What is the recommended culture volume for HeLa cells and seeding number for 384 well plate wells?

What is the recommended culture volume for HeLa cells and seeding number for 384 well plate wells?

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

For 24 well plate wells, thermofischer recommends to seed 0.05E6 cells with 0.5 to 1 ml of medium. What are recommended numbers for a 384 well plate well?

The Privalsky [email protected] UCDavis writes:

Growth Area of Tissue Culture Plates and Dishes

Growth Area (cm2) Media Volume (ml) Maximum Volume (ml) Plates 35 mm 8 1.6- 2.4 60 mm 21 4.2- 6.3 100 mm 44 11 -16.5 150 mm 148 30 -45 245 mm (square) 500 100-150 Dishes 6 well 9.5 1.9 -2.0 16.8 12 well 3.8 0.76 -1.14 6.9 24 well 1.9 0.38 -0.57 3.4 48 well 0.95 0.19 -0.28 1.6 96 well 0.32 0.1 -0.2 0.36 384 well 0.056 0.025-0.05 0.125 Flasks T25 25 5- 7.5 T75 75 15-22.5 T150 150 30-45

Cell seeding density - (Sep/29/2008 )

Hello people,
I am new to cell culture and working with smooth muscle cells.
Can anyone tell me how to calculate seeding density of cells? To be more specific, say I harvest cells from T 75 flask and count is 500,000/ml. Now, how should I calculate number of cells per well if seeded in 12 well plates.
Please, help me out.

Hey Thanks Minnie Mouse. The pdf is really good and helpful. But I want to know how to calculate number of cells per well. Anyone plz.

If you have 20ml in T75 flask, then you will have a total of 1x10^7 (500,000 x 20= 10,000,000)

If you seeded all the cells into 12 wells, then

1x10^7 cells/ 12wells = 833,333cells / well

I think it is too much cells per well.

Would you tell me the volume of the cell suspension in T75flask and the required seeding density in the 12 well plate?

The vol. of media in T75 was 10 ml. The cells were trypsinized, pelleted and suspended in 4ml of media. Counted by coulter counter and count was 500,000/ml. I want to put these cells in equal amount in each well in 12 well plates(1 ml/well) or say 96 well plates (200 ul/well)..What should I do?

Now, Minnie mouse you have divided the number of cells by 12well. So is it no. of wells that matters or amount of suspension (in ml/ul).

Thanks for your response, but I need more

If you have 20ml in T75 flask, then you will have a total of 1x10^7 (500,000 x 20= 10,000,000)

If you seeded all the cells into 12 wells, then

1x10^7 cells/ 12wells = 833,333cells / well

I think it is too much cells per well.

Would you tell me the volume of the cell suspension in T75flask and the required seeding density in the 12 well plate?

. [The vol. of media in T75 was 10 ml. The cells were trypsinized, pelleted and suspended in 4ml of media. Counted by coulter counter and count was 500,000/ml. I want to put these cells in equal amount in each well in 12 well plates(1 ml/well) or say 96 well plates (200 ul/well)..What should I do?

i'll suggest what minnie mouse suggested but in other words:
you have 500.000X4ml=2000000 cells.
for 12well plates:
bring the cells from 4ml to 12.5ml.
then put 1ml in each well. discard remaining 0.5 ml. you'll have 166.666 cells/well

for96well plates:
bring the cells from 4ml to 20 ml.
put 200ul in each well. discard remaining 800ul . you'll have 20.8cells/well

Hey Thanks man for your effort, but how you get 166.666 cells/well, and 20.8 cells/well. Can yo explain it little bit more. Plz.

