A Dynamic Live Cell Assay Platform

A dynamic live cell assay platform to elucidate the mechanisms underlying autophagy

A Dynamic Live Cell Assay Platform

Autophagy (self-eating) is a complex cellular process essential for cell survival under stressed conditions. Recent evidence has further suggested that tumor cells use the autophagic pathway to promote survival under stressed conditions, such as nutrient deprivation or hypoxia. However, due to use of static, end-point assays, the dynamics of autophagy during the cell’s stress and recovery phases are not fully understood.

Here, we report a dynamic, live cell assay to monitor both the rate of autophagosome formation and changes in lysosomal degradative processes during autophagy. By combining live cell analysis with cell lines stably expressing fluorescently-tagged markers specific for autophagosomes (LC3-GFP), the cellular dynamics of autophagy can be visualized at the single cell level in real time. Using a microfluidic system capable of temporally regulating media and gas exchange within a culture chamber, LC3-GFP CHO reporter cells were exposed to starvation or hypoxic conditions in the presence of various stressor compounds. Changes in LC3 levels (as measured by autophagosome counts) were monitored throughout culture duration using a standard fluorescence microscope. The real-time analysis capabilities of the CellASIC® ONIX platform demonstrated here for autophagy and associated pathways will not only provide new insights into how cancer cells may avoid therapeutic intervention, but also have larger implications for the investigation of such dynamic cellular processes as migration, apoptosis, proliferation, and cell-cell interactions.


Autophagy is an intracellular process leading to the lysosomal degradation of cytosolic components and organelles. The best understood role for autophagy is in cellular housekeeping; this activity, present at low levels in normal cells, directs the removal of damaged or unwanted products 1 . However, autophagy can also be induced in response to cancer therapies. In such situations, autophagy functions as a survival mechanism and thus potentially limits drug efficacy 2,3 . In established tumors, malignant progression and tumor maintenance have been linked to physiological adaptations resulting in upregulated or constitutively active autophagic pathways 2 . Further, there are many stimuli that have been shown to activate autophagy. These stimuli include nutrient starvation (a well-characterized inducer of autophagy), reactive oxygen species (ROS) 4 , stress on the endoplasmic reticulum, and ammonia 5. Once stimulated, unwanted cytosolic proteins and aging organelles are sequestered by a double membrane vesicle known as an autophagosome (Figure 1). ATG protein complexes coordinate vesicle formation and enable the recruitment of LC3 (also known as ATG8/microtubule- associated protein 1 light chain 3) into the inner and outer membranes of the autophagosome. LC3-labeled vesicles are trafficked to the lysosome. During this last phase, autophagosomes fuse with lysosomes to form autolysosomes, where unwanted nutrients are reduced to basic molecular building blocks (amino acids, fatty acids, and nucleotides) and ultimately released back into the cytoplasm.


Figure 1. Four stages of autophagy. Autophagy can be induced by nutrient depletion or inhibition of mTOR pathway. During autophagy, cytosolic proteins and aging organelles are sequestered by a double-membraned autophagosome. One of the hallmarks of autophagy is translocation of LC3 from the cytoplasm to the autophagosome. Autophagosomes then fuse with lysosomes to promote breakdown of the vesicle and all contents, including LC3. This process can be visualized using either a LC3-GFP fusion protein or an anti-LC3 antibody.

Measurement and tracking of autophagy are essential for elucidating this process under normal physiological conditions as well as its role in cancer and metastasis. Many newer autophagy assays rely on the expression of stably transfected green fluorescence protein (GFP)-LC3 fusion proteins; in this case, autophagosome activity is visually identified by changes in GFP puncta6. Lysosomal inhibitors, such as chloroquine (CQ), have also been invaluable in determining the relative autophagic response to cellular stress. CQ blocks the last step of autophagy, lysosomal degradation; the resulting buildup of intermediates can serve as a quantifiable marker of autophagic activity 7 . By visualizing live cells transduced with a GFP-tagged autophagosome marker (LC-3) in the presence of CQ, researchers can monitor autophagosome formation process on a fluorescent microscope in real time. However, little is yet known about the latter stages of autophagy and the dynamics of lysosomal degradation.

