Info
Background:
The genetic heterogeneity inherent to brain tumors makes it challenging to treat brain cancer since distinct sub-populations of tumor cells within the same brain tumor microenvironment (BTME) can express divergent drug targets1, interact differentially with blood vessels2 or immune cells3, and invade the brain via unique mechanisms4. Nonetheless, our understanding of how genetic heterogeneity drives tumor cell invasion in vivo remains limited due to our inability to track the long-term fate, signaling, and role of these cells in brain tumor formation. For example, two-photon (2P) microscopy studies using in vitro assays of cell migration5 might not recapitulate tumor cell behavior in vivo, while the application of large, benchtop-based 2P systems for imaging preclinical mouse models is restricted to acquiring brief snapshots under anesthesia6 that in turn could under-sample tumor cell behavior and incompletely characterize BTME remodeling. “Miniscopes”7 or miniaturized head-mounted microscopes for preclinical imaging, offer a promising solution for continuous anesthesia-free in vivo brain tumor imaging. However, their use in assessing BTME heterogeneity is limited by a lack of multicolor-imaging8, i.e., an inability to simultaneously probe multiple tumor cell genotypes tagged with fluorescent (FL) reporters in conjunction with other critical in vivo BTME variables such as vascular function9.

Hypothesis/Objective:
To develop a color-sensing miniscope capable of simultaneous in vivo dual-FL and vascular imaging within the BTME.

Methods:
As shown by the schematics in Fig. 1a-b, we combined micro-optical elements within a 3D-printed casing. Illumination for imaging green FL protein (GFP) expression was provided via an excitation module (Ex-Mod #1), consisting of a blue light-emitting diode (LED) and an excitation filter (470±20 nm). The broad excitation spectrum of red FL proteins (RFP) allowed the same blue light source to be used for simultaneous imaging of RFP expression. A second Ex-Mod provided RFP-only excitation via a lime-colored LED. A dual-band emission filter and lens (f=4.6 mm) were used for image formation on an RGB-color CMOS image sensor (OV5647). We used a 530±20 nm LED and a 680 nm laser as light sources for simultaneous imaging of intrinsic optical signals10 and laser speckle contrast11 for visualizing the BTME vasculature and its blood flow, respectively. We validated the miniscope’s performance by conducting in vitro imaging of GFP and RFP microbeads and in vivo imaging of a mouse brain inoculated with murine (GL261) and human (HSR-GBM1B) glioma cells engineered to express GFP and RFP, respectively.

Results:
Fig. 1c shows the color-sensing miniscope (~3.5 g). Fig. 1d-g shows the result of imaging green and red FL microbeads (diam=10 µm) with it, where we were able to probe weak and strong FL signals within the same field of view via short (250-600 ms, Fig. 1d-e) and long (2000 ms, Fig. 1f-g) exposure image acquisitions. Fig. 1h shows a dual-color in vivo image from a 3×3 mm2 field of view in the murine cortex seeded with GL261-RFP and HSR-GBM1B-GFP brain tumor cells. Fig. 1i shows the RFP-only image acquired under a separate lime-colored excitation (i.e., Ex-Mod #2), which served as a negative control for distinguishing GFP-expressing cells. Fig. 1j-l shows an example of imaging the BTME vasculature (Fig. 1k) and perfusion (Fig. 1l) concurrently in the presence of two distinct brain tumor cell populations (Fig. 1j). Finally, Fig. 1m-n shows a representative example that illustrates our ability to track the movement of individual brain tumor cells with a distinct genetic background, e.g., GL261 cell C1 (GFP expression) vs. HSR-GBM1B cell C2 (RFP expression), over 16 hours in the brain of a freely behaving animal.

Conclusion:
Color-sensing miniscopes can track genetically heterogeneous brain tumor cells in vivo. Such tools have the potential to transform our understanding of genetic heterogeneity in the BTME.

Image/Figure:
Click to view full size
Click to view full size

Image/Figure Caption:
Fig. 1: Color-sensing miniscopes enable in vivo characterization of the genetic heterogeneity of brain tumors. (a-b) Schematics showing the color-sensing miniscope design from the front (i.e., a) and rear (i.e., b). An excitation module (Ex- Mod #1, marked in b), comprising of a blue (470±20 nm) excitation source and excitation filter (i.e., Ex Filter #1), provides illumination for imaging of green and red fluorescence (FL). An LED cap (marked in a) blocks light leakage from the Ex-Mod to the image sensor and improves the signal-to-noise ratio (SNR) of acquired FL images. Another Ex-Mod containing a lime (570±20 nm) excitation source was used in some prototypes to provide RFP-only excitation. Two more light sources, a 530±20 nm LED and a laser diode (wavelength=670 or 680 nm), provide illumination for acquiring images of cerebral blood volume and blood flow. Illumination sources were turned on and off one after the other, and images for each contrast mechanism were acquired sequentially. Light scattered back from the object plane (e.g., the mouse cortex), passes through an aperture (diameter = 2.5 mm, not visible in the schematics), a fixed-focus lens (f=4.6 mm), and a dual-bandpass emission filter before hitting the CMOS image sensor. The CMOS image sensor features red-green-blue optical filters deposited in a Bayer pattern, which allows it to record color images (i.e., images with red, green, and blue channels). (c) Picture of a prototype of the color-sensing miniscope. The prototype weighs ~3.5 grams. (d and f) Images of an aqueous preparation of green and red FL microbeads (diameter=10 µm) acquired via the miniscope with short (250-600 ms, d) and long (2s, f) exposure times. (e and g) Corresponding images of red FL beads acquired with lime excitation. Images of GL261-GFP and HSR-GBM1B-RFP cells acquired in vivo following superficial inoculation of those cells in a male C57BL/6 mouse under blue (h) and lime (i) light excitation. In vivo images acquired from the color-sensing miniscope that show the dispersion of GFP- and RFP-tagged brain tumor cells (j) within an in vivo BTME characterized by a well-perfused (l) microvasculature (k). (m-n) Color-sensing allowed us to continuously track (e.g.,>16 hours) the migration or lack thereof of individual brain tumor cells with distinct genetic backgrounds, e.g., mouse (GL261, cell C1) and human (HSR-GBM1B, cell C2) tumor cells, within the same BTME in the brain of a freely behaving mouse. H=high, L=low.

Author
Janaka Senarathna
Johns Hopkins University School of Medicine