitonin.
Gonadotropins are glycoprotein hormones composed of two subunits: alpha and beta. The alpha subunit is shared by thyrotropin, follicle-stimulating hormone, luteinizing hormone, and chorionic gonadotropin (hCG). The first three hormones are pituitary products, while the latter is a product of the syncytiotrophoblast of the placenta. Although hCG is expressed in nearly all normal tissues, it does not circulate to any appreciable extent except in pregnancy. Trophoblastic tumors like hydatidiform mole and gonadal and nongonadal choriocarcinoma often secrete excessive hCG, which thus is a useful tumor marker in these conditions. Trophoblastic tumors are derived from tissues that have the capacity to make hCG normally; therefore, their production of this hormone is not usually considered as “ectopic.” Tumors that do make sufficient quantities of hCG ectopically to raise circulating levels of the hormone include ovarian, prostatic, and testicular tumors (such as seminomas), pinealomas, lung cancers (particularly large cell cancers), gastrointestinal cancers (colon, pancreas, esophagus), breast cancers, melanomas, and hepatoblastomas. Recently, histologic analysis of a spectrum of primary lung neoplasms showed that those of a more differentiated neuroendocrine cell type (eg, small cell lung cancer and carcinoid tumors) were more likely to produce the hCG-α subunit. In contrast, hCG-β subunit expression in trophoblastic and nontrophoblastic tumors typically correlated with a less well differentiated tumor cell phenotype.
Depending on the age and gender of the patient, the clinical signs and symptoms may vary. Importantly not all patients manifest signs of their hCG excess. In susceptible individuals, however, clinical findings may be clearly due to the elevated hCG. Children, for example, with malignant hepatoblastoma can present with precocious puberty. Women may have dysfunctional uterine bleeding. Men with high hCG levels may have signs of hypogonadism with impotence and gynecomastia. Because of the thyroid-stimulating effects of high hCG levels, these patients occasionally demonstrate hyperthyroidism.
Since the beta subunit is unique to each of the glycoprotein hormones, the best means of detecting excessive production of hCG is by measuring the beta hCG subunit by a highly specific radioimmunoassay or immunofluorometric assay. Since hCG can serve as an important tumor marker for initial diagnosis and for recurrence, this value can be followed as an indicator of tumor activity. In contrast to hCG, ectopic production of the gonadotropins FSH and LH is extremely rare.
Etiology & Clinical Features
Oncogenic osteomalacia is a syndrome seen in association with unusual mesenchymal tumors and rarely with prostate cancer. These patients have hypophosphatemia, renal phosphate wasting, and low serum levels of 1,25-dihydroxyvitamin D. Alkaline phosphatase activity, reflecting bone turnover, is often elevated. Levels of calcium and parathyroid hormone are typically normal. Hypophosphatemia in this syndrome is due to reduced renal phosphate reabsorption. The defect in phosphate reabsorption is due to proximal tubular dysfunction and may be accompanied by glucosuria and aminoaciduria. Clinical symptoms include bone pain, muscle weakness, fractures, back pain, waddling gait, and progressive debility. The syndrome often poses a significant diagnostic dilemma to clinicians since the tumors responsible for it may be very small, obscurely situated, and difficult to identify. Phosphate depletion and low 1,25-dihydroxyvitamin D levels lead to poor bone mineralization and osteomalacia. Typically, the diagnosis of osteomalacia is often not even suspected for years despite the presence of classic symptoms. The humoral basis for the syndrome is supported by the observation that the attendant biochemical abnormalities remit and the rickets heals when the responsible tumor is removed.
Pathology & Pathogenesis
Tumors responsible for this form of acquired osteomalacia are usually small and grow slowly. Because the histology of these tumors is unusual and their locations often unanticipated, this syndrome has been dubbed
“strange tumours in strange places” by Weiss and colleagues. The range of locations for these tumors includes the lower extremities (45%), head and neck (27%), and upper extremities (17%). In a review of head and neck tumors that cause oncogenic osteomalacia, Gonzalez-Compta and coworkers noted that in 57% and 20% of cases, respectively, tumors were in the sinonasal and mandibular areas. Because these tumors are often small and in obscure locations, careful physical examination combined with cranial, chest, and abdominal CT scanning are usually needed for diagnosis. In several instances, MRI skeletal surveys—and in some instances indium-111 pentreotide scanning—have been instrumental in localizing the tumors at skeletal sites.
Most tumors causing this syndrome are benign, but malignant lesions have also been reported. The histologic spectrum of tumors has included hemangiomas, hemangiopericytomas, angiosarcomas, chondrosarcomas, prostate cancer, schwannomas, neuroendocrine lesions, and mesenchymal tumors. Many of these tumors have been classified pathologically as mixed connective tissue tumors. They are often located in bone. Osteoclast-like giant cells and stromal cells as well as highly vascular features characterize these tumors. On microscopic analysis, these tumors do not commonly demonstrate neurosecretory granules. As evidence of the slow growth of these tumors, delays in diagnosis of as long as 19 years have been reported.
