Modernizing the MTT assay with microfluidic technology and image cytometry

Authors

  • Michael Halter

    Corresponding author
    1. Biochemical Science Division, National Institute of Standards and Technology, Gaithersburg MD20899
    • Cell Systems Science Group/Biochemical Science Division, National Institute of Standards and Technology, 100 Bureau Drive, MS 8313, Gaithersburg, MD20899-8313
    Search for more papers by this author

  • Published 2012 Wiley-Periodicals, Inc. This article is a US government work and, as such, is in the public domain in the United States of America.

The combination of microfluidic technology and imaging provides a new set of cytometric tools that can be quantitative and highly informative. Microfluidic devices can be used to capture small numbers of cells and expose those cells to a unique set of conditions (e.g., growth factors, toxins, and shear forces) by controlling the chemical and physical environment around the cells. The processes used to capture cells and manipulate the environment can all be controlled “on-chip,” reducing the number of manual handling steps, allowing for automation, minimizing the quantities of reagents required. Because the microfluidic device can be frequently placed on a microscope stage, quantitative, time lapse imaging can be used to record the dynamic responses of the cells. In the extreme, large numbers of individual cells can be specifically manipulated all while images are acquired on an automated microscope.

In this issue of Cytometry, Lim et al. (page 691) adapt the standard methyl tetrazolium (MTT) assay (1) to a microfluidic device and use automated imaging and image analysis to obtain single cell measurements. The MTT assay is a colorimetric assay that is typically performed in 96-well microtiter plates and at the end of the assay an absorbance measurement is made from each well using a plate reader. In routine applications, the MTT assay is used to assess cell viability or to measure cell proliferation. Because pharmaceutical agents or toxins can stimulate or inhibit cell growth, the MTT assay is frequently used as cytotoxicity assay. Typically, the treated and control cells are incubated for 2–4 h with the MTT reagent, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, at concentrations in the range of 1 mM. Metabolically active cells reduce the MTT reagent into purple formazan crystals, presumably through mitochondrial enzymes such as succinate dehydrogenase. Before reading the plate, the crystals are dissolved with sodium dodecyl sulfate, and then the absorbance is measured for each well.

The results of the standard MTT assay are ambiguous because the absorbance measurement alone cannot be used to distinguish between changes in average enzyme activity of the cell population from changes in the number of cells that can occur by either cell proliferation or death. Furthermore, the complex set of reactions within cells and the heterogeneity at the level of the cell population is reduced into a single number. Nonetheless, the MTT assay is one of the most popular assays for assessing toxicity and measuring proliferation. The MTT assay is probably used so frequently because of how easy it is to implement.

In this issue (page 691), Lim and coworkers develop an MTT cell toxicity assay that reduces the ambiguity and increases the information content of the standard MTT assay. Their assay is developed on a polydimethylsiloxane (PDMS) microfluidic platform that allows the toxin, Cd2+, to be introduced to the cells over a large range of concentrations using an on-chip gradient generator. Cells are loaded into 30 isolated chambers on-chip, cultured in the presence of the Cd2+ toxin for 48 h, and then the MTT reagent is added. Instead of dissolving the formazan crystals, Lim and coworkers image the cells in situ in the microfluidic device and derive absorbance and morphology measurements for each cell. Imaging of the formazan crystals is rarely performed but can be highly informative. Another study that imaged the formazan crystals formed during the standard MTT assay was published in a previous issue of Cytometry. Using confocal imaging, Bernas et al. (2) determined that only 25–45% of the formazan crystals were associated with the mitochondria. Their study suggests that non-mitochondrial associated enzymes are primarily responsible for the reduction of the MTT reagent into formazan crystals. Although this suggests that the optical absorbance will not indicate mitochondrial activity specifically, the MTT results will nonetheless indicate the overall redox potential of the cell.

In the study by Lim and coworkers (page 691), an image analysis procedure is developed using an ImageJ macro to segment each cell, and then to determine its optical absorbance and its circularity. They use the circularity, which is a scalar derived from the ratio of the cell area to the square of its perimeter, to measure changes in cell morphology in response to the Cd2+ toxin. The authors use these two derived parameters to interpret the cellular response to toxin as four subpopulations: +MTT absorbance/−morphology change; +MTT absorbance/+morphology change; −MTT absorbance/−morphology change; and −MTT absorbance/+morphology change. One of the outcomes of this analysis is that the morphology change is found to be a more sensitive indicator of Cd2+ toxicity. In contrast to the standard MTT assay, the cell-by-cell analysis of this assay is capable of distinguishing between changes in cellular enzyme activity and cell number.

