Pallet arrays enable cells to be separated while they remain adherent to a surface and provide a much greater range of cell selection criteria relative to that of current technologies. However there remains a need to further broaden cell selection criteria to include dynamic intracellular signaling events. To demonstrate the feasibility of measuring cellular protein behavior on the arrays using high resolution microscopy, the surfaces of individual pallets were modified to minimize the impact of scattered light at the pallet edges. The surfaces of the three-dimensional pallets on an array were patterned with a coating such as fibronectin using a customized stamping tool. Micropatterns of varying shape and size were printed in designated regions on the pallets in single or multiple steps to demonstrate the reliability and precision of patterning molecules on the pallet surface. Use of a fibronectin matrix stamped at the center of each pallet permitted the localization of H1299 and mouse embryonic fibroblast (MEF) cells to the pallet centers and away from the edges. Compared to pallet arrays with fibronectin coating the entire top surface, arrays with a central fibronectin pattern increased the percentage of cells localized to the pallet center by 3-4-fold. Localization of cells to the pallet center also enabled the physical separation of cells from optical artifacts created by the rough pallet side walls. To demonstrate the measurement of dynamic intracellular signaling on the arrays, fluorescence measurements of high spatial resolution were performed using a RhoA GTPase biosensor. This biosensor utilized fluorescence resonance energy transfer (FRET) between cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) to measure localized RhoA activity in cellular ruffles at the cell periphery. These results demonstrated the ability to perform spatially resolved measurements of fluorescence-based sensors on the pallet arrays. Thus, the patterned pallet arrays should enable novel cell separations in which cell selection is based on complex cellular signaling properties.

A method to select and separate viable cells based on the results of a cell-lethal assay was developed. Cells were plated on an array of culture sites with each site composed of closely spaced, releasable micropallets. Clonal colonies spanning multiple micropallets on individual culture sites were established within 72 h of plating. Adjacent sites were widely spaced with 100% of the colonies remaining sequestered on a single culture site during expansion. A laser-based method mechanically released a micropallet underlying a colony to segment the colony into two genetically identical colonies. One portion of the segmented colony was collected with 90% efficiency while viability of both fractions was 100%. The segmented colonies released from the array were fixed and subjected to immunofluorescence staining of intracellular phospho-ERK kinase to identify colonies that were highly resistant or sensitive to phorbol ester-induced activation of ERK. These resistant and sensitive cells were then matched to the corresponding viable colonies on the array. Sensitive and resistant colonies on the array were released and cultured. When these cultured cells were reanalyzed for phorbol ester-induced ERK activity, the cells retained the sensitive or resistant phenotype of the originally screened subcolony. Thus, cells were separated and collected based using the result of a cell-lethal assay as selection criteria. These microarrays enabling clonal colony segmentation permitted sampling and manipulation of the colonies at very early times and at small cell numbers to reduce reagent, time, and manpower requirements.