Glioma expansion in collagen I matrices: analyzing collagen concentration-dependent growth and motility patterns

Abstract

We study the growth and invasion of glioblastoma multiforme (GBM) in three-dimensional collagen I matrices of varying collagen concentration. Phase-contrast microscopy studies of the entire GBM system show that invasiveness at early times is limited by available collagen fibers. At early times, high collagen concentration correlates with more effective invasion. Conversely, high collagen concentration correlates with inhibition in the growth of the central portion of GBM, the multicellular tumor spheroid. Analysis of confocal reflectance images of the collagen matrices quantifies how the collagen matrices differ as a function of concentration. Studying invasion on the length scale of individual invading cells with a combination of confocal and coherent anti-Stokes Raman scattering microscopy reveals that the invasive GBM cells rely heavily on cell-matrix interactions during invasion and remodeling.

FIGURE 1

Microscopy setup. Collimated light enters the inverted scope and is directed to the objective by a beamsplitter (NT 80/20) for confocal reflectance, or a shortpass (SP) dichroic mirror (at 650 nm) for CARS. Light reflected from the sample passes through the beamsplitter, and is deflected by a longpass (LP) dichroic mirror (at 488 nm) through a pinhole (PH), bandpass filter (BP), and to a photomultiplier tube (PMT). CARS light is collected by the condenser in the transmitted direction and directed through a bandpass filter to a second PMT. A halogen lamp is used for brightfield images. Most transmitted light from the brightfield microscopy passes through the SP 650 and the LP 488 and is collected on a third PMT.

FIGURE 2

(a) Brightfield image of cells in a collagen I matrix. (b) CARS image of cells in the same collagen matrix.

FIGURE 3

(a) Invasive distance of GBM in collagen I gels (0.5 mg/ml (○), 1.0 mg/ml (□), 1.5 mg/ml (▵), and 2.0 mg/ml (▹◃)). (b) MTS growth over 94 h. (c) GBM 4 h after implantation in 0.5 mg/ml gel. (d) GBM 4 h after implantation in 2.0-mg/ml gel and definitions of invasive distance and MTS radius. The MTS radius is defined by the extent of the dense cells in the center of the GBM system. The invasive distance is defined as the distance between the periphery of the MTS and a circle that circumscribes the invasive cells. (e) GBM 24 h after implantation in 0.5 mg/ml gel. (f) GBM 24 h after implantation in 2.0-mg/ml gel. The contrast in c and d differs from that in e and f in that the former were taken with a 10× phase-contrast objective, whereas the latter were taken with a different (5×) phase-contrast objective.

FIGURE 4

Confocal reflectance images of a quadrant of the MTS and surrounding collagen fibers 3–4 h after implantation. (a) MTS in a 2.0-mg/ml gel. (b) MTS in a 0.5-mg/ml gel. (c) Projection of confocal reflectance images of MTS in a 1.5-mg/ml collagen matrix several hours after implantation. Note the buildup of collagen around the edge of the MTS due to the growth of the MTS into the matrix.

FIGURE 5

Confocal reflectance images of four collagen matrices: (a) 0.5 mg/ml; (b) 1.0 mg/ml; (c) 2.0 mg/ml; and (d) 2.5 mg/ml. All images are 153.6 × 153.6 μm. Panel b contains both the confocal reflectance image and red lines overlying the collagen fibers identified by the analysis procedure employed. Some pixels in the center of the image are removed to eliminate speckle from the image.

FIGURE 6

(a) Mesh-size distribution for 0.5 mg/ml (•), 1.0 mg/ml (▪), and 2.0 mg/ml (▴) bare collagen gels with exponential fits. (Inset) Mesh size (•, left axis) and elastic modulus (▪, right axis) as a function of concentration. (b) Row density distribution (see text for details) for 0.5 mg/ml (•), 1.0 mg/ml (▪), and 2.0 mg/ml (▴) bare collagen gels. (c) Histogram of lengths of fibers identified via the procedure illustrated in Fig. 5 b for the 0.5-mg/ml (•), 1.0-mg/ml (▪), and 2.0-mg/ml (▴) bare collagen gels and fits to exponentials. (Inset) Characteristic fiber length versus concentration. (d) Angular distribution from −60° to 60° for the 0.5-mg/ml (•), 1.0-mg/ml (▪), and 2.0 mg/ml (▴) bare collagen gels.

FIGURE 7

CARS images at the periphery of the MTS taken 20 min apart. All images are 97.2 μm (x) × 95.7 μm (y). t* is the time elapsed from the beginning of the observation time, ∼3 h after implantation. Arrows point to two particular cells at t* = 0, 100, and 160 min to show that paths initially filled with thin invasive cells are later filled with thicker proliferative cells.

FIGURE 8

Confocal reflection images of collagen fibers subtracted from images of the same location at a later time. All images are 179.4 × 179.4 μm. The difference images have negative (black) and positive (white) components, and thus the black portions of the images correlate with the early time fiber arrangement whereas the white portions of the images correlate with the fiber arrangement at the latter time point. (a) Time zero (defined by start of the observation time) and time 40 min. Dotted line indicates the location of the cell. (b) Minute 40 − minute 72. (c) Minute 72 − minute 146.

FIGURE 9

(a and b) Confocal reflectance images of remodeled collagen matrices superimposed with simultaneously collected CARS images of invasive cells attached to the MTS. (c) Confocal reflectance image of remodeled collagen matrix and cell detached from the MTS.

FIGURE 10

(a) Row (•) and column (▪) density distributions for the 1.5-mg/ml remodeled matrix shown in Fig. 9 b. (b) Angular distribution from −60° to 60° for the 1.5-mg/ml remodeled matrix shown in Fig. 9 b. (c) Row- (•) and column (▪)-density distributions for the 1.5-mg/ml matrix shown in Fig. 9 c. (d) Angular distribution from −60° to 60° for the 1.5-mg/ml remodeled matrix shown in Fig. 9 c.