Another consequence of the cell by cell analysis is that the assay can detect variations in enzyme activity between individual cells. Indeed, the authors report more than an order of magnitude difference in MTT absorbances across the population even when all the cells in the population are exposed to the same level of toxin. This suggests that in the typical MTT assay, population heterogeneity can be completely obscured by the population averaged measurement and that the results may be dominated by the large activity of a small number of cells. Cell to cell differences are always present in a population of cells, and this heterogeneity is becoming increasingly recognized as important to measure in cell biology assays (3).

Any cell-based assay is complex and can yield analytical results that are challenging to interpret. When the assay includes microfluidic and imaging components, several additional complicating factors need to be considered: the materials used in the microfluidic device, design of the flow system, procedure for loading cells, image acquisition and analysis, and others. The study presented by Lim et al. illustrates several procedures used to assure the quality of the assay. Measurements of cell density are reported for determining the protocol for loading cells and the time provided for cells to adhere and grow in the device. The incubation time of the MTT reagent was determined by identifying when the formazan absorbance stopped increasing, presumably because the cytosol was fully reduced. They also validated the toxin concentration gradient, and that the device can maintain a stable concentration gradient throughout the assay lasting more than 48 h.

A recent study by Cooksey et al. focused on the issues related to assay quality in a microfluidic device (4). Similar to Lim et al., Cooksey and coworkers adapted a cell-based assay to a microfluidic platform. Cooksey and coworkers used a green fluorescent protein (GFP)-based imaging assay for ribosome activity (5) and a PDMS microfluidic device with 64 chambers. In their assay, the GFP intensity and number of cells in each chamber was determined. The sensitivity of the assay readout to the following microfluidic specific parameters was examined: response in a microfluidic environment compared to a standard tissue culture polystyrene dish, the effect of the tubing material used to interface the device with the cells and media, under static conditions or in the presence of flow, illumination source, and cell attachment substrate. The authors intentionally inhibited the ribosome activity of the cells using cycloheximide. Interestingly, the proteosome degradation indicated by the GFP intensity was insensitive to any of the parameters they studied. In contrast, the proliferation rate was highly sensitive to whether cells were culture in the microfluidic chamber compared to bulk culture. The control experiments performed by both studies illustrate the added layer of effort required to validate and qualify these new cytometric methods that include microfluidic and imaging components.

Although these advanced cytometric methods have a large number of parameters that need to be controlled, it is clear that they can provide unique and quantitative information and enable new biological understanding. One excellent example of this was reported by Lecault et al. where they determined the time dependent effects of stem cell factor (a.k.a. steel factor) on the cycling potential of hematopoietic stem cells (6). Lecault et al. used a 6,144-chambered device to isolate 769 single cells and collect 5 days of time lapse imaging data. Using programmable control of medium exchanges to systematically investigate the timing of steel factor stimulation, they found that 24 h of incubation with steel factor caused irreversible effects on cell fate decisions. Critical features of their microfluidic cell-based assay include the large number of chambers, non-perturbing cell culture conditions compared with bulk culture, dynamic control of medium exchanges, robust long-term cell culture, and automated imaging and image analysis. Another example where the marriage between microfluidics and imaging can yield high quality, quantitative information to gain new biological insight was reported by Figueroa et al. (7) Using a microfluidic device, they were able to routinely screen 20,000 single olfactory sensory neurons freshly harvested from a mouse for olfactory function. By imaging the time-dependent response of a fluorescent calcium dye, they could identify 2,900 olfactory sensory neurons from the larger pool of primary cells. They could further determine the set of particular stimulants that these cells would respond. These studies are possible because of the advanced cytometric methods that have only recently become available by combining microfluidic technology and optical imaging. Interestingly, almost identical technology has enabled the high throughput evaluation of small, multicellular organisms such as nematodes, fruit flies, and zebrafish. In this field that has been termed “wormometry,” a number commercial instruments are already available (8).

For cell-based assays, devices that combine microfluidics and imaging are clearly of interest to the cytometry community. These technologies will have increasing clinical importance in the near future. For example, Takao and Takeda recently described a simple, sensitive, and accurate circulating tumor cell assay based on microfluidics (9). Their assay has several advantages over existing tests used in hospitals to guide therapeutic regiments for cancer patients. As these technologies develop, there is likely a unique role for societies such as the International Society for Advancement of Cytometry (ISAC) to add value to this field by supporting studies that will help validate and qualify these assays. By developing and reporting methods that help assure the measurement repeatability, reproducibility, and robustness of these assays, the analytical rigor needed for the widespread acceptance of these tools by the cytometry world will be furthered. The study by Lim et al. takes the standard MTT assay, one of the most common cell-based assays, and develops it into an assay that is likely to be more informative and robust. Translating their assay into the new standard MTT assay could be a benefit to many cell biology laboratories all over the world.

Ancillary