Most tumors responsible for this syndrome overproduce fibroblast growth factor (FGF) 23, a protein implicated in phosphate wasting. FGF-23 is mutated in cases of autosomal dominant hypophosphatemic rickets. These mutations appear to render the mutant protein less susceptible to proteolytic cleavage and inactivation. Whether FGF-23 is responsible for the changes in vitamin D hydroxylation in oncogenic osteomalacia and whether the osteomalacia is secondary to problems in vitamin D availability or phosphate wasting is unknown.
Investigators in the field have noted common features between tumor-induced osteomalacia and X-linked hypophosphatemic rickets. The latter condition is a dominant disorder characterized by rickets or osteomalacia, hypophosphatemia, and low 1,25-dihydroxyvitamin vitamin D levels. Despite these similarities, there are several unresolved differences between the two syndromes. One is that levels of 1,25-dihydroxyvitamin D are inappropriately normal in patients with X-linked hypophosphatemic rickets and frankly low in oncogenic osteomalacia. In addition, patients with X-linked hypophosphatemic rickets demonstrate osteosclerosis and enthesopathy (calcification of tendons and ligaments).
X-linked hypophosphatemic rickets is probably due to defective functioning or synthesis of the PEX gene product, or PHEX, a protein that is homologous to neutral, membrane-bound endopeptidases. PHEX is thought to activate or inactivate a circulating factor involved in phosphate metabolism, which was classically termed “phosphatonin” by investigators in this field. It has long been proposed that the normal function of phosphatonin was to block renal phosphate reabsorption. It has been shown that FGF-23, the product of tumors that cause oncogenic osteomalacia, inhibits phosphate uptake in kidney cells. PHEX, the endopeptidase product of the PEX gene, can degrade wild-type FGF-23 but not mutant FGF-23 such as is found in autosomal dominant hypophosphatemic rickets.
Vasoactive intestinal peptide (VIP) is a peptide hormone with a number of important functions in the regulation of neuronal activity, differentiation, and survival, particularly in the sympathetic nervous system. A variety of neuroendocrine tumors express and release VIP. Pancreatic islet cell tumors that make VIP are quite rare. They can cause voluminous secretory diarrhea, achlorhydria, and hypokalemia. Other tumors, including small cell carcinomas of the lung, carcinoids, pheochromocytoma, medullary carcinomas of the thyroid, and some colonic adenocarcinomas, also produce VIP which may or may not be responsible for related symptoms. VIP is also made in non-small cell lung carcinomas and may function as an autocrine regulator of cell growth based on in vitro studies.
Hypercalcemia of Malignancy
Finley RS: Bisphosphonates in the treatment of bone metastases. Semin Oncol 2002;29(Suppl 4):132.
Hofbauer LC et al: Receptor activation of nuclear factor-kappaB ligand and osteoprotegerin: potential implications for the pathogenesis and treatment of malignant bone diseases. Cancer 2001;92:460.
Levine PH et al: A study of adult T-cell leukemia/lymphoma incidence in central Brooklyn. Int J Cancer 1999;80:5662.
Strewler GJ: The parathyroid hormone-related protein. Endocrinol Metab Clin North Am 2000;29:529.
Syndrome of Inappropriate Secretion of ADH
Adrogue HJ, Madias NE: Hyponatremia. N Engl J Med 2000;342: 1581.
Ferlito A, Rinaldo A, Devaney KO: Syndrome of inappropriate antidiuretic hormone secretion associated with head neck cancers: review of the literature. Ann Otol Rhinol Laryngol 1997;106:878.
Hirshberg B, Ben-Yehuda A: The syndrome of inappropriate antidiuretic hormone secretion in the elderly. Am J Med 1997; 103:270.
Kamoi K et al: Osmoregulation of vasopressin secretion in patients with the syndrome of inappropriate antidiuresis associated with central nervous system disorders. Endocr J 1999;46:269.
Vanhees SL, Paridaens R, Vansteenkiste JF: Syndrome of inappropriate antidiuretic hormone associated with chemotherapy-induced tumour lysis in small-cell lung cancer: case report and literature review. Ann Oncol 2000;11:1061.
Kaltsas GA et al: A critical analysis of the value of simultaneous inferior petrosal sinus sampling in Cushing's disease and the occult ectopic adrenocorticotropin syndrome. J Clin Endocrinol Metab 1999;84:487.
Tabarin A et al: Usefulness of somatostatin receptor scintigraphy in patients with occult ectopic adrenocorticotropin syndrome. J Clin Endocrinol Metab 1999;84:1193.
Baxter RC: The role of insulin-like growth factors and their binding proteins in tumor hypoglycemia. Horm Res 1996;46: 195.
Furrer J et al: Carcinoid syndrome, acromegaly, and hypoglycemia due to an insulin-secreting neuroendocrine tumor of the liver. J Clin Endocrinol Metab 2001;86:2227.
