Hypoglycemia due to Paraneoplastic Secretion of Insulin-Like Growth Factor-I

Abstract

Nonislet cell tumor hypoglycemia (NICTH) resulting from the hypersecretion of IGF-II by a number of different types of mesenchymal tumors has been reported by numerous authors (1). This issue of JCEM contains the first report of a woman with recurrent hypoglycemia resulting from the hypersecretion of IGF-I by an undifferentiated lung carcinoma (2). RIA of this patient’s serum demonstrated only a modest increase in total IGF-I but a 10-fold increase in free IGF-I. Measurements of free and total IGF-II were normal. After filtration through a sizing gel, the IGF-I present in the patient’s serum eluted in the position similar to the IGF-I present in normal serum. Hypoglycemic symptoms were controlled first by treatment with corticosteroids and GH and subsequently by chemotherapy of her tumor. The uniform molecular size of IGF-I observed in this patient’s serum differs remarkably with the heterogeneity of IGF-II-related peptides in the sera of patients with NICTH with large mesenchymal tumors (3). Most of these peptides in the serum of these patients retain the first 21 amino acids of the E domain of pro-IGF-II. Normally, this E-21 of pro-IGF-II is glycosylated (4). This may be a targeting signal for subsequent cleavage of the 21-amino-acid E domain peptide. Therefore, deficient glycosylation of IGF-II E may contribute to the size heterogeneity observed in IGF-II associated with NICTH. We found that big IGF-II of serum of four patients with NICTH were not retained on a Jaclin column, suggesting defective glycosylation (5). We suggested that this failure of glycosylation may have contributed to defective processing of pro-IGF-II. A comparable defect in processing of pro-IGF-I does not appear to occur with tumor-related pro-IGF-I. The E-21 fragment derived from pro-IGF-II appears to be cleared from the plasma very slowly. Much is excreted in the urine and is retained in the serum in uremia. The proper test for a physician to order when faced with a hypoglycemic patient with a tumor and normal concentration of serum insulin is not total serum IGF-I or -II but the free IGF peptide. In the serum of this patient, serum IGF-I was measured by the method of Frystyk et al. (6), which is the gold standard with which others are compared. It involves ultracentrifugal sedimentation of the IGF binding protein complexes under carefully controlled conditions of temperature with subsequent sampling of the supernatant. It is possible to obtain comparable separation of free IGF from bound IGF of serum by precipitation in neutral 80% ethanol and centrifugation in a simpler desktop high-speed centrifuge (7). Fortunately for physicians, several commercial laboratories are now able to measure free IGF-I and free IGF-II in serum. Interpretation of results so obtained requires knowledge of the method used and results in normal sera. With these resources available, it is reasonable to predict that additional cases of tumor-related hypoglycemia due to IGF-I will be recognized. nonislet cell tumor hypoglycemia.