Lenalidomide Induced Cytokine Release Syndrome in Chronic Lymphocytic Leukemia (CLL): Clinical and Laboratory Correlates of Immune Activation.

Lenalidomide is active in lymphoid malignancies, but its mechanism of action remains ill defined. One possible mechanism is immune activation due to increased expression of costimulatory molecules on tumor cells. In CLL lenalidomide treatment has been uniquely complicated by tumor flare reactions (TFR: pain and lymph node swelling) resulting in treatment related mortality. To investigate effects of lenalidomide on CLL cells we exposed PBMC from 17 CLL patients enrolled in a phase II clinical trial of single agent lenalidomide and normal donors (n=10) in-vitro to 2μM lenalidomide for 48 hours and measured costimulatory molecules CD80 and CD86 on B-cells and activation marker CD69 on T-cells by flow cytometry. CD80 expression increased on average 2-fold on CLL cells but remained unchanged on normal B-cells (p=0.01 for log2 MFI CLL vs normal). CD69 expression on T-cells followed a similar pattern, albeit with more interindividual variability among CLL samples (p=0.03 for log2 MFI CLL vs. normal). Next we wished to correlate the degree of in-vitro activation with the clinical effects of lenalidomide treatment in the same patient. Our observations in several patients suggested that the dominant feature of lenalidomide treatment is a cytokine release syndrome (CRS). Indeed, on day 8 of treatment we detected increased serum levels of TNFa, IL-1ra, CCL2, CCL3, CCL4 and IL-8. To correlate the CRS with in-vitro measurements, we applied uniform criteria for diagnosis. Patients who experienced at least 2 of the following symptoms (% of patients with the symptom, n=18) were considered to have a CRS: increase in lymph node size and or ALC by >25% (50%), fever >38C (44%), pain (61%), fatigue (72%), chills (33%), hypotension/dehydration (39%), and rise of creatinine (33%). Onset of symptoms was within 8 to 72 hours (average 38 hours) after initiation of therapy (20mg patients 1–10, 10mg patients 11–18). The CRS score, summarizing number and severity of symptoms in each patient, averaged 3.14 (range 0–10) with no difference between the 20mg and10mg cohort. The CRS score correlated (Pearson r-value, p<0.01) with: lenalidomide induced increase in CD80 on CLL cells (r=0.67), and increase of CD69 on T-cells (r=0.74), rise in CRP on day 4 (r=0.63), and inversely with treatment induced changes in T-cell numbers (r=−0.7). No correlation was found between the CRS score and the decrease in leukemic cells (mean decrease 40% day 8, 54% day 21). It has been suggested that during the TFR, T-cell numbers in lymphoid organs increase. We therefore scored 11 matched lymph node biopsies taken from the same patient pre-treatment and on day 8 of the first cycle for T-cell content. Only 2 patients had an increase of >50% in CD3+ cells in the lymph node and the average pre/post ratio of T-cells was 1.14 (p=0.37). In summary, lenalidomide upregulated expression of CD80 on B-cells and of CD69 on T-cells from CLL patients but not on normal B and T-cells. The in-vitro response correlated with the clinical onset and severity of a CRS in-vivo. However, the degree of T-cell activation or of the CRS did not predict the effect on the leukemic cell count. Our data suggest the possibility that immune activation in CLL may be primarily responsible for side effects but not be required for disease control. This interpretation is consistent with the good clinical activity of lenalidomide in several lymphoid malignancies in the absence of notable immune effects and if confirmed, has important implications for future use of lenalidomide and the formulation of combination regimens.