(702a) Characterizing Interstitial Fluid Flow and the Effects of Shear Stress in the Brain Tumor Microenvironment

CHARACTERIZING INTERSTITIAL FLUID FLOW AND THE
EFFECTS OF SHEAR STRESS IN THE BRAIN TUMOR MICROENVIRONMENT

Cornelison RC¹*,
Kingsmore KM², Brennan CE², Tom S¹, Munson JM^1,
2

¹Virginia Tech, Blacksburg, VA 24060; ²University
of Virginia, Charlottesville, VA, 22908

*presenting author; rcorneli@vt.edu

Co-author contacts: kmk2yt@virginia.edu, ceb7rb@virginia.edu, stom18@vt.edu, jm4kt@vt.edu

Introduction

Glioblastoma
(GBM) is the most invasive and deadly brain tumor. During tumorigenesis,
heightened interstitial fluid pressure within the tumor drives fluid flow at
tumor-parenchyma interfaces, directly enhancing glioma cell invasion [1]. It is
known in other tissues that interstitial flow and increased shear stress also activate
stromal cells, which further exacerbates cancer cell invasion and worsens disease
[2]. We hypothesized that interstitial fluid flow and shear stress stimulate
astrocytes and microglia in the brain tumor microenvironment (TME) to enhance
glioma cell invasion. We first sought to quantitatively characterize interstitial
fluid flow at tumor-parenchyma interfaces in vivo and then assess the
effects of interstitial flow and shear stress on glial cell activation in
vitro in relation to cancer cell invasion.   

Methods

Four different patient-derived glioma stem
cell (GSC) lines (G2, G34, G62, G528) were individually implanted in
immune-compromised mice to generate in vivo xenograft models of GBM. Approximately
10-11 days post-inoculation, gadolinium contrast agent was injected
intravenously, and dynamic contrast-enhanced magnetic resonance imaging (MRI)
was used to visualize changes in contrast intensity over thirteen minutes [3].
The resulting series of images were then processed using a MATLAB model
developed in our lab that solves for a number of solute and fluid transport
parameters. After imaging, the tracer dye Evans blue was injected intravenously
to assist in endpoint flow region analysis, and the following day the mice were
sacrificed via intracardial perfusion. The brain tissue was then prepared for
cryosectioning and immunohistochemical analysis.

In
vitro,
patient-derived GSCs were combined with human astrocytes and microglia within a
hyaluronan matrix to generate a high-throughput (96-well) tissue culture insert
model of the TME [4]. This model enabled examination of interstitial flow via
application of a pressure head on top of the hydrogel. Invasion of three GSC
lines (G34, G528, and G2) was quantified ± TME ± flow by counting the number of
cells/mm² on the underside of the porous tissue culture insert after
overnight culture. Additionally, 2D cultures of either astrocytes, microglia,
or both were subjected to low shear (~0.1 dyne/cm²) to generate
serum-free conditioned media. These media were then used to assess induction of
GSC chemotaxis in a standard Matrigel chemotaxis assay. Ibidi chambers were
used to subject glia to physiological (0.01 dynes/cm²) or
pathological (0.1, 0.5, 1 dynes/cm²) shear rates compared to static
conditions, and expression of sphingosine-1-phosphate receptor 3 (S1P₃)
was examined as a known marker of activation [5].

Results

The speed of interstitial flow was extremely
heterogeneous within tumors as well as at tumor-parenchyma interfaces, ranging
from ~0.1-1.5 µm/s. The mean speed was similar across the four tumor models and
did not correlate with tumor size or blood vessel density. Surprisingly, vector
direction – opposed to magnitude – at the tumor border was more indicative of
high flow regions, as evidenced by comparison to Evans blue labeled tissue (Fig.
1A_i-ii and 1B_i-ii). Regions where the overall vector
direction was relatively perpendicular to the tumor border correlated with
higher extravasation of Evans blue, while regions of acute-angled vectors showed
lower dye extravasation. We then used immunohistochemistry to label these
tissues for cells of the TME such as astrocytes (GFAP⁺) and
microglia (Iba1⁺). We observed a higher presence of these cell types
in the tumor in regions of high flow compared to low flow (Fig. 1A_iii
and 1B_iii), leading us to examine the effects of interstitial
flow and shear stress on glial cells in vitro.

We
found that the addition of astrocytes and microglia to 3D, static cultures of patient-derived
GSC lines increased invasion of two out of three lines (G34 and G2), with the
TME decreasing G528 invasion (p<0.01, n=5). The addition of interstitial
flow exacerbated these effects, further increasing G34 and G2 invasion and
decreasing G528 invasion (Fig. 1C). We next examined if the effects of the
TME under shear were mediated via soluble factors using static/sheared glial
conditioned media. GSC migration only increased toward conditioned media from glial
co-cultures, but not individual cultures, under shear, revealing that all three
elements were required to induce chemotaxis (Fig. 1D). To examine the
impacts of shear stress on glial activation, we seeded astrocytes and microglia
in Ibidi µSlide chambers and applied physiological and pathological fluid shear.
One hour of shear was sufficient to increase expression of S1P₃, a
known marker of glial activation. Cells exposed to physiological shear (0.01
dyne/cm²) showed marker expression roughly equivalent to cells
cultured in static conditions, but cells exposed to pathological rates of shear
(0.1, 0.5, and 1 dyne/cm²) exhibited increased S1P₃
expression. Cells at 0.1 dynes/cm² exhibited the highest upregulation.
Together, these data indicate that shear stress can independently induce glial cell
activation, which in turn stimulates glioma cell chemotaxis through release of
soluble factors.

Conclusions

We have developed a method to identify peri-tumoral
regions of high interstitial fluid flow, a known stimulator of glioma cell
invasion, in live animals using non-invasive brain imaging techniques. Vector
maps of interstitial flow velocity were generated by processing of dynamic
contrast-enhanced MR images, revealing that vector directionality at the tumor
border correlated well with identifiable regions of interstitial flow in fixed
tissue samples. Additionally, using multiple in vitro models, we showed
that interstitial flow and shear stress can activated astrocytes and microglia,
in turn promoting glioma cell chemotaxis. Hence, we have ultimately identified
a novel intersection between a mechanical force (interstitial fluid flow) and a
chemical force (paracrine signaling) in the human brain tumor microenvironment
which enhances cancer cell invasion.

References

[1]
Kingsmore, KM, DK Logsdon, DH Floyd, SM Peirce,
BW Purow, and JM Munson (2016) “Interstitial flow differentially increases
patient-derived glioblastoma stem cell invasion via CXCR4, CXCL12, and
CD44-mediated mechanisms.” Integrative Biology, 8(12), 1246-1260.

[2]
Shieh, AC, HA Rozansky, B Hinz, and MA Swartz (2011) “Tumor cell invasion is
promoted by interstitial flow-induced matrix priming by stromal fibroblasts.” Cancer Research, 71:790-800.

[3]
Kingsmore et al., (2018). submitted.

[4]
Harris, AR, JX Yuan, and JM Munson (2017) “Assessing multiparametric drug
response in tissue engineered tumor microenvironment models.” Methods, 134-135:20-31.

[5]
Fischer, I, C Alliod, N Martinier, J Newcombe, C Brana, and S Pouly (2011) “Sphingosine
kinase 1 and sphingosine 1-phosphate receptor 3 are functionally upregulated on
astrocytes under pro-inflammatory conditions.” PloS one, 6(8), e23905.