Cells are quite valuable, especially when used for regenerative research, diagnostics or research. But harvested cells do not come presorted and need to be separated from a heterogeneous mixture of cells. There are already numerous methods to sort cells according to biophysical properties such as size, density, morphology, and dielectric or magnetic susceptibility. Cell sorting based on labels can have a higher specificity, but introduces extra steps to add and remove labels, which can affect the phenotype of the cell. Rohit Karnik of MIT has developed a cell sorting method based on cell rolling. The continuous, label-free process is described in “Cell sorting by deterministic rolling” in Lab on a Chip.
Cell rolling is a phenomenon where a cell is constantly forming and releasing adhesive bonds with a surface under fluid flow. The continuous creation and release of bonds by the cell induces rolling and is an integral role in the movement of lymphocytes, platelets, stem cells and metastatic cancer cells. To induce cell rolling, a surface needs to be coated with a ligand specific to the target cell type. The rolling target cells need to be focused, so slanted ridges are added to the bottom of the channel. When the target cell comes into contact with the surface of the coated ridges, it will begin to roll along it and eventually turn the corner into the space between ridges. By following the path of the direction of the ridges, the targeted cells will be focused on one side of the channel known as the gutter. The non-target cells should not adhere and roll along the ridges, allowing them to be spatially differentiated from the target cells. But the ridges actually serve an additional purpose: Acting as mixers, these ridges introduce circulation to the axial flow. This flow would normally be laminar, which would prevent the majority of cells from coming in contact with the surface of the channel and rolling.
Karnik validated this cell sorting method by processing the leukemia cell lines HL60 and K562. The surfaces were coated with P-selectin, which is a known ligand for the target HL60. HL60 and K562 were injected in a single inlet at a ratio of 2:3, respectively. Outlet A held 95.0 ± 2.8% HL60 cells, and outlet B had 94.3 ± 0.9% K562 cells. The cell sorting was extremely successful and 87.2 ± 3.7% and 76.7 ± 14.2% of the HL60 and K562 cells were recovered at the end of the process. Cell loss was most likely due to settling in the syringe at the inlet and cells remaining in the channel and dead volumes at the end of the process. Karnik also investigated the effect of the ligand concentration on cell sorting. Higher concentrations generated stronger cell-surface adhesion, but this came at the expense of cell rolling, so an operating point had to be determined for an ideal cell rolling concentration; at a flow rate of 70 µL/min, the channel was incubated with P-selectin concentration of 1.5 µg/mL.
I really like this method of cell sorting because it is both passive and label-free. Although an extra section of channel must be added with coated ridges, no other major components are necessary. This method does not need any more equipment or chambers, making it simple to integrate into a project. With its small footprint, it can also be highly parallelized, negating the need for it to operate at high flow rates which could hinder cell rolling. This could either function as a standalone sorting device or integrated into a device processing a mixture of cells. Similar to other cell sorting procedures, widespread usage of this particular method is limited to availability of information: only cell lines for which we’ve characterized the rolling behavior for can be sorted this way.
Choi, S., Karp, J., & Karnik, R. (2012). Cell sorting by deterministic cell rolling Lab on a Chip, 12 (8) DOI: 10.1039/c2lc21225k