. [The vol. of media in T75 was 10 ml. The cells were trypsinized, pelleted and suspended in 4ml of media. Counted by coulter counter and count was 500,000/ml. I want to put these cells in equal amount in each well in 12 well plates(1 ml/well) or say 96 well plates (200 ul/well)..What should I do?

i'll suggest what minnie mouse suggested but in other words:
you have 500.000X4ml=2000000 cells.
for 12well plates:
bring the cells from 4ml to 12.5ml.
then put 1ml in each well. discard remaining 0.5 ml. you'll have 166.666 cells/well

for96well plates:
bring the cells from 4ml to 20 ml.
put 200ul in each well. discard remaining 800ul . you'll have 20.8cells/well

[quote name='ambition' date='Sep 30 2008, 01:00 PM' post=�']
Hey Thanks man for your effort, but how you get 166.666 cells/well, and 20.8 cells/well. Can yo explain it little bit more. Plz.

you have to pick a cell and then spit it in 1000 parts. then you pick 666 parts and put them in the well.
come on.
it's just the result of the calculation . it's like saying: in springfield 1.3men/10 are gay or 13men/100 or 130/1000. and so on.

Please check this post for Sigma's cell culture protocol

Thank you guys, for your patience, may be its dumbest thing to ask but still its not very clear.
Let me put in my words, for calculating no. of cells per well, I have to divide number of cells per ml (say 10,000cells per ml) by the total amount of media (say 12.5ml for 12 well plate) required. Is this the right way.

How to calculate the cell seeding density

To make the cellular growth and proliferation data homogeneous and independent of the vessel used, reference has not to be made to the total number of cells in a well but rather to the cells density (cells / cm 2 ), that is the number of cell in 1 cm 2 of growing surface.

I explain better with an example: in a 96 well, that generally has a surface of 0.3 cm 2 , the total number of cells to be plated will be much smaller than the number of cells to be plated in a 25 flask if the same cell density has to be mainteined. So if for a certain cell type you establish that you have to seed 10000 cells in a well of 96, then in a flask of 25 you will have to seed 10000 / 0.3 * 25 cells in total. In this way the density of cells will be kept constant.

If you know that in a well from 96 at confluence you have about 40,000 cells, then in a flask of 25 you will have about 3 million cells. This can be useful to calculate the yield in terms of the number of cells and then make a subsequent experiment.

The formula for calculating the cell number in 1 cm 2 , that is the seeding density (SD) is:

where N = total number of cells (NOT cell/ml)

A = area (expressed in cm 2 ) of the well in which cells are grown

If it is necessary to calculate the number of cells to be plated in a well of different format, then the formula to be used to maintain the same seeding density (SD) will be:

where N = total number of cells

A1 = area (expressed in cm 2 ) of well in which cells are grown at the moment

A2 = area (expressed in cm 2 ) of the well in which I want to grow the cells (in the future) so that the same cell density is maintained.


Optimization of Cell Seeding Density, EDTA Concentration, and Cell Density during Analysis Results in a 12-Fold Increase in Single-Cell Retrieval

The first goal in the development of this HTFC protocol was to find a strategy to optimize reproducible cell retrieval, using the adherent GIMEN neuroblastoma cell line. Initially, we adapted the cell detachment protocol of Kaur and Esau to a 384-well format 10 but were unable to achieve sufficient and reproducible cell retrieval (Fig. 1A, before optimization).

First, cell seeding density was evaluated by seeding increasing numbers of cells per well. As expected, cell retrieval markedly improved when more cells were plated (Fig. 1B). However, reproducibility of cell retrieval decreased when seeding density exceeded 5,000 cells/well, as observed by an increase in SD. Based on these data, it was concluded that a cell seeding density of 4,500 cells/well was optimal. Second, microscopic evaluation of the cell suspensions after different lengths of incubation periods with increasing EDTA concentrations revealed a minimum incubation time of 45 min and a minimum EDTA concentration of 3 mM (data not shown). Further increase in the EDTA concentration to 3.2 and 3.4 mM did not result in further improvement of cell retrieval (Fig. 1C). Finally, we assessed the effect of sample dilution prior to flow cytometric analysis. Dilution of the samples with 50 μL PBS supplemented with 2% FCS + 2 μM EDTA resulted in a 3.8-fold increase in cell retrieval (Fig. 1D). Optimal cell seeding density, EDTA cell detachment concentration and incubation times, and final cell suspension density of the flow cytometry sample per well resulted in an over 12-fold increase in single-cell retrieval (Fig. 1A). More than 90% of the nondebris cell population were single cells and flow rate was constant. 7-AAD and TMRM staining confirmed the cells were alive (Supporting Information S1).