Here, we exploited a microfluidic live cell analysis platform (the CellASIC® ONIX Microfluidic Platform with M04S plates) to develop a dynamic cell-based assay for monitoring the whole autophagy process. This platform offers temperature and gas control as well as media perfusion for precise environmental control within the associated culture chamber. Using this system, LC3-GFP CHO reporter cells were subjected to nutrient starvation or hypoxic stresses for a designated time period followed by reintroduction of normal growth conditions. The time course of autophagy was visualized under a fluorescent microscope.

By this assay, lysosomal degradation in the LC3-GFP CHO reporter cells was first observed within four hours after the recovery growth medium was added. In addition, cells responded very rapidly to hypoxia (within 30 minutes of the gas switch) and were unable to survive beyond six hours of hypoxic treatment. However, studies involving LC3-GFP CHO reporter cells transduced with a LAMP-RFP construct to permit visualization of lysosomal function in concert with autophagosomes, were hampered by an apparent stress-related response to the transfection process. We are currently seeking alternative labeling methods for the dual-color hypoxia-induced autophagy assay.

Combining live cell analysis, fluorescently-tagged reporter cells, and the unique microenvironmental control capabilities of the CellASIC® ONIX platform, we were able to expose cells to different stressors (starvation and hypoxia) and monitor the entire process of autophagy for the very first time. The assay method described here provides quantitative information on both autophagosome formation and lysosomal degradation machinery, and thus provides a platform for the discovery of new targets and therapeutic compounds in cancer as well as other diseases.

Materials and Methods

Cell culture, reagents, gas tanks and live cell analysis systems

The FlowCellect® CHO GFP-LC3 Reporter Autophagy Assay Kit (Cat. No. FCCH100170) was used according to the manual. Cells were maintained in F12-K medium containing 2 mM L-glutamine, 10% heat-inactivated fetal bovine serum (EmbryoMax® FBS, Cat. No. ES-009-B), 100 Units/mL penicillin, 100 µg/mL streptomycin (Cat. No. TMS-AB2- C), and 250 µg/mL Geneticin® (G418) under 5% CO 2 at 37 °C for up to 10 passages after cells were thawed from the cryopreserved vial.

Earle’s Balanced Salt Solution (EBSS) for inducing starvation stress was purchased from Sigma-Aldrich, and the lysosome inhibitor chloroquine (CQ) was purchased from Life Technologies. Pre-mixed gas tanks of different compositions were purchased from Science and Technical Gases Ltd. Live cell analysis was performed using the CellASIC® ONIX Platform, consisting of the CellASIC® ONIX System (Cat. No. EV262), Tri-gas Mixer (Cat. No. GM230), Microincubator Controller (Cat. No. MIC230) and CellASIC® ONIX M04S Microfluidic Plates (Cat. No. M04S-03- 5PK). A LAMP1-RFP BacMam transduction kit was purchased from Life Technologies.

Cell seeding on M04S plates

To seed the cells, the cell suspension was prepared according to the manual at the density of 1 x 106 cells/ mL, allowing close cell-cell contact. Before introducing the cells into the microfluidic plate, the liquid in inlet wells 1, 6, 7, and 8 was first aspirated. The inner ring in wells 6 and 7 was then carefully aspirated (extended aspiration results in bubbles being introduced into the microfluidic channel). Ten µL of cell suspension was added into the inner ring of well 6. The plate was placed in a laminar cell culture hood for 30 minutes to allow cells to load into the microchambers by capillary action. The plate was placed inside a standard CO 2 incubator. To promote cell attachment, 300 µL of the culture medium was added into well 1 and gravity-driven perfusion was used to slowly deliver culture medium to the microchamber. For prolonged culture in the incubator, medium was aspirated in well 1 and 7 every 48 hours and 300 µL of the culture medium was added into well 1 to re-establish gravity-driven perfusion. The medium was replenished every two days.

Live cell analysis for the dynamic starvation-induced autophagy assay

Phosphate-buffered saline (PBS) was aspirated into wells 1, 2, 3, 6, 7, and 8, but the liquid in the inner holes was not disturbed. 300 µL of desired treatments (CQ of different doses in 1X EBSS in this case) and culture medium was added into well 2 and 3, respectively. The plate was sealed to the heater manifold and microincubator controller as described in the manual. The flow program was created using the CellASIC® ONIX FG Software using the following parameters:

V3, 1 psi, 135 minutes
V2, 1 psi, 255 minutes
V3, 1 psi, 240 minutes

As a general guideline, 300 µL in each well, when set to flow at a pressure of 1 psi, can provide ample nutrients to the cell cultures for up to 18 hours.

