High-Throughput Live Cell Screening

A Novel Multiplexed Live Cell Screening Technology

Poster presented at ASCB 2008
S. Henderson, A. Cabasug, A. Stokes, D. Poon
CT-1700 cellTRAY for Live Cell Screening

Screening of siRNA libraries is a complex and expensive pursuit. Current technology usually involves a large number of 96 or 384 well plate assays, with a fluorescent, or colorimetric end point readout. However, a novel real time microscope based screening technology has emerged, which utilizes up to 7680 micro-etched wells in either a disposable plastic, or reusable glass  slide. This system, called the cellTRAY® (CT-1700) can be spotted with different siRNA oligonucleotides or cDNA encoding shRNA constructs in each well, which can then be transfected into cells that are seeded over the spotted cellTRAY.

Using this method, live-cell experiments or assays can be run in a high-throughput manner. This ability opens the possibility for the screening of large nucleotide based libraries for functional analysis on live cells.

The following experiments provide proof-of-concept for spotted, addressable and isolated cDNA arrays allowing for differential expression/knockdown of genes in live cells with implications in high-throughput screening of gene libraries.


Transfection Optimization: An eGFP construct was pre-spotted onto the device then transfected in HEK293T cells under various conditions in order to optimize the transfection protocol. Cell counts of eGFP expressing cells and total cells (stained with Hoescht 33343 nuclear stain) using ImageJ image analysis software was employed to predict transfection efficiencies.

shRNA Inhibition Validation: A cDNA encoded anti-eGFP shRNA construct was co-transfected with an eGFP expressing construct into HEK293T cells. eGFP expressing cells were then counted using ImageJ image analysis software compared to a control transfected using only an eGFP construct.

Apoptosis Assay: Initial knockdown experiments were conducted using JC1 apoptosis stain (Invitrogen). shRNA constructs of anti-BAX, anti-Bcl-2, and anti-ASK1 were pre-spotted into specific regions of the device and transfected into HEK293T cells. Cells were then incubated for 48 hours then labeled with JC1 stain. Fluorescence measurements were taken using a ~488nm filter and ~600nm filter. Relative intensity values (600nm/488nm) were calculated using ImageJ image analysis software.

Transfection Protocol Optimization

Transfection conditions were optimized for the spot area of each well of the platform using a total transfection mixture volume of 50 µL. The optimized transfection parameters used 2.5 µg of DNA plasmid, a lipid volume of 7.5 µL (LT1, Mirus Bio), a DNA to lipid ratio of 1:3, and a cell seeding density of 3.0*105 cells/mL.

Diagram of transfection optimization
Transfection Optimization: Transfection Efficiency vs, Increasing Concentration/Ratio. Shaded column) DNA µg concentrations (0 µg, 1 µg, 2.5 µg, 5 ug., 7.5 µg, 10 µg) in 50 µL transfection mixture. Open column) Lipid concentrations (0 µL, 5 µL, 7.5 µL, 10 µL, 12.5 µL, 15 µL) in 50 µL transfection mixture.  Line) DNA:Lipid Ratio (0:1, 1:2, 1:3, 1:4, 1:5, 1:6).

Each active well was hand-spotted with 1.5 µL transfection mixture, however, the possibility for machine optimized spotting using a spotting pin set-up or other automated technology could be easily adapted. The basic transfection procedure was based upon Ziauddin et. al. 2002, which used a filtered 1% gelatin solution. This allows for a non-cytotoxic substrate for the lipid-DNA complexes to settle within a spotted area without the need for any substrate-linking reactions. The matrix formed by the 1% gelatin upon drying seems to hold the lipid-DNA complexes in the spotted area even after cell seeding , which seems to localize transfection and subsequent protein expression to the spotted area exclusively.

Transfected Cells
Two wells containing transfected HEK293T cells showing well-exclusivity. Brightfield image overlayed with green fluorescence-captured image using 488nm filter taken from the same location.   A) Unspotted well.  B) Spotted well transfected with, and expressing eGFP construct.

The proprietary software used for image-capturing was specifically designed for the cellTRAY® platform. This enabled the capture of one image-per-well with complete customization through a bookmarking function allowing the imaging of specific wells-of-interest anywhere on the cellTRAY®. The ability to perform whole-tray or partial-tray scans is the basis for the platform’s high-throughput functionality, and is necessary for a platform containing >640 wells. In a high-throughput screen when a small fraction of wells could turn up as possible “hits”, the ability to scan through every well quickly and pick a fraction of wells to monitor over time is a significant advantage for this particular application.

shRNA Inhibition Validation & Apoptosis Assay

With the ability to monitor specific wells over time, an apoptosis assay was designed using lentiviral shRNA constructs (Open Biosystems) against specific genes in the apoptosis pathway. The shRNA construct encodes a 30 base pair (bp) mircro-RNA (miRNA) which, when transcribed, folds into a hairpin structure containing a 22 nucleotide (nt) double-strand RNA (dsRNA) encoding the target silencing sequence with a 19nt loop and 125nt flanking sequence from human miRNA-30 (miR30) which is shown to increase knockdown efficiency (Boden et. al. 2004).