The reproducibility of the cell numbers retrieved with the protocol can be concluded from a HTFC compound screen we have performed utilizing this protocol, in which over 10,000 wells were analyzed with an average retrieval of 1,600 (±SD 294) alive single cells/well. This corresponds to 74% alive single cell retrieval when corrected for sampled volume.

Optimized Antibody Staining Allows for (Multiplexed) Staining with Minimal Nonspecific Background Signal

Elimination of all washing steps from the HTFC protocol contributes to the HT nature of the protocol by decreasing the labor intensity, while minimizing cell loss inherent to washing. On the other hand, elimination of these steps also clearly indicates the need for antibody concentration titration. TNF-α and IFN-γ are involved in the upregulation of cell surface protein expression of MHC-I 11 , CD54 (ICAM-1) ( 13, 14 ), and CD274 (PD-L1) 15 in several tumor cell types. Effects of these cytokines on protein expression have been validated for all utilized cell lines using conventional flow cytometry (shown for GIMEN Fig. 2A and Supporting Information S2).

Decreasing the antibody concentration caused a clear decrease in background staining in medium-treated HLA-ABC stained MHC-I lacking GIMEN cells (Fig. 2B). However, a marked decrease in the stain index between untreated and TNF-α + IFN-γ treated samples was observed when decreasing the antibody concentration (Fig. 2C). This indicated a delicate balance between nonspecific background staining and discriminative ability of the antibody staining. Titration showed that HLA-ABC antibody concentration was optimal at a concentration of 8 ng/well (final concentration of 0.27 ng/μL). The HTFC staining protocol allows for distinct discrimination between HLA-ABC expression in untreated versus TNF-α-treated (Z = 49) and versus TNF-α + IFN-γ treated cells (Z = 94) (n = 8 per group), comparable with results from a typical conventional staining protocol (Fig. 2A). The same effect is observed for the other utilized antibodies (Supporting Information S2).

A unique aspect of flow cytometry is the opportunity to multiplex expression analysis of proteins on/in the same cell. Combining this aspect with HTS allows for an opportunity to increase the possibilities and informativity of HTS analysis. Combining antibody staining of HLA-ABC with the nucleic acid dye 7-AAD and antibody staining against PD-L1 and ICAM-1 revealed no noticeable differences in individual staining efficacy of the HTFC protocol, as shown for HLA-ABC antibody staining (Fig. 3A).

We previously generated a GIMEN NFkB reporter cell line and found that TNF-α upregulates MHC-I expression in an NFkB-dependent manner, whereas IFN-γ-induced MHC-I upregulation is independent of NFkB 11 . Utilizing our HTFC protocol combining HLA-ABC antibody staining with evaluation of the intrinsic NFkB reporter expression confirmed NFkB (in)dependency of the observed MHC-I upregulations (Fig. 3B). This shows that the protocol is also suitable to study the effects of compounds on intracellular transcription factors using reporter cell lines.

The Staining Protocol is Translatable to Multiple Cell Lines without any Modifications

The optimized HTFC protocol was subsequently performed using five additional cell lines, including breast cancer, cervical cancer, hepatocellular cancer, and human embryonal kidney cells. Retrieved single cell counts were lower (HepG2), comparable (MCF-7, SKBR3), or superior (HeLa, HEK293T) to the counts obtained in GIMEN neuroblastoma cells (Supporting Information S3). Individual and multiplex HTFC staining was performed and validated with conventional flow cytometry staining (data not shown). The HepG2 and MCF-7 cell lines were selected based on their trypsinization resistant nature. The MCF-7 line shows sufficiently high and reproducible cell retrieval, even though we do observe more doublets, which were excluded from analysis (Supporting Information S3A). In contrast, the slower growing, clumping HepG2 cells showed decreased cell counts, with an average single cell retrieval of 814 (n = 44, two individual experiments), but the counts were still sufficiently high and reproducible for reliable results (Supporting Information S3E). This indicates the potential to translate the HTFC protocol universally to other adherent cell lines of interest, contributing to the versatility of the protocol.