Validation of the gas exchange in the microfluidic culture chambers

Oxygen concentration within culture chambers was measured using oxygen-sensitive spots from PreSens (SP-PSt3- NAU-D2- YOP) and a fiber optic transmitter to allow reading of the oxygen concentration through the glass bottom of the plate. This noninvasive oxygen sensor uses sinusodially modulated light and calculates the phase angle difference between excitation and emission light. The spots are 2 mm diameter, less than 200 µm tall and capable of measuring O 2 concentration from 0 to 100% O 2 , with a resolution 0.1% O 2 or less and accuracy within 0.4% O 2. Response time is less than six seconds.

We then constructed a control calibration dish by attaching an oxygen sensor in an enclosed chamber and performed a two-point calibration in air and nitrogen. We fabricated a custom test microfluidic plate by placing the probes inside the cell culture chambers of an M0 4 S plate after oxygen plasma modification and prior to bonding to the glass slide. Channels and chambers were primed with water as normal. The test plate was sealed to the manifold of the CellASIC® ONIX Microincubator Controller, and the temperature set to 37 °C. Multiple gas switches from air to N 2 and back at an air flow of 100 mL/min were then applied, allowing the gas concentration to stabilize for over two hours under each condition. We later found that the O 2 concentration within the microfluidic chamber consistently reached within 10% of the newly delivered gas concentration in under 60 minutes following a switch and would eventually reach within 0.4% O 2 of the supplied gas mixture’s O 2 concentration.

Live cell analysis for the dynamic hypoxia-induced autophagy assay

Before the experiment, the Microincubator Controller was connected to a pre-mixed gas tank containing 95% air and 5% CO 2 (normoxic gas line) and purged for 30 minutes at a flow rate of 3 mL/minute. Similarly, the hypoxic gas line, configured with a pre-mixed gas tank containing 94.8% N 2 , 0.2% O 2 , and 5% CO 2 , connected to the Tri-gas Mixer, was purged as above. To start the experiment, PBS was aspirated from wells 1, 2, 3, 6, 7, and 8; liquid was left in the inner holes. Next, 300 µL of the desired treatments (CQ of different doses in growth medium, leaving hypoxia as the only source of stress in the experiment) and culture medium was added to wells 2 and 3, respectively. Then the plate was configured as previously described. The flow program was created using the CellASIC® ONIX FG Software using the following parameters:

V3, 1 Psi, 120 minutes
V2, 1 Psi, 180 minutes
V3, 1 Psi, 960 minutes

Live cell image analysis

An Olympus IX-71 inverted microscope was used for live cell viewing. All images were taken under the 40x objective. The number of autophagosomes for each image was determined using a custom developed image processing sequence for object identification in CellProfiler Software (version r11710, Broad Institute). Autophagosome-sized areas of locally higher intensity were first enhanced using a white top-hat transform, then identified and counted by applying a fixed threshold and an intensity-based declumping algorithm (see technical application note, manuscript in preparation).

Results and Discussion

To validate the media exchange capability of the CellASIC® ONIX platform as well as the ability to monitor and quantify autophagy through autophagosome counting, LC3-GFP CHO cells (70% confluent) were perfused with EBSS + 50 µM CQ for 100 minutes followed by regular culture medium for 200 minutes. As shown in Figure 2 (and the image analysis results in Figure 3), the dynamic changes of autophagy in both the stress and recovery phase could be quantified through autophagosome counting.


Figure 2. Schematic of live cell images for autophagy of LC3-GFP expressing CHO cells. First, medium was perfused to establish fluorescence baseline. A stressor (serum starvation) and the lysosome inhibitor CQ1 were then introduced to trigger autophagosome accumulation within cells. When cells were returned to standard growth medium, autophagosomes underwent lysosomal degradation.


Figure 3. Validation of the autophagosome counting assay. The number of autophagosomes in each image in Figure 2 was determined using a custom-developed image processing sequence for object identified with CellProfiler Software. Rate of autophagosome formation and degradation was successfully monitored with the proposed assay.

