Stand-Sit Microchip for High-Throughput, Multiplexed Analysis of Single Cancer Cells

Cellular heterogeneity in function and response to therapeutics has been a major challenge in cancer treatment. The complex nature of tumor systems calls for the development of advanced multiplexed single-cell tools that can address the heterogeneity issue. However, to date such tools are only available in a laboratory setting and don’t have the portability to meet the needs in point-of-care cancer diagnostics. Towards that application, we have developed a portable single-cell system that is comprised of a microchip and an adjustable clamp, so on-chip operation only needs pipetting and adjusting of clamping force. Up to 10 proteins can be quantitated from each cell with hundreds of single-cell assays performed in parallel from one chip operation. We validated the technology and analyzed the oncogenic signatures of cancer stem cells by quantitating both aldehyde dehydrogenase (ALDH) activities and 5 signaling proteins in single MDA-MB-231 breast cancer cells. The technology has also been used to investigate the PI3K pathway activities of brain cancer cells expressing mutant epidermal growth factor receptor (EGFR) after drug intervention targeting EGFR signaling. Our portable single-cell system will potentially have broad application in the preclinical and clinical settings for cancer diagnosis in the future.

. DNA barcode design and validation. (a) AutoCAD design for the barcode mold yielding 10 different barcodes, each stripe with width of 50 µm. (b) Validation of 10 barcodes by hybridization with Cy3-tagged complementary DNA sequences. Fluorescence intensity was measured for the selected rectangular area of the barcode pattern. Figure S3. Cross-reactivity of assayed proteins. 6-element barcode arrays were used for sandwich ELISA using a mixture of capture antibody-DNA conjugates and corresponding detection antibodies for all conditions, while using one standard recombinant protein per row (shown in red Cy5 signals) at 5 ng/mL concentration. The green-colored array elements serve as reference signals for the spatial address of the barcodes and were developed by hybridization of Cy3-labeled DNA. The major Cy5 signals for protein detection are confined to the designated barcodes with minor crosstalk. Differences in fluorescence intensities of the Cy5 signals may be attributed to variations in the performance of the recombinant proteins and/or antibody pairs. Figure S4. Assessment of on-chip cell viability for the integrated Aldefluor/functional proteomic assay. A column of microchambers (240 chambers with around 50 cells) was chosen randomly and imaged right after loading Aldefluor-stained cells ( t = 0 h) in Aldefluor assay buffer. The assay buffer was exchanged with cell culture medium, then cells were incubated for 4 h. The medium was supplemented with 2 µM calcein-AM to facilitate imaging of viable cells. At t=4 h, viability was still >90%, and this suggests that majority of cells retained viability within the incubation period required for co-detection of ALDH and signaling proteins.

Characterization of On-Chip Mass Transport Processes (Sit State)
The chip underwent several "sit' states (sit-closed or SC, sit-open or SO) for the dual assay and we sought to characterize the mass transport processes associated with each state as well as the transition from sit-closed (SC) to sit-open (SO) states.

Sit-closed
We visualized the convective mass transfer process for the flow of solutions from the inlets to the outlets of the chip using blue and orange dyes. We first flooded the chip with a blue dye solution at the stand state, then set the chip to SC at t=0, after which a second dye solution was fed to inlets and the chip was oriented vertically. Images were taken every two minutes in the span of 10 minutes (Fig. S2).

Transition from sit-closed to sit-open
This transition is associated with opening the ducts and bridging the channels and chambers. We first filled the chambers with PBS, followed by the channels with 1 ug/mL Dylight 488 at SC. The adjustment screw was then rotated by 180 degrees to reach SO, then images of the chip were taken in a span of 15 minutes.

Sit-open
a. Diffusion of small molecules (~1 kDa) from channels to microchambers, then from microchambers to channels Dylight 488 was used as a model small molecule, with a molecular weight of ~1 kDa. The chip was first filled with PBS at the stand state, then set to SC, followed by SO at t = 0.
Dylight solution (1µg/mL) was allowed to flow through the channels in 10 minutes to exchange PBS from the channels. Afterwards, the chip was oriented horizontally to allow simple diffusion. Images were taken every 20 minutes.
Diffusion from microchambers to channels was visualized using the same method, except that chambers were first filled with Dylight, and channels with PBS.
b. Diffusion of small molecules (~1kDa) from microchambers to channels Streptavidin conjugated with Alexa 647 fluorophore (average molecular weight of 60 kD) was used to visualize diffusion of large molecules into microchannels while the chip was in the SO state. The chip was filled with 40 µg/mL SA-647 in PBS while in the stand state, after which the chip was converted to SC then SO where PBS was used to wash away the SA-647 from the microchambers. After washing for 10 minutes, images were taken at 1hour intervals to visualize the diffusion of SA-647 from microchambers into microchannels.
The diffusion of secreted and intracellular proteins in the microchambers was calculated using

Einstein-Smoluchowski equation for diffusion:
(1) Where L is the mean displacement of the molecule within time interval t, and D is the diffusion coefficient (in water). The diffusion time was calculated for distances 100 µm up to 500 µm, which is the distance from the center of a microchamber to the duct openings. For the diffusion of Dylight 488 (~1 kDa), we used D = 3 x 10 6 cm 2 /s based on the diffusion coefficient of Cy5 (~0.7 kDa) in water a 25 °C, which was estimated to be 3.6 x 10 6 cm/s. Thus, the diffusion throughout the 1 mm microchamber only needs ~0.5 h, and 3 min through ducts. However, the small cross-section ducts significantly limit the amount of molecules exchanging between microchannels and cell chambers, as the diffused amount can be roughly estimated by movement rate x cross section area. The typical proteins (10-50 KDa) in our assay need 1-2 hours to diffuse from one end to the other end. They might have been captured by the antibody array once released by the cells. Thus, after cell lysis, we set 2 h for proteins sufficiently depleted by the antibody array.