Serum CD26 (Dipeptidyl Peptidase IV, DPP IV) compared with Immunoglobulin Heavy-Chain Mutation Status, ZAP-70 and CD38 as a Predictor of Time to First Treatment in Early B-CELL Chronic Lymphocytic Leukemia

We analyzed the correlation between well-established biological parameters of prognostic relevance in B-chronic lymphocytic leukemia [CLL] (i.e, mutational status of the immunoglobulin heavy chain variable region [IgV_H], ZAP-70- and CD38-expression) and serum levels of CD26 (dipeptidyl peptidase IV, DPP IV) by evaluating the impact of these variables on the time to first treatment [TFT] in a series of 69 previously untreated Binet stage A B-CLL patients. By using a commercial ELISA (Human DPPIV/CD26 Quantikine R D Systems, USA) we found that circulating levels of CD26 did not reflect Rai sub-stages (P=0.52), β2-microglobulin (P=0.68), LDH (P=0.06), mutational status of IgV_H (P=0.28), ZAP-70 (P=0.52) and CD38 (P=0.48). We used an optimal cut-point search to determine how to best split soluble CD26 data. Maximally selected logrank statistics plots identified a CD26 serum concentration of 371 ng/mL as the best cut-off. Accordingly, patients who had CD26 levels higher than 371 ng/mL experienced a shorter TFT (median 40 months) in comparison to patients whose CD26 levels were lower than 371 ng/mL (median 120 months; P=0.003). The univariate Cox proportional hazard model identified higher serum concentration of CD26 and absence of mutation in IgV_H as predictor of shorter TFT. Again in multivariate analysis all these parameters maintained their discriminating power: mutational status of IgV_H (HR= 0.23; CI 95% 0.10−0.51, P<0.0001), soluble CD26 (HR= 0.38; CI 95% 0.17−0.85, P=0.02). The effects of CD26 on TFT were masked by mutational status of IgV_H in patients with unmutated IgV_H. In contrast, serum levels of CD26 and mutational status of IgV_H had a joint effect on TFT in patients with mutated IgV_H which translates into a segregation of patients with mutated IgV_H in two sub-groups with different TFT according to CD26 levels (HR= 0.18; CI 95% 0.05−0.60, P=0.005). Our results indicate that in early B-cell CLL biological profile including among other parameters soluble CD26 may provide a useful insight into the complex interrelationship of prognostic variables. Furthermore, CD26 along with mutational status of IgV_H can be adequately used to predict clinical behaviour of patients with low biological risk.