Tumor recovery by angiogenic switch from sprouting to intussusceptive angiogenesis after treatment with PTK787/ZK222584 or ionizing radiation

Abstract

Inhibitors of angiogenesis and radiation induce compensatory changes in the tumor vasculature both during and after treatment cessation. To assess the responses to irradiation and vascular endothelial growth factor-receptor tyrosine kinase inhibition (by the vascular endothelial growth factor tyrosine kinase inhibitor PTK787/ZK222854), mammary carcinoma allografts were investigated by vascular casting; electron, light, and confocal microscopy; and immunoblotting. Irradiation and anti-angiogenic therapy had similar effects on the tumor vasculature. Both treatments reduced tumor vascularization, particularly in the tumor medulla. After cessation of therapy, the tumor vasculature expanded predominantly by intussusception with a plexus composed of enlarged sinusoidal-like vessels containing multiple transluminal tissue pillars. Tumor revascularization originated from preserved alpha-smooth muscle actin-positive vessels in the tumor cortex. Quantification revealed that recovery was characterized by an angiogenic switch from sprouting to intussusception. Up-regulated alpha-smooth muscle actin-expression during recovery reflected the recruitment of alpha-smooth muscle actin-positive cells for intussusception as part of the angio-adaptive mechanism. Tumor recovery was associated with a dramatic decrease (by 30% to 40%) in the intratumoral microvascular density, probably as a result of intussusceptive pruning and, surprisingly, with only a minimal reduction of the total microvascular (exchange) area. Therefore, the vascular supply to the tumor was not severely compromised, as demonstrated by hypoxia-inducible factor-1alpha expression. Both irradiation and anti-angiogenic therapy cause a switch from sprouting to intussusceptive angiogenesis, representing an escape mechanism and accounting for the development of resistance, as well as rapid recovery, after cessation of therapy.

Figure 1

The graph (A) and bar graph (B) showing tumor growth delay and central necrosis. A, B: PTK/ZK and radiation delayed tumor growth (A) and induced central necrosis (estimated as a ratio “volume tumor necrosis/entire tumor volume”) (B). Mean values (n ≥3) are represented together with the SD. *Value is significantly different from the control (*P ≤ 0.02). Tumor growth was inhibited by both treatment strategies. C, C′: Vascular casts of MMTV/c-neu mammary carcinoma xenografts, on day 5 (C) and day 9 (C′) after onset of treatment with PTK/ZK. Large avascular areas (asterisks) are apparent within the medullary region. Multiple tiny capillary sprouts emanate from the preserved cortical vessels and invade the avascular medulla (arrowheads). D–F: Transmission electron micrographs illustrating changes in the vessels. In (D), a vessel in a non-treated tumor with intact endothelium is shown. E, F: Attenuation with partial denudation (see inserts) and vacuolization (asterisks) of endothelial cells in the tumor vessels after treatment with PTK/ZK (E) or irradiation (F). The magnification is ×3600; the scale bar is presented in (D).

Figure 2

Consecutive sections through MMTV/c-neu mammary carcinoma xenografts showing immunostaining for CD31 (A, B, C) and SMA (A′, B′, C′). Multiple vessels within the cortex (arrowheads in B) bear a covering of SMA-positive cells (arrow in B′), whereas those within the medulla (arrowheads in C) are mainly SMA-negative (positive is marked with arrow in C′). The spatial distribution of SMA accords with the distribution revealed by immunoblotting (D). Scale bars = 200 μm (A, A′) and 50 μm (B, B′, C, C′) presented in the right lower corner of (A′) and (C′) accordingly.

Figure 3

Proliferative activity of tumor cells on day 5 visualized by BrdU immunostaining revealed treatment-specific spatial tumor recovery (A). The outer “ring” of the tumor with the breadth of up to 1 mm was characterized by the maximal proliferating activity. The counting of BrdU-positive cells on days 5 and 14 indicated a specific time-course of tumor recovery (B). On day 14 the overall proliferative rate of tumor cells increased and the difference between cortex and medulla diminished (C). The values significantly different from control values are marked with *(P ≤ 0.02) or ^#(P ≤ 0.05).

Figure 4

A–F: Vascular casts of MMTV/c-neu mammary carcinoma xenografts on day 14 revealing a change in the mode of vascular growth from sprouting (in control animals) to intussusception (after treatment with PTK/ZK or irradiation). The revascularization process involves mainly intussusceptive angiogenesis, as evidenced by the presence of numerous tiny holes in the casts (arrows). In the control tumors, vascular growth occurs predominantly by sprouting (arrowheads). G, H: Bar graphs showing the number of newly-formed pillars (with a diameter of less than 2.5 μm) and sprouts per vascular surface on days 9 (G) and 19 (H). The values significantly different from control values are marked with *(P ≤ 0.02) or ^#(P ≤ 0.05).

Figure 5

Micrographs and schema illustrating time-course of the angiogenic switch during tumor recovery. The scanning electron micrographs representing the vascular pattern of the PTK/ZK treated tumors on day 6 (A), 14 (A′), and 19 (A″). On day 6 and 14 pillars (arrowheads) and meshes (asterisks) dominate. In contrast to days 6 and 14, on day 19 (A″) the sprouts (arrows) prevail in the vascular pattern of the PTK/ZK-treated tumors with some pillars (arrowhead) or meshes (asterisks) present. The bar in (A–A″) is 100 μm. The scheme in B represents the above-described changes: in non-treated tumors the dominating mode of angiogenesis is sprouting. After the short-term therapy the intussusception prevails: so called, intussusceptive phase of recovery (days 6, 9, and 14). On day 19 the second wave of intussusception is there what is characterized by numerous sprouts along with pillars or meshes.

Figure 6

A–C: Immunohistochemical staining of MMTV/c-neu mammary carcinoma xenografts for CD31 revealing a multitude of tiny vessels in control tumors and fewer, but larger sinusoidal ones in the treated groups. D–F: Changes in the IMD; number of vessels per ×200 field of view) and vessel area density on days 5, 9, and 19 reflect a switch in the mode of angiogenesis: intussusception on day 9 and a second wave of sprouting on day 19. The values significantly different from control values are marked with *(P ≤ 0.02) or ^#(P ≤ 0.05). The HIF-1α staining (G–I) reveals the differences in oxygenation level of the tumor cells on day 14: in non-treated tumors (G) there is apparent hypoxia demonstrated by plentiful HIF-1α-positive cells on the distance of more than 100 to 130 μm from the vessels (arrowheads in G–I). There are necrotic areas on similar distances from the vessels (asterisks in G). On day 14 in both treated groups the level of oxygenation is apparently better with only some HIF-1α-positive cells (H and I). The vessels are marked with arrowheads (G–I). The magnification is ×200.

Figure 7

Spatial distribution of SMA on day 9. A, B: Laser scanning microscopy (A) of double-immunostaining [CD31-red (endothelium); SMA-green (SMA-positive cells)] and light microscopic inspection of toluidine-blue-stained consecutive sections (B–B‴) revealed that SMA-positive cells are located in the vicinity of pillars (arrows in A and B) and also as peri-endothelial covering (arrowheads in A and B). The double arrows in A and B indicate transluminar tissue pillars. C: Estimation of SMA by immunoblotting revealed higher levels in the medullary than in the cortical tumor regions. (C = cortex, M = medulla).