Cardiovascular and systemic microvascular effects of anti-vascular endothelial growth factor therapy for cancer

Abstract

Objectives: This study sought to evaluate the contribution of microvascular functional rarefaction and changes in vascular mechanical properties to the development of hypertension and secondary ventricular remodeling that occurs with anti-vascular endothelial growth factor (VEGF) therapy.

Background: Hypertension is a common side effect of VEGF inhibitors used in cancer medicine.

Methods: Mice were treated for 5 weeks with an anti-murine VEGF-A monoclonal antibody, antibody plus ramipril, or sham treatment. Microvascular blood flow (MBF) and blood volume (MBV) were quantified by contrast-enhanced ultrasound in skeletal muscle, left ventricle (LV), and kidney. Echocardiography and invasive hemodynamics were used to assess ventricular function, dimensions and vascular mechanical properties.

Results: Ambulatory blood pressure increased gradually over the first 3 weeks of anti-VEGF therapy. Compared with controls, anti-VEGF-treated mice had similar aortic elastic modulus and histological appearance, but a marked increase in arterial elastance, indicating increased afterload, and elevated plasma angiotensin II. Increased afterload in treated mice led to concentric LV remodeling and reduced stroke volume without impaired LV contractility determined by LV peak change in pressure over time (dp/dt) and the end-systolic dimension-pressure relation. Anti-VEGF therapy did not alter MBF or MBV in skeletal muscle, myocardium, or kidney; but did produce cortical mesangial glomerulosclerosis. Ramipril therapy almost entirely prevented the adverse hemodynamic effects, increased afterload, and LV remodeling in anti-VEGF-treated mice.

Conclusions: Neither reduced functional microvascular density nor major alterations in arterial mechanical properties are primary causes of hypertension during anti-VEGF therapy. Inhibition of VEGF leads to an afterload mismatch state, increased angiotensin II, and LV remodeling, which are all ameliorated by angiotensin-converting enzyme inhibition.

Figure 1. Ascending Thoracic Aorta Histology

Mean (±SEM) (A) thickness of the tunica media, (B) number of elastic lamina, and (C) elastin layer thickness for control and G6-31−treated mice after 5 weeks of therapy. (D) Elastic lamina fragmentation scores, expressed as a proportion for each score. p = NS by Mann-Whitney and chi-square analysis. (E) Elastin staining of the ascending aorta illustrating normal (top) and mild fragmentation (bottom, arrow) of the elastic laminae. Scale bar = 20 µm. Four measurements were made for 8 mice in each group.

Figure 2. Mean (±SEM) MBV in the Myocardium and Skeletal Muscle

Volumes were measured at baseline (Wk 0) and after 5 weeks of therapy (Wk 5). Microvascular blood volume (MBV) for skeletal muscle represents values calculated for the capillary compartment.

Figure 3. Skeletal Muscle MBF

(A) Mean (±SEM) microvascular blood flow (MBF) at rest. (B) Mean (±SEM) MBF during contractile exercise measured after 5 weeks of therapy. (C) Examples of contrast-enhanced ultrasound images from the proximal hindlimb obtained after the high-power destructive pulse sequence and corresponding time-intensity data at rest and during contractile exercise. BG = background image.

Figure 4. Renal Function and Histology

(A) Mean (±SEM) urinary albumin excretion at weeks 2 to 5 of therapy. (B) Serum creatinine at the end of 5 weeks of therapy. p < 0.05 versus controls. (C) Histopathology scores for glomerular disease; p < 0.001 by Kruskal-Wallis test for treatment-related differences. *p < 0.05 versus controls for intergroup comparisons; n = 6 for all treatment groups.

Figure 5. Examples of Histopathology From the Outer Renal Cortex From Control and G6-31−Treated Mice

(A, C, E, and G) Control mice and (B, D, F, and H) G6-31−treated mice. Examples include periodic acid–Schiff (PAS) staining (A to D), endothelial staining with CD31 (E and F), and fibrin staining (G and H representing sections that were directly adjacent to corresponding CD31 staining sections). The PAS staining examples show glomeruli that were graded as normal (A and C) and severe mesangial change (B and D). Arrows (G) illustrate the location of nonstaining glomeruli that are clearly visible by CD31 staining (E). Examples of PAS staining for G6-31 mice and fibrin/CD31 overlay images with confocal microscopy are provided in the Online Appendix.