In order to validate the ability for the system to transfect and subsequently knockdown apoptosis related genes, a co-transfection of an eGFP construct and an anti-eGFP shRNA construct was performed.

Relative knockdown efficiency of eGFP by anti-eGFP shRNA
Relative knockdown efficiency of eGFP by anti-eGFP shRNA. Transfection efficiencies produced through analysis of eGFP fluorescent images (488nm filter) compared against total cells counted using Hoescht 33343 nuclear staining (350nm filter). A) No DNA spotted wells. B) eGFP spotted wells. C) eGFP + anti-eGFP shRNA spotted wells. Significant knockdown was observed (p-value=0.008). Cells counted using image analysis software.

Two known and one unknown apoptosis related genes were chosen.  The first gene, Bcl-2, is a known anti-apoptotic gene.  The second gene, BAX, is a known pro-apoptotic gene.  Knockdown of each would provide clear indication of either pro or anti-apoptotic effects.  A third gene, ASK1, whose apoptotic activity is currently under debate, would provide a good test for the type of applications used on the system.

A JC-1 apoptosis assay (Invitrogen) was used to determine the extent of shRNA induced knockdown of the different apoptotic-related genes.  The JC-1 stain is a cationic dye that acts as an indicator of mitochondrial membrane potential.  JC-1 aggregates in the mitochondrial membrane of live cells emitting in the ~600nm range (orange-red).  The onset of apoptosis leads to the loss of mitochondrial membrane potential resulting in the dispersal of JC-1 aggregates.  The non-aggregated JC-1 molecules, in turn, fluoresce in the ~525nm (green) range.  The shift from red to green, as a cell enters apoptosis, provides a method to quantify the extent of apoptosis occurring in a population of cells.

The three shRNA constructs were spotted into separate regions of a cellTRAY and transfected into HEK293T cells.  The cells were then incubated for 48 hours.  JC-1 red and green fluorescence was measured over a time course of approximately 200 minutes. Death slopes were calculated in order to compare rates of apoptosis between each knockdown.

Cell population apoptosis over time

Figure 6. JC-1 apoptosis assay performed on a cellTRAY® spotted with A) No DNA, B) Non-silencing shRNA, C) anti-Bcl-2 shRNA, D) anti-BAX-shRNA, E) anti-ASK1-shRNA. Red/Green intensity is a measure of relative rate of cell population apoptosis over time.

Death slopes for JC-1 apoptosis assay
Death slopes for JC-1 apoptosis assay. No DNA m= -0.0006, Non-silencing shRNA m= -0.0007, anti-Bcl-2-shRNA m= -0.0005, anti-BAX shRNA m= -0.0002, anti-ASK1 shRNA m= -0.0003


Based upon the JC-1 apoptosis data collected, significant knockdown of BAX (p-value= 1.18 x 10-6) and ASK1 (p-value= 1.7 x 10-4) was observed compared to controls. However, the result of the Bcl-2 knockdown was inconclusive as no increase in the rate of apoptosis was observed compared to control (p-value= 0.244). Some possibilities for this particular outcome could be that only partial inhibition of Bcl-2 due to an ineffectiveness of the anti-Bcl-2 shRNA construct was causing no noticeable increase in the rate of apoptosis. Another possibility could be that other anti-apoptotic genes could be masking the expression of Bcl-2 inhibition causing no increase in the rate of apoptosis. Furthermore, future experiments are planned to provide evidence of shRNA knockdown and labeling using fluorescence-conjugated antibodies specific to the proteins targeted above.

These experiments provide an early proof-of-concept for this platform’s ability to perform high-throughput screening and analysis for spotted, addressable and isolated cDNA arrays allowing for differential expression/knockdown of genes in live cells.


  1. Boden, D et. al. Enhanced Gene Silencing of HIV-1 Specific siRNA Using microRNA Designed Hairpins. Nucleic Acids Res. 32. 1154-1158 (2004).
  2. Ziauddin, J., Sabatini, D.M. Microarrays of Cells Expressing Defined cDNAs. Nature 411, 107-110 (2001).
Request More Information

Related Information

CT-1000 Live Cell Imaging System cellTRAY Live Cell Containment Device