The assay with which we started this study failed to demonstrate any consistent effects of glutaminase inhibition on either glutamine metabolism or proliferation in these two cancer cell lines: addition of the inhibitor to cells (A549) known to be sensitive to the inhibitor, had no apparent effect on their proliferation (Supplementary Figure S2a). Conversely, the glutamine metabolism by cells (H358) that are insensitive to the same inhibitor was reduced to a much greater extent than that by sensitive (A549) cells (Fig. 2). We then demonstrated that these inconsistencies were artifacts, for one, because most glutamine had been depleted in the pre-incubation period (Figs 2b and 3b) leaving too little glutamine for effects of the inhibitor to become statistically noteworthy.

With our improved assay conditions we were able to show that the glutaminase inhibitor (GLS1i) does have an effect on glutamine metabolism of both cell lines, but that only the proliferation of the A549 cells is reduced (Fig. 5). For inhibitors of intracellular targets there is always the question of whether these are taken up by cells. Our inhibitor led to an increase in intracellular glutamine and a decrease in intracellular glutamine (Fig. 6a and b), suggesting that the compound entered the cells and affected its target: the glutaminase enzyme.

That the inhibitor used in this paper has been designed so as to be an inhibitor of glutaminase, working at a concentration of 1 micromolar, is not of much importance: it suffices to know that the inhibitor affects cell proliferation. Indeed, we expect that inhibitors of cell proliferation that work through any target should be compromised by the changes in the cells’ environment during the assay, if the conditions are not designed to keep these to a minimum. We used an inhibitor that is of importance to the pharmaceutical industry, because this pinpoints the likely implications of our findings. To be able to do this, we have had to accept that we cannot now disclose the chemical identity of the inhibitor we used.

Our findings highlight the importance of in vitro assay optimization for the assessment of the potential of metabolic, and probably also other, inhibitors as anti-cancer drugs that impact on cellular metabolism. Variability in the metabolic state during the assay may well create false positives and false negatives because intermediary and energy metabolism is full of pleiotropic implications. Importantly perhaps, the implications of our findings are unlikely to be limited to studies of metabolic inhibitors. Our observation that cell proliferation assays for drug effectiveness are readily compromised because the metabolic conditions around the cells are changing during these assays should not only impact assays for drugs that act on metabolism. Assays of drugs targeting other aspects of cell biology such as cell division or cell anatomy will also be compromised when cells are dying or changing metabolic state during the assay: changes in metabolic state, e.g. in ATP levels, will affect the dependence of viability on almost any drug. This touches on two important concepts: (i) the various ‘layers’ in cell function are strongly connected, such that one should always consider the other layers when one manipulates only one layer and (ii) much of the ‘irreproducibility’ of science that is being reported is due not to any experimental mistakes, but due to failure to sufficiently control for these strong interactions, in particular feedback via effects on the wider physiological state of the system under study.

Other inhibitors, such as those of cell signaling or transcription require even longer cell incubations, and may therefore be compromised even more by changes in the levels of metabolites such as ATP, NADH, acetyl-CoA and glutamate that cross-talk widely. Even though metabolism may not be the drug target in many cases, its perturbation due to inappropriate culture conditions, might produce a false response. And since the drug may well affect metabolism indirectly, the impact of metabolic status could be overlooked in both control and treated conditions. Our results suggest that cell assays are even more readily compromised by metabolic changes when the cells are metabolically very active. Naturally, this spells even more trouble for cell types that are relevant for cancer. For robust cell line assays it may be justified that new high-throughput technologies are developed that maintain the extracellular milieu constant over time. Microfluidic technologies may prove suitable for this.

Of course, as highlighted by Haanstra et al. 35 finding drugs that robustly kill cancer cells is not necessarily a good idea. First, all drug actions should be concentration dependent and, second, drugs will also affect normal (non-cancer) cells. In this sense, no drug will kill cancer cells at concentrations that do not affect normal cells 36 . Good drugs are those that kill cancer cells at lower drug concentrations than normal cells. Therefore, assays of anti-cancer drugs need to be both precise and robust.