Hizuka N et al: Serum high molecular weight form of insulin-like growth factor II from patients with non-islet cell tumor hypoglycemia is O-glycosylated. J Clin Endocrinol Metab 1998; 83:2875.
Hoekman K et al: Hypoglycaemia associated with the production of insulin-like growth factor II and insulin-like growth factor binding protein 6 by a hemangiopericytoma. Clin Endocrinol (Oxf) 1999;51:247.
Le Roith D: Tumor-induced hypoglycemia. N Engl J Med 1999; 341:757.
Mizuta Y et al: Acinar cell carcinoma of the pancreas associated with hypoglycemia: involvement of “big” insulin-like growth factor-II. J Gastroenterol 1998;33:761.
Seckl MJ et al: Hypoglycemia due to an insulin-secreting small-cell carcinoma of the cervix. N Engl J Med 1999;341:733.
Teale JD, Marks V: Glucocorticoid therapy suppresses abnormal secretion of big IGF-II by non-islet cell tumours inducing hypoglycaemia (NICTH). Clin Endocrinol (Oxf) 1998;49:491.
Dirnhofer S et al: Selective expression of trophoblastic hormones by lung carcinoma: neurendocrine tumors exclusively produce human chorionic gonadotropin alpha-subunit (hCG alpha). Hum Pathol 2000;31:966.
Lundin M et al: Tissue expression of human chorionic gonadotropin beta predicts outcome in colorectal cancer: a comparison with serum expression. Int J Cancer 2001;95:18.
Syrigos KN et al: Beta human chorionic gonadotropin concentrations in serum of patients with pancreatic adenocarcinoma. Gut 1998;42:88.
Vartiainen J et al: Preoperative serum concentration of hCGbeta as a prognostic factor in ovarian cancer. Int J Cancer 2001;95: 313.
GHRH and Growth Hormone
Beuschlein F et al: Acromegaly caused by secretion of growth hormone by a non-Hodgkin's lymphoma. N Engl J Med 2000; 342:1871.
Doga M et al: Ectopic secretion of growth hormone-releasing hormone (GHRH) in neuroendocrine tumors: relevant clinical aspects. Ann Oncol 2001;12(Suppl 2):S89.
Drange MR, Melmed S: Long acting lanreotide induces clinical and biochemical remission of acromegaly caused by disseminated growth hormone-releasing hormone-secreting carcinoid. J Clin Endocrinol Metab 1998;83:3104.
Matsuno A: Pituitary somatotroph adenoma producing growth hormone (GH)-releasing hormone (GHRH) with an elevated plasma GHRH concentration: a model case for autocrine and paracrine regulation of GH secretion by GHRH. J Clin Endocrinol Metab 1999;84:3241.
Othman NH et al: Growth hormone-releasing hormone (GHRH) and GHRH receptor (GHRH-R) isoform expression in ectopic acromegaly. Clin Endocrinol (Oxf) 2001;55:135.
Kelley MJ et al: Calcitonin elevation in small cell lung cancer without ectopic production. Am J Respir Crit Care Med 1994; 149:183.
Ghillani PP et al: Identification and measurement of calcitonin precursors in serum of patients with malignant diseases. Cancer Res 1989;49:6845.
Bowe AE et al: FGF-23 inhibits renal tubular phosphate transport and is a PHEX substrate. Biochem Biophys Res Commun 2001;284:977.
DiMeglio LA, Econs MJ: Hypophosphatemic rickets. Rev Endocr Metab Disord 2001;2:165.
Econs MJ: New insights into the pathogenesis of inherited phosphate wasting disorders. Bone 1999;25:131.
Fukumoto S et al: Diagnostic utility of magnetic resonance imaging skeletal survey in a patient with oncogenic osteomalcia. Bone 1999;25:375.
Gonzalez-Compta X et al: Oncogenic osteomalacia: case report and review of head and neck associated tumours. J Laryngol Otol 1998;112:389.
Jan de Beur SM et al: Localisation of mesenchymal tumors by somatostatin receptor imaging. Lancet 2002;359:761.
Rowe PSN: The role of the PHEX gene (PEX) in families with X-linked hypophosphataemic rickets. Curr Opin Nephrol Hypertens 1998;7:367.
Weiss D et al: Oncogenic osteomalacia: strange tumours in strange places. Postgrad Med J 1985;61:349.
White KE et al: Autosomal-dominant hypophosphatemic rickets (ADHR) mutations stabilize FGF-23. Kidney Int 2001;60:2079.
Hejna M et al: Serum levels of vasoactive intestinal peptide (VIP) in patients with adenocarcinomas of the gastrointestinal tract. Anticancer Res 2001;21:1183.
Smith SL et al: Vasoactive intestinal polypeptide secreting islet cell tumors: a 15-year experience and review of the literature. Surgery 1998;124:1050.