Once the assay was validated, profiling of the CQ dose response in CHO cell lines was conducted. Once established in the microchamber, exposure conditions involved 3 phases: standard culture medium for 135 minutes, continuous CQ (10 µM, 100 µM, or 1 mM) perfusion for 255 minutes, and culture medium for the final 240 minutes to permit visual capture of the lysosomal degradation process. Images were taken every 15 minutes on a group of roughly 20 cells at three positions per chamber, for all four culture units.

Overall, the rate of autophagosome formation was proportional to the CQ concentration applied. However, at 1 mM, cells ceased committing to the autophagy pathway, and the number of autophagosomes stayed constant for the rest of the experiment. We also observed more dead cells in this treated group, indicating either that the maximal levels of autophagy in this cell line had been achieved, or that the cells committed to apoptosis or necrosis at the high CQ dose. Furthermore, degradation of autophagosomes occurred at a faster rate than the accumulation


(Figure 4).

Figure 4. Assessment of CQ dose response using a dynamic autophagy assay. Three levels of CQ (10 µM, 100 µM, and 1 mM) were perfused through independent culture chambers of the same microfluidic plate at the same time. Time-lapse analysis was performed on 3 different positions each of the microchambers. The fluorescence intensity was counted and averaged per frame using CellProfiler Software and normalized to the background to measure flux. Error bars represent standard deviation (S.D.) of the number of counted puncta in approximately 60 cells per time point.

To further explore the dynamics of stress-induced autophagy, we exploited the CellASIC® ONIX system’s ability to regulate gaseous microenvironments to introduce severely hypoxic conditions within the cell chamber. Prior to analysis, the system’s control of oxygen content was validated. For gas flow rates of 20 mL/min and 3 mL/min, we consistently found that switch times occurred in less than one hour. For these two gas flow rates, steady-state concentrations were achieved with less than 2% and 10% deviation from the supplied gas, respectively.

In traditional static cultures, achieving equilibrium following defined gas switching is impractical due to incubator size and differences between the measured pericellular oxygen tension (within the flask) and that in the ambient air 8,9 . However, the CellASIC® ONIX Microfluidic Platform features a significantly reduced culture vessel size (10,000 cells per chamber) and restricted fluid volume (a few nanoliters), together leading to a faster gas exchange during our hypoxia studies.

Results from initial hypoxia experiments supported this fact; specifically, we found that, compared to the typical hypoxic response of cells cultured in traditional petri dishes 8-12 , LC3-GFP CHO reporter cells in the microfluidic perfusion environment were far more sensitive to gas switching, demonstrating autophagosome formation within three hours of hypoxic treatment. Following a six-hour exposure, a large percentage of cells failed to recover and underwent apoptosis (data not shown).

Based on these preliminary results, we performed dynamic profiling of autophagosome formation in reporter cells in response to CQ (10 µM, 100 µM, or 1 mM) under hypoxia conditions. Similar to results of starvation-induced autophagy, the rate of autophagosome appearance accelerated with respect to increasing CQ dose. As for the recovery phase, cells treated with 100 µM of the CQ responded almost instantaneously, while those treated with the highest dose (1 mM) demonstrated a far more protracted recovery profile (Figure 5 and 6).


Figure 5. Images of the LC3-GFP CHO reporter cells cultured on the CellASIC® ONIX system taken during each phase of the hypoxia-induced autophagy assay. Cultures were first perfused with standard growth medium under normoxic conditions for 60 minutes, followed by continuous CQ (10 µM, 100 µM, or 1 mM) perfusion in the presence of severe hypoxia (0.2% O 2 ) treatment for 3 hours, followed by removal of the stressors and reestablishment of normoxia in standard growth medium for another 16 hours. An Olympus IX-71 inverted microscope was used for the entire process; all images were taken under the 40x objective.


Figure 6. Dynamic, quantitative live cell analysis of autophagosome formation with respect to CQ treatment and hypoxia. Three levels of CQ inhibitor (10 µM, 100 µM, and 1 mM) were perfused through independent culture units at the same time. Time-lapse analysis was performed on 3 different positions in each of the microchambers. The fluorescence intensity was counted and averaged per frame using CellProfiler Software and normalized to the background to measure flux. Error bars represent S.D. of the puncta in around a total of 60 cells per time point.

