Calpain 2 is required for the invasion of glioblastoma cells in the zebrafish brain microenvironment

Abstract

Glioblastoma is an aggressive primary brain tumor with a 5-year survival rate of less than 5%. The ability of glioblastoma cells to invade surrounding brain tissue presents the primary challenge for the success of focal therapeutic approaches. We previously reported that the calcium-activated protease calpain 2 is critical for glioblastoma cell invasion in vitro. Here, we show that expression of calpain 2 is required for the dispersal of glioblastoma cells in a living brain microenvironment. Knockdown of calpain 2 resulted in a 2.9-fold decrease in the invasion of human glioblastoma cells in zebrafish brain. Control cells diffusely migrated up to 450 μm from the site of injection, whereas knockdown cells remained confined in clusters. The invasion study was repeated in organotypic mouse brain tissues, and calpain 2 knockdown cells demonstrated a 2.3-fold lower area of dispersal compared with control cells. In zebrafish brain, glioblastoma cells appeared to migrate in part along the blood vessels of the host. Furthermore, angiogenesis was detected in 27% of zebrafish injected with control cells, whereas only 12.5% of fish receiving knockdown cells showed the formation of new vessels, suggesting a role for calpain 2 in tumor cell angiogenesis. Consistent with the progression of glioblastoma in humans, transplanted tumor cells were not observed to metastasize outside the brain of zebrafish. This study demonstrates that calpain 2 expression is required for the dispersal of glioblastoma cells within the dynamic microenvironment of the brain, identifying zebrafish as a valuable orthotopic system for studying glioblastoma cell invasion.

Fig. 1

Calpain 2 expression is required for invasion of human glioblastoma cells in the zebrafish brain. A: Real-time analysis of human glioblastoma cell invasion in vitro. Control and calpain 2 knockdown glioblastoma cells were added to the Matrigel matrix at 0.8 (top), 0.4 (middle), and 0.2 (bottom) mg/ml concentrations in the upper chamber of transwells and invasion stimulated with ±10% FBS in the lower chamber. The impedance of microelectrodes on the lower surface of membrane was measured every 15 min for 23 hr and represented as cell index. C, control; K, knockdown; S, 10% FBS. The curves are representative of two independent experiments. B: Control and calpain 2 knockdown (KD) glioblastoma cells labeled with CMDiI were microinjected in the brain of 4-day-old zebrafish. Monochrome images of cells in the brain of zebrafish captured 1 and 6 days postinjection (dpi). The insets show the initial position of the transplanted cells in the brain at 1 dpi. C: The bar graph represents the average area occupied by control and KD cells in the brain quantified in Metamorph 6.2; n = 25; mean ± SEM. D: Box plot showing percentage increase in the area occupied by cells after 6 dpi between control and KD groups (*P < 0.001). E: Monochrome images of CMDiI stained cells maintained under standard cell culture conditions; dps, days poststaining.

Fig. 2

Human glioblastoma cells disperse along blood vessels in the zebrafish brain. Ten days postfertilization, Tg(fli1:egfp) zebrafish, expressing GFP in endothelial cells, were microinjected with CMDiI-labeled human glioblastoma cells (red) and imaged by confocal microscopy. Invasion was assessed in three dimensions by capturing z-stacks at a thickness of 5.6 μm. Ortho analysis of z-stacks was performed to monitor the distribution of migrating cells along blood vessels (green). The image in the X-Z plane (below) shows relative distribution of tumor cells along the primordial midbrain channel, whereas that in the Y-Z plane (right) shows cells spreading along the anterior cerebral vein. The yellow regions identify close association of cells with the blood vessels. Images are representative of 30 injected fish. Blood vessel nomenclature: PMBC, primordial midbrain channel; ACeV, anterior cerebral vein; MtA, metencephalic artery; CtA, central artery.

Fig. 3

Calpain 2 knockdown (KD) glioblastoma cells remain confined in clusters and demonstrate decreased localization with blood vessels. A: Human glioblastoma cells transplanted in the brain of 10-day-old Tg(fli1:egfp) zebrafish were imaged for 6 dpi by confocal microscopy. The yellow regions represent close association of cells (red) with blood vessels (green). Blood vessel nomenclature: PMBC, primordial midbrain channel; ACeV, anterior cerebral vein; CtA, central artery. B: Colocalization of tumor cells with blood vessels was quantified using Zeiss Physiological Software v3.2; n = 6 (control) and 5 (knockdown); mean ± SEM.

Fig. 4

Knockdown of calpain 2 attenuates tumor cell angiogenesis induced by injected glioblastoma cells. Tg(fli1:egfp) zebrafish, transplanted with control and calpain 2 knockdown glioblastoma cells, were imaged for 6 dpi by confocal microscopy. Three-dimensional reconstruction of tumor cells invading the brain of zebrafish at 1 and 6 dpi shows blood vessels in green and tumor cells in red. The newly formed blood vessels at 6 dpi are shown with arrows; n = 11 (control) and 8 (knockdown).

Fig. 5

Human glioblastoma cells do not invade in the yolk sac of zebrafish. A: CMDiI-labeled control and calpain 2 knockdown cells (KD) were microinjected into the yolk sac of 4-day-old zebrafish. Animals were imaged for 5 days using a Zeiss axiovert fluorescence microscope. Cells did not show significant dispersal and remained confined in clusters in the yolk sac. B: Area occupied by cells at 2 and 5 dpi was quantified using Metamorph 6.2; n = 5; mean ± SEM.

Fig. 6

Knockdown of calpain 2 results in decreased invasion of glioblastoma cells in organotypic mouse brain slices. A: Human glioblastoma cells (10,000), stained with CMDiI, were added to the brain tissue obtained from 62-day-old mice and imaged for 7 days by confocal microscopy. B: Area of dispersal was quantified using Zeiss Physiology Software v3.2; n = 3; mean ± SEM.