Perhaps even more so than this, our results should warn against the straightforward implementation of historically-fixed sets of conditions for drug assays in cell lines. Living cells are complex enough to engage in all sorts of metabolic changes, these changes may well differ between individual cell lines, and the metabolome is sensitive to such changes well before the metabolic fluxes produced by the cells are 37 . We therefore advocate that reports on drug assays are accompanied by a thorough description of the experimental procedure used as well as by a metabolic characterization of the cells during the assay, such as in the workflow we demonstrated here. After all, such characterization has become possible over recent years.

Indeed, reports in the literature regarding the characteristics of cell lines under basal and perturbed conditions may have overlooked changes in the metabolic environment of cells or high cell density. Such aspects are typically not reported or measured and may contribute to the irreproducibility of the results when repeating the assay in a different laboratory. Such irreproducibility is fueling the reproducibility debate 1,2 .

Not only do our results highlight the need for reporting experimental ‘details’ concerning culture conditions, they also show a way towards rationalizing and standardizing these. Required details would include, but not be limited to, information on the source of cell lines and passage number (or at least whether all cells used were below a certain passage number), number of cells per well at the time of seeding and throughout the assay, density/confluence throughout the assay, volume of culture media used, details of cell culture flasks used, length of assay (from time of seeding), choice of cell culture medium (including concentrations of all components) and sera (concentration used and source), and how concentrations of key nutrients (e.g. glucose and glutamine) and pH change throughout the assay. This would complement existing efforts for standardization across biomedical research 38,39,40 and improve reproducibility, transparency and evaluation of the experimental data, points that are of critical concern 41 .

A number of reporting guidelines for the results of biological assays have been in existence for some time, e.g. the Minimum Information About a Microarray Experiment (MIAME) standard. MIAME is now a reporting requirement for a number of funding agencies and journals 42 . Similarly, there are now minimum reporting standards in use for metabolomics 39 , proteomics 40 and systems biology models 43 . Since 2008, the Minimum Information for Biological and Biomedical Investigations (MIBBI) project has acted as a repository for the many minimum reporting guidelines that have since been created there are now over 40 for the biological and biomedical sciences (, accessed 04 July 2016).

The assay developed and discussed here, as well as the contention that it should be accompanied by metabolic analyses of the assay cell lines, should contribute to improved assay reproducibility in cell biology and drug discovery.

Practical advice for seeding the cells

Start preparations one week in advance. For HEK293 cells, 100% confluency is approximately 10–12 million/75cm 2 flask. Seeding Suggestion: Seed 1 million cells (from passage) in 15 ml DMEM + 10% FBS on a Friday so as to have 90% confluency after the weekend.

On Monday trypsinize cells and dilute to a density of 1 million cells/ml. Set up four flasks of cell culture:

  • The first flask is used to test cell densities with Forskolin and to determine the best concentration of Forskolin. This flask is going to be used on Tuesday. Seed a sufficient number of cells in order to reach 80-90 % confluency overnight. This can be achieved by adding 4ml of HEK293 cells at a density of 1 million cells/ml to 11 ml DMEM + 10% FBS.
  • The second flask will be used for the actual experiment (treatment) with the agonist and antagonist treatments. This flask is going to be used on Wednesday. Add 2.5ml of 1 million cells/ml of HEK293 cells into 12.5 ml DMEM + 10% FBS.
  • The third flask is a reserve flask. Mix 1.5 ml of 1million cells/ml HEK293 cells into 13.5 ml DMEM + 10% FBS.
  • The forth flask is used for passage of cells on Friday for the following Monday. Add 0.75ml of cells at a density of 1 million cells/ml to 14.25 ml DMEM + 10% FBS.

This protocol describes the production of hepatocyte-like cells (HLCs) from human pluripotent stem cells and how to induce hepatic steatosis, a condition characterized by intracellular lipid accumulation. Following differentiation to an HLC phenotype, intracellular lipid accumulation is induced with a steatosis induction cocktail, allowing the user to examine the cellular processes that underpin hepatic steatosis. Furthermore, the renewable nature of our system, on a defined genetic background, permits in-depth mechanistic analysis, which may facilitate therapeutic target identification in the future.

For complete details on the use and execution of this protocol, please refer to Sinton et al. (2021).