To simultaneously monitor the two most important organelles involved in autophagy, we further transduced the LC3-GFP reporter CHO cells with a fluorescently tagged, lysosome-specific fusion protein construct, LAMP1-RFP. Transduced cells were incubated under mildly hypoxic conditions (3% O 2 ) in the presence of CQ at various concentrations for 180 minutes, followed by prolonged culture (660 minutes) under normoxia in the presence of standard medium. The data indicate that autophagosome formation started immediately after the switch to hypoxic conditions and lasted for three hours in the cells treated with 1 mM of CQ. In these cultures, lysosome degradation did not occur until almost 11 hours after gas exchange (Figure 7). However, we did not observe any conclusive response in the lysosomal activity during either autophagy or recovery phases except for the observation that lysosomes were instantly condensed under the hypoxic stress. We speculate that the LAMP1- RFP transduction process (or the LAMP1-RFP construct itself) might be another source of cellular stress, hence effecting overall autophagic activity. We are currently exploring alternative labeling methods for dual-color assays for hypoxia-induced autophagy.


Figure 7. Two-color images of transduced LAMP1-RFP/ LC3-GFP CHO reporter cells showing autophagosomes (green) and lysosomes (red) throughout the entire hypoxia-induced autophagy assay. Cells were first treated with mild hypoxia (3%) for 3 hours in the presence or absence of CQ, followed by the removal of the stressors and the treatment of normoxia and the growth medium for another 16 hours. An Olympus IX-71 inverted microscope was used for the entire process, and all images were taken under the 40x objective.


We have used the CellASIC® ONIX live cell analysis platform to create a dynamic assay that not only has the potential to simultaneously monitor multiple intracellular components throughout the entire autophagic process without disruption but also allows precise manipulation of culture parameters (medium flow, inducer/inhibitor concentration, gas content), thus exposing cells to more physiologically relevant conditions. This platform may be capable of simulating conditions of pulse exposure to drug compounds and could provide additional information on dose response for compound profiling by revealing rate of autophagosome formation and degradation.



  1. Cuervo, A. et al. (2004). Autophagy: in sickness and in health. Trends in Cell Biology; 14(2); 70-77.
  2. Kimmelman, A. et al. (2011). The dynamic nature of autophagy in cancer. Genes and Development 25: 1999- 2010.
  3. Amaravadi, R. et al. (2011). Principles and current strategies for targeting autophagy for cancer treatment. Clinical Cancer Research 17: 654-666.
  4. Neufeld, T. et al. (2010). TOR-dependent control of autophagy: biting the hand that feeds. Current Opinions of Cell Biology 22: 157-168.
  5. Cheong, H. et al. (2011). Ammonia-induced autophagy is independent of ULK1/ULK2 kinase. Proc Natl Acad Sci 108: 11121-11126.
  6. Mizushima, N. et al. (2010). Methods in mammalian autophagy research. Cell 140 (3): 313-326.
  7. Yamamoto A. et al. (1998). Bafilomycin A1 prevents maturation of autophagic vacuoles by inhibiting fusion between autophagosomes and lysosomes in rat hepatoma cell line, H-4- II-E cells. Cell Structure and Function 23: 33-42.
  8. Baumgardner, J. et al. (2003). In vitro intermittent hypoxia: challenges for creating hypoxia in cell cultures. Respiratory Physiology and Neurobiology 136: 131-139.
  9. Bambrick, L. et al. (2011). In vitro cell culture pO2 is significantly different from incubator pO2. Biotechnology Progress 27: 1185-1189.
  10. Song Y. et al. (2013). Autophagy contributes to the survival of CD133+ liver cancer stem cell in the hypoxic and nutrient-deprived tumor microenvironment. Cancer Letters (in press).
  11. Grkovic S. et al. (2013). IGFBP-3 binds GRP78, stimulates autophagy and promotes the survival of breast cancer cells exposed to adverse microenvironments. Oncogene 32: 2412-2420.
  12. Hubbi M. et al. (2013). Chaperone-mediated autophagy targets hypoxia-inducible factor-1 α (HIF-1 α) for lysosomal degradation. The Journal of Biological Chemistry 288: 10703-10714.


Related Links