Prognostic Significance of CXCR4 Overexpression in Pediatric Acute Myeloid Leukemia with Low-Risk: A Report from the Japanese Pediatric Leukemia/Lymphoma Study Group

Background:

The outcome of acute myeloid leukemia (AML) has been improved by advanced risk-stratified chemotherapy based on prognostic factors. However, the long-term survival rate remains 60-70% and further investigation is needed. Core binding factor (CBF)-AML, characterized by t(8;21) or inv(16)/t(16;16), is the most frequent subtype in pediatric AML. Although CBF-AML is generally classified as a low-risk (LR) group, approximately 30% of the patients relapsed within 3 years in the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG) AML-05 trial. This finding suggests that the LR patients had other risk factors that could account for their adverse outcome. CXC chemokine receptor 4 (CXCR4) is a G protein-coupled receptor that helps leukemia cells adhere to bone marrow stromal cells that affect the cell survival and proliferation. There are several reports that CXCR4 overexpression is an adverse prognostic factor in adult AML; however, few studies have examined their significance in pediatric AML. Here, we retrospectively examined whether CXCR4 overexpression can be an adverse prognostic factor in pediatric AML with LR.

Patients and Methods:

From 2006 to 2010, 485 de novo pediatric AML patients aged <18 years were registered in the AML-05 trial. Overall, 42 patients were excluded, mainly because of misdiagnosis. The risk groups were defined as follows: all of the CBF-AML patients with complete remission (CR) after the initial induction course without FLT3-ITD were included in the LR group, while patients with -7, 5q-, t(16;21)(p11;q22), Ph1, or non- CR after the initial induction course or with FLT3-ITD were included in the high-risk (HR) group. Patients with no LR or HR features were in the intermediate-risk. CXCR4 mRNA expression was analyzed by quantitative RT-PCR using diagnostic bone marrow samples. CXCR4 expression levels were dichotomized based on the average expression level.

Results and Discussion:

A total of 248 samples were available for this study. No significant differences in the clinical outcomes were observed between these 248 patients and the 195 patients who did not have CXCR4 data (EFS, P=0.55; OS, P=0.87). CXCR4 overexpression (CXCR4+) was present in 81 patients (32.7%). CXCR4+ was not correlated with age, sex, and WBC count. CBF-AML was similarly distributed between CXCR4+ and CXCR4- patients. The frequency of FLT3-ITD in CXCR4+ patients was significantly higher in CXCR4- patients (P=0.049), while there were no significant differences in that of KIT and CEBPA mutation (P=0.08 and 0.71, respectively). In total (n=248), no significant differences were observed between CXCR4+ and CXCR4- patients in 3-year OS (68.6% vs. 75.1%, P=0.42) and 3-year EFS (65.6% vs. 58.9%, P=0.27). However, in the LR group (n=93), a significant difference was observed between CXCR4+ and CXCR4- patients in 3-year OS (79.2% vs. 98.3%, P=0.008), whereas no significant difference was observed in 3-year EFS (83.3% vs. 69.6%, P=0.33). We have revealed that KIT mutations are adverse prognostic factors in t(8;21) pediatric AML in our recent study (Leukemia, 2015). In t(8;21) AML without KIT mutations (n=44), a significant difference was observed between CXCR4+ and CXCR4- patients in 3-year OS (76.1% vs. 100.0%, P=0.01), whereas no significant difference was observed in 3-year EFS (77.8% vs. 74.2%, P=0.85). Multivariate Cox regression analysis, including CXCR4+, KIT mutation, and age identified CXCR4+ as the only prognostic factor predicting poor OS in the cohort of AML with LR (hazard ratio, 11.01; P=0.01), but not EFS (hazard ratio, 0.54; P=0.21). The results suggest that CXCR4+ is associated with an adverse prognosis in pediatric AML with LR. The adverse outcome was restricted in OS, not in EFS, suggesting that CXCR4+ may be associated with the poor prognosis after recurrence, although the underlying mechanism remains unknown. Furthermore, it is noteworthy that the OS of LR patients with CXCR4- was quite high, which should be confirmed in larger studies.

Conclusions:

We identified the adverse prognostic significance of CXCR4 overexpression in pediatric AML with LR. The adverse prognosis may be restricted after recurrence. Further studies of the underlying biology are required. There are several promising CXCR4 antagonists including AMD3100, therefore, our findings will help to develop better therapeutic approaches in pediatric AML with LR.