Vascular endothelial growth factor (VEGF), the most important angiogenic factor in tumor biology, contributes to acute myeloid leukemia (AML) growth via multiple paracrine interactions. Expression of VEGF receptors (VEGFR) by some AML cells also allows VEGF to function as a direct cell-autonomous autocrine growth factor promoting survival and chemoresistance. Although VEGF isoforms and VEGFR have been previously been identified in AML patient samples, very few studies have examined the prognostic impact of concomitant VEGF/VEGFR expression in this disease or in subsets of AML by prognostic karyotype. We examined expression of two VEGF ligands (VEGF-A, -C) and three VEGFR (VEGFR-1, R-2, R-3) in diagnostic marrow samples from 91 consecutive AML pts (46% male, 54% female) with normal karyotype (NK) treated at our institute from 1995–2005. Median patient age was 66 years (range 21–87). Less than half (n=46, 49%) achieved complete remission (CR) following induction chemotherapy. Median overall survival (OS) was 9.6 months. Levels of VEGF-A, -C, VEGFR-1, R-2, and R-3 mRNA were measured by Q-PCR and expressed relative to normal bone marrow controls (level of mRNA expression=1). VEGF-A and -C protein expression was also evaluated by immunohistochemistry (IHC). We found that the majority of AML samples overexpressed VEGF-A (median 1.35 fold increase, range 0.06–12.18) and underexpressed VEGF-C (median 0.49, range 0.03–27.93) by Q-PCR. IHC confirmed that most marrow AML cells were positive for VEGF-A (n=69/84, 82%) but were not positive for VEGF-C (n=56/86, 65%) protein. VEGFR mRNA expression was found in a minority of NK-AML samples but was highly concordant with co-expression of all three receptors (VEGFR1+R2+R3) identified in one third (27/84, 32%) of all patients. Significant prognostic factors for OS were age, CR, VEGFR-1 (p=0.0146) and VEGFR-2 (p=0.0055). Significant factors for EFS were the same: age, CR, VEGFR-1 (p=0.0146), and VEGFR-2 (p=0.0055). Predictors for leukemia-free survival (LFS, defined as time from CR to relapse/death) were VEGFR-1 (p=0.0289) and elevated WBC at diagnosis (p=0.0265). VEGF-A, and VEGFR-3 expression did not predict for outcome in any analyses. Multivariate analysis confirmed that age and CR were independent prognostic factors in all NK-AML patients; however, the impact of individual VEGFR status on outcome was unable to be determined due to the high concordant sample expression, and concordant expression was not significant.

Conclusions: We conclude that the majority of NK-AML marrow samples expressed higher levels of VEGF-A than normal bone marrow. One third of NK-AML samples also expressed one or more VEGFR with a high degree of concordant expression. Both VEGFR-1 and VEGFR-2 overexpression were prognostic factors for OS and EFS, while VEGFR-1 overexpression predicted for leukemia-free survival. These data demonstrating the prognostic significance of VEGFR-1 and VEGFR-2 expression in normal karyotype AML (constituting 40–49% of new AML diagnoses) suggest that the presence of autocrine VEGF-A/VEGFR-1 and VEGF-A/VEGFR-2 mediated loops in these abnormal myeloblasts may confer therapy resistance and/or an increased risk of relapse. Accordingly, targeted therapy directed against multiple VEGFR with novel receptor tyrosine kinase inhibitors, alone or in combination with chemotherapy, could be considered in order to improve clinical outcomes in NK-AML. Potential correlations between VEGFR expression and FLT-3/NPM-1 gene mutations in NK-AML patients are under evaluation.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author