Background: This study explores the ability of the combined detection of soluble IL-2 receptor (sIL-2r) and interleukin-10 (IL-10) to predict treatment failure in patients with aggressive non-Hodgkin's lymphoma (NHL) and to evaluate the modifications in cytokine measurements induced by the therapeutic administration of recombinant human granulocyte-macrophage colony-stimulating factor (GM-CSF).
Methods: Serum levels of sIL-2r and IL-10 were measured serially in 93 patients with newly diagnosed aggressive NHL treated with four courses of a multiagent chemotherapy regimen. GM-CSF was administered subcutaneously in 39 of these patients from day +5 to day +18 after each chemotherapy course.
Results: Pretreatment levels of sIL-2r were greatly elevated in patients with NHL compared with control subjects (P < 0.001), significantly correlating with the Ann Arbor stage (P < 0.001) and beta 2-microglobulin (beta 2-m) concentrations (r = 0.552, P = 0.004). IL-10 was detected in 37 patients at diagnosis, with no correlation with clinicohematologic parameters, and was not detected in the control sample (P < 0.001). Cytokine and receptor levels progressively declined to normal ranges in responding patients, whereas they remained elevated in nonresponders. During administration of GM-CSF, the authors observed an increase of sIL-2r, whereas lower elevations were recorded for IL-10. However, on completion of the induction treatment, cytokine/receptor levels were comparable in patients with the same type of response, whether or not they had received GM-CSF. In the five patients who were investigated at relapse, the levels of sIL-2r, beta 2-m, and lactic dehydrogenase were found to be elevated. IL-10 concentrations were high in three of these patients: two already had detectable levels at presentation, whereas one tested positive only on recurrence. No single parameter was associated with response to therapy, but the combination of elevated IL-10 and sIL-2r concentrations greater than 3000 U/ml resulted in a subset of eight patients who failed induction chemotherapy (P < 0.001). In addition, six of eight patients with high IL-10 and beta 2-m concentrations greater than 3.3 mg/l had an unfavorable outcome (P = 0.003). A multivariate regression model was used to identify sIL-2r (P = 0.004) and beta 2-m (P = 0.043) as the covariates that amplified the prognostic ability of IL-10.
Conclusions: sIL-2r and IL-10 measurements provide valuable information for better management of patients with NHL as markers to monitor disease activity and as prognostic indicators.