Abstract
The cure rate in pediatric acute lymphoblastic leukemia (ALL) is about 85%. Therapy failures are mainly due to relapse. Relapse rates could possibly be influenced by the anti-leukemic activity of the immune system, e.g. by the Th1/Th2 balance. It would be desirable to increase the cure rate if the immune system could be induced to participate in the elimination of leukemia cells. However, leukemia cells, especially ALL cells are poor stimulators for T cells and do not induce a Th1 response believed to be necessary for tumor cell elimination. Missing costimulatory molecules on the leukemia cells, especially CD80 and CD86, are commonly accepted as the main reason for their poor immunostimulatory activity. On the other hand, efficient costimulation does not guarantee improved cure rates. On the contrary, in a mouse model inoculation of CD86 transfected lymphoma cells led to tumor progression possibly by interaction with CD152 or induction of a Th2 response. Indeed, in an earlier study, we examined the mRNA expression of CD80 and CD86 in marrow samples of ALL patients and found an increased expression of CD86 and IL-4 in patients with a late leukemic relapse raising the possibility of a Th2 shift predisposing to relapse (Stachel et al, Eur J Med Res, 2006). It becomes increasingly clear that members of the costimulatory family other than CD80 and CD86 also influence the immune response. ICOS/ICOS ligand and PD-1/PD1 ligand both play an important role in the second step of Th2 induction. To determine whether increased or decreased expression of costimulatory molecules play a role in the pathogenesis of relapse in pediatric ALL we examined the expression of various costimulatory molecules in leukemic marrow samples in a prospective study. Samples from 49 consecutive pediatric patients with B cell precursor acute lymphoblastic leukemia (BCP ALL) were analyzed by semiquantitative RT-PCR for CD28, CD152 (CTLA-4), ICOS and ICOS ligand (B7RP-1).
A total of 29 patients (60.4 %) remained relapse free and 19 (39.6 %) relapsed after a median follow up of 24.5 months (range: 6 to 50 months). Of those relapsing five patients suffered a very early relapse (VER, within 18 months from diagnosis), seven an early relapse (ER, between 18 and 30 months from diagnosis) and seven patients a late relapse (LR, later than 30 months from diagnosis). We found that in the subgroup of ALL patients experiencing a VER expression of mRNA ICOS ligand was significantly increased (2.0 +/− 0.7 (mean +/− SD relative intensity units) compared to non-relapsing patients (1.1 +/− 0.8, p=0.02). This trend was also visible in the late relapsing patient group (1.4 +/− 0.8 vs 0.9 +/− 0.6, ns). By comparative densitometry we estimated the copy number of ICOS ligand in the VER group to be 500,000 per 12,500 BM cells vs 350,000 per 12,500 BM cells in non-relapsing ALL patients.
Since ICOS ligation on T-cells by ICOS ligand on B-cells appears to favor a Th2 polarization of the immune system our results could provide further evidence that a Th2 shift induced by leukemic blasts could confer an increased risk for very early ALL relapse. These data could provide a rationale for immunotherapeutic strategies aimed at strengthening a Th1 polarized immune response in the treatment for childhood BCP ALL.
Nucleotide excision repair is a predictor of early relapse in pediatric acute lymphoblastic leukemia