Clinical and biologic correlates of insulin binding by leukemia lymphoblasts

The percentage of total 125I-labeled insulin specifically bound to lymphoblasts was measured in 46 children with leukemia. Among 35 children with newly diagnosed acute lymphoblastic leukemia (ALL), specific insulin binding ranged from 0.09 to 14.8% per 10(6) blasts. A lower level of insulin binding was correlated with T-cell surface markers (P less than 0.003), higher hemoglobin level (P less than 0.005), presence of a mediastinal mass (P less than 0.01), lower glucocorticoid receptor level (P less than 0.02), higher platelet count (P less than 0.04), age less than 2 or greater than 10 yr (P less than 0.05), white blood cell count greater than or equal to 100 X 10(3)/mm3 (P less than 0.06) and higher labeling index (P less than 0.07). It was not correlated with the presence of central-nervous-system disease, FAB classification, or sex. With a follow-up of 24 to 33 + months, insulin binding was not correlated with treatment outcome. Six patients with relapsed ALL and three with acute nonlymphoblastic leukemia showed insulin binding levels similar to those in newly diagnosed ALL patients. Blasts from one patient with B-cell ALL and one with chronic myelogenous leukemia were characterized by lower insulin binding, while lymphoblasts from a patient with T-cell lymphoma bound insulin at marginally detectable levels. In vitro studies with IM-9, NALM-1 and NALM-16 cell lines showed that changes in insulin binding caused by dexamethasone treatment were not correlated with hormone-induced cell death. Although study of insulin binding by malignant lymphoid cells may be important in understanding the biology of leukemic cells, it does not appear to have any obvious clinical utility.