Myelosuppression of thrombocytes and monocytes is associated with a lack of synergy between chemotherapy and anti-VEGF treatment

Abstract

Purpose: Chemotherapeutic agents that have shown improved patient outcome when combined with anti-vascular endothelial growth factor (VEGF) therapy were recently identified to induce the mobilization of proangiogenic Tie-2-expressing monocytes (TEMs) and endothelial progenitor cells (EPCs) by platelet release of stromal cell-derived factor 1α (SDF-1α). VEGF blockade was found to counteract cell mobilization. We aimed to determine why agents like gemcitabine do not elicit TEM and EPC recruitment and may therefore lack synergy with anti-VEGF therapy.

Experimental design: Locally advanced pancreatic cancer patients (n = 20) were monitored during 16 weeks of neoadjuvant therapy. Treatment was based on gemcitabine with or without the addition of bevacizumab. Blood levels of proangiogenic cell populations and angiogenesis factors were determined in 2-week intervals.

Results: The lack of EPC mobilization during gemcitabine therapy was associated with severe thrombocytopenia and reduced SDF-1α blood concentrations. Furthermore, myelosuppression by gemcitabine correlated significantly with loss of TEMs. With respect to angiogenic factors stored and released by platelets, plasma levels of the angiogenesis inhibitor thrombospondin 1 (TSP-1) were selectively decreased and correlated significantly with thrombocytopenia in response to gemcitabine therapy.

Conclusions: A thorough literature screen identified thrombocytopenia as a common feature of chemotherapeutic agents that lack synergy with anti-VEGF treatment. Our results on gemcitabine therapy indicate that myelosuppression (in particular, with respect to thrombocytes and monocytes) interferes with the mobilization of proangiogenic cell types targeted by bevacizumab and may further counteract antiangiogenic therapy by substantially reducing the angiogenesis inhibitor TSP-1.

Figure 1

Blood levels of VEGF and CA 19-9 in pancreatic cancer patients undergoing neoadjuvant therapy. (A) Schematic representation of study design and blood sampling time points 1 to 8: The onset of bevacizumab administration is indicated by dashed and solid arrows for treatment groups 1 and 2, respectively. Plasma concentrations of VEGF (B, C) and CA 19-9 (D, E) according to blood collection schedule are illustrated by box plot for all patients n = 20 (B, D) or for the individual treatment groups n₁ = 9 and n₂ = 11 (C, E).

Figure 2

Effect of neoadjuvant treatment on EPC counts, blood platelets, and SDF-1α plasma levels of pancreatic cancer patients. Fluctuations of EPCs (CD34⁺ CD133⁺ blood cells) are presented by box plot for the entire study collective n = 15 (A) or separately for treatment arms n₁ = 6 and n₂ = 9 (B). C, D, and E illustrate the time course of SDF-1α (n = 7), blood platelets (n = 19), and SDF-1α levels normalized to a platelet count of 300 G/L (n = 7) irrespective of treatment group.

Figure 3

Changes in leukocyte, monocyte, and TEM counts during pancreatic cancer therapy. Total populations of blood leukocytes (A, n = 17), CD14⁺ monocytes (B, n = 7), and the subset of CD14⁺ CD16⁺ Tie2⁺ TEMs (C, n = 7) were monitored in patient blood by flow cytometry. Data are illustrated by box plot irrespective of treatment group.

Figure 4

Effect of chemotherapy on circulating proangiogenic and antiangiogenic factors. The plasma levels of antiangiogenic TSP-1 (A, n = 20), proangiogenic PD-ECGF (B, n = 20), and bFGF (C, n = 6) were determined by ELISA and are presented by box plot in relation to blood sampling time points, irrespective of treatment group.

Figure 5

Concurrent parameter fluctuations in accordance with chemotherapy schedule. Median values of platelets, TSP-1 (multiplied by 5), SDF-1α (divided by 7), and TEMs (divided by 20) as established throughout neoadjuvant cancer therapy are depicted in relation to gemcitabine administration. Therapy breaks (preceding blood withdrawal) are indicated by arrowheads.