Serum levels of
vascular endothelial growth factor (
VEGF-S) have been reported to correlate with
tumor stage and prognosis in various human
malignancies. The source of soluble
VEGF in peripheral blood remains obscure. We therefore measured the concentration of immunoreactive
VEGF in 241 serum samples and 61 plasma samples (
VEGF-P) from 20 subjects undergoing myeloablative
chemotherapy and from 3 normal platelet donors. A significant correlation between the peripheral blood platelet count (PC) and
VEGF-S (r = 0.86) but not
VEGF-P was found.
VEGF-S levels were 58.43 +/- 42.50 pg/ml (mean +/- SD) in patients with a PC < 50 x 10(9)/l, 203.29 +/- 176.56 pg/ml for a PC of 50-150 x 10(9)/l, and 457.42 +/- 475.41 pg/ml for a PC > 150 x 10(9)/l. Interestingly,
VEGF-P levels were substantially lower than the corresponding
VEGF-S values, namely below the detection limit in most cases. Supernatants from platelet-rich plasma contained no
VEGF, but after in vitro lysis of the platelets very high
VEGF levels were found. The
VEGF content per 10(9) platelets was calculated at 2.51 +/- 2.39 pg and was dependent on the mean platelet volume. In summary,
VEGF release from platelets during blood clotting was found to be the main source of
VEGF in serum samples.
Cancer patients in clinical remission have negligible amounts of soluble
VEGF in peripheral blood, and myeloablative
chemotherapy causes a significant drop in
VEGF-S levels corresponding to the decrease in PC. Thus, studies addressing the diagnostic and prognostic value of
VEGF-S in
cancer patients must be interpreted with caution. Our data provide the basis for predicting
VEGF-S in relation to PC in vivo, and for reevaluating former studies of
VEGF-S in patients with malignant or nonmalignant disease.