Quantifying the Brain Metastatic Tumor Micro-Environment using an Organ-On-A Chip 3D Model, Machine Learning, and Confocal Tomography

Confocal tomography was developed to obtain the full gamut of cancer cell behaviors, as they interact with the cellular barrier and niche. This method provides a basis for studying metastases as they occur. This approach is useful for quantifying live cell behavior.

Confocal tomography, when combined with machine learning, is useful for a variety of organ-on-a-chip platforms. Analyzing primary or recurrent tumor or circulating tumor cells can constitute an important diagnostic tool to predict brain metastases and to discern brain metastatic from non brain metastatic cells. To assemble the microfluidic BBN device, transfer a device from the vacuum desiccator onto an appropriate surface with the inlets facing down and place the entire setup into a plasma chamber.

Place the entire setup into a plasma chamber and pull a vacuum before treating the device with plasma for 30 seconds at 80 Watts. At the end of the treatment, use a guide to quickly place the device, inlet side up onto the glass slide on the lab bench to create a permanent bond between the PDMs of the device and the slide. Next, insert 200 microliter pipette tips, cut two millimeters from the tip into all of the inlets and outlets and place the device into the plasma chamber for an eight minute, 200 watt plasma treatment.

After the device has cooled, place the device into a sterile secondary container and within 15 minutes of plasmid treatment, transfer 120 microliters of a collagen astrocyte solution into the device through the pipette tip for the bottom chamber. Allow the solution to wick across the chamber into the opposite pipette tip and fill the next channel of the device. When all four channels of the device have been filled, place the chip in the cell culture incubator for one hour.

When the collagen has set, bill all of the pipette tips feeding the bottom chamber with the appropriate complete cell culture medium and coat the upper chamber with 2%growth factor reduced matrigel, in complete endothelial medium. When both chambers have been coded, place the device into the incubator for one hour before rinsing the upper chamber with the appropriate medium. Alternating tips, see 30 microliters of one times 10 to the sixth endothelial cells per milliliter of appropriate medium, through the tips into the upper chamber every 15 minutes, for a total of four aliquots of cells per tip.

Incubate the device in between seedings. When all of the endothelial cells have been seeded, fill all of the tips with the medium and return the device to the cell culture incubator for 48 hours. When the endothelial layer has matured, replace the medium in the chip to replenish the cell culture medium and seed each top chamber channel with 30 microliters of one times 10 to the six cancer cells per milliliter of appropriate medium.

After each 30 microliter aliquot of cells has been seeded, return the device to the incubator for 15 minutes then fill all tips with the appropriate medium return the device to the cell culture incubator for 24 to 48 hours. After the device has been cultured and imaged, to measure the phenotype of the cancer cells, open the provided software and read the confocal image into memory. The software will save each color channel from the 3D confocal image into a separate TIFF file.

Select Change opacity values for each microscope channel'and adjust the channel alpha value for the color channels present in the image, such that the background is removed and only the fluorescents within the cells remains. To visualize the effect, select View a 3D rendering'to verify the threshold are set correctly and if the image looks correct, save the opacity values in the experiment tracker spreadsheet. Then, use marching cubes to convert the volume image into individual 3D triangular meshed objects saved in the VTK data format.

To fit a plane to the endothelial barrier, first locate the cell centroids. Use the polydata connectivity filter to iterate through the list of meshes in the VTK file to extract the regions that are not connected. Calculate the centroid of each mesh and add the measurement to a list of centroids filtering for meshes that are too large or too small.

To fit a plane to the list of centroids for the endothelial cells, use a minimization of error method, run Visualize RFP centroids and plane fit'to generate and plot the plane fit and inspect the fit of the plane, adjusting the fit manually as necessary. When the plane has been properly fitted, save the normal of the plane to the experiment tracker file. To measure each cancer cells phenotype after characterizing the endothelial layer with a plane, load the cell analysis function and run Read in the experiment tracker information and analyze the channels that exist"to iterate over and analyze each region in the VTK file.

For each region, the function will clip each cell so that the mesh below the membrane is calculated. To measure the cellular phenotype, calculate the shape, volume and position of each cancer cell. Then measure the volume and position of each clipped cancer cell to calculate the percentage of cell that extravasated through the endothelial barrier.

After performing these steps for several experiments run Export the data as a single xlsx file'to save the data in a spreadsheet. A microfluidic BBN chip with a confluent endothelial barrier is acceptable for experimentation, while a microfluidic BBN chip with especially poor endothelial coverage, is not. In this representative analysis the brain seeking cancer cell line, exhibits a subpopulation of cells that extravasates across the endothelial barrier and migrates deep into the brain niche space of the microfluidic BBN chip At two and nine days after exposure, the parental cancer cells maintained a substantial proportion of cells on top of the barrier away from the brain niche space, while the brain metastatic cancer cell population maintained a proportion of cells that were greater than 100%extravasated.

Morphological quantification of the cancer cell shape indicated that parental cells demonstrated fewer spherically shaped cells day one post-seeding, while after two and nine days of interaction both cancer cell lines trended toward decreasing their sphericity. In addition, the cancer cell subpopulations that extravasated into the astricydic niche were smaller in size, compared to the cancer cells that remained in interaction with the endothelial barrier without extravasating into the brain. Brain seeking cancer cell lines and patient derived xenografts, exhibit a phenotypic pattern in the microfluidic BBN chip that could be exploited to differentiate between the brain metastatic and non brain metastatic cancer cells by machine learning.

Self-selection is important. Endothelial cells with a higher, low passage number exhibited variable barrier behaviors. In addition, cancer cells and their fluorescent expression may change over time.

After this procedure, researchers may employ molecular profiling of the secret's home or cells in the device in order to study molecular mechanisms of metastases. This technique enables the exploration of complex interactions between cancer cells and their microenvironments, by combining an engineered microenvironment with a quantitative imaging of niche components over time.