Sample Preparation for Fluorescence Microscopy: An Introduction - Concepts and Tips for Better Fixed Sample Imaging Results

Multiple processing steps are required to prepare tissue culture cells for fluorescence microscopy. Experiments are generally classified as being either live or fixed cell microscopy. Fluorescence microscopy of live cells uses either genetically encoded fluorescent proteins (e.g. GFP, mcherry, YFP, RFP, etc.) or cell membrane-permeable, non-toxic fluorescent stains. Fluorescence microscopy of fixed cells uses a fixative agent that renders the cells dead, but maintains cellular structure, allowing the use of specific antibodies and dyes to investigate cell morphology and structure. Appropriate sample preparation is necessary to ensure high quality images are captured. Here we describe a number of concepts and considerations regarding the sample preparation process that can assist with automated digital fluorescence microscopy of fixed cells.

Cell Fixation

The goal of fixation is to maintain cellular structure as much as possible to that of the native or unfixed state during the processing steps and subsequent imaging. There are a number of fixation methods suitable for fluorescence microscopy that fall into two basic categories: aldehyde fixatives and alcohol fixatives. Organic solvents such as alcohols and acetone remove lipids and dehydrate the cells, while precipitating the proteins on the cellular architecture. Cross-linking aldehyde reagents form intermolecular bridges, normally through free amino groups, creating a network of linked antigens. Cross-linkers preserve cell structure better than organic solvents, but may reduce the antigenicity of some cell components, and require a permeabilization step to allow the antibody access to the specimen. Fixation with both methods may denature protein antigens, and for this reason, antibodies prepared against denatured proteins may be more useful for cell staining. As each method has its advantages and disadvantages and because different fixative methods can destroy or mask different epitopes, several different methods may need to be tested for optimal results.

Aldehyde fixatives cross link proteins and generally do an excellent job of preserving cell morphology. While they are generally slower acting than organic based fixatives, 4% formaldehyde for 10 minutes at room temperature is a good starting point. Glutaraldehyde will also cross link proteins, but results in significant autofluorescence and generally should be avoided or used in low concentration in conjunction with formaldehyde. Note that glutaraldehyde is the preferred fixative of choice for electron microscopy.

Formaldehyde vs. Paraformaldehyde vs. Formalin

Paraformaldehyde is a polymer of formaldehyde. It exists as a dry powder suitable for storage, and needs to be broken down into its monomeric component formaldehyde. This is accomplished by heating under basic conditions until it becomes solubilized. Commercial formaldehyde solutions generally referred to as formalin, often contain methanol to prevent re-polymerization into paraformaldehyde. A saturated formalin solution has formaldehyde content in water or buffer of 37- 40% and is sometimes referred to as &ldquo100% formalin&rdquo. Therefore using 10% formalin (1:10 dilution) is equal to using 3.7-4% formaldehyde for fixation. Since &ldquo100% formalin&rdquo contains up to 15% of methanol as a stabilizer, it can have a significant impact on fixation. This is mainly the result of membrane permeabilization by methanol, which results in the interference in the staining of membrane bound proteins.

Prepare a fresh 4% formaldehyde solution by dissolving paraformaldehyde powder into PBS (PH 7.4) using a stirring hot plate with the heater set to a medium setting until the liquid reaches approximately 60 °C. As the paraformaldehyde breaks down to formaldehyde it will dissolve. Once the solution is completely dissolved (approx 30 min), the solution is quickly chilled on ice back to room temperature prior to use. Typically, tissue culture cells are fixed for 10 minutes at room temperature with 4% paraformaldehyde in PBS followed by 2-3 washes with PBS to remove excess formaldehyde and stop the fixing reaction.

Organic solvents such as methanol rapidly precipitate proteins, maintaining structure when doing so. While the cellular protein cytoskeletal structure is maintained with methanol fixation, small molecules within will be lost during the subsequent processing steps because they are not precipitated. Because methanol precipitates proteins never use this method if genetically encoded fluorescent proteins are to be imaged or detected - activity from these proteins will be abolished with methanol fixation. Methanol is used cold (-20 °C) for 10-20 minutes. Using a combination of methanol and acetone (1:1) can sometimes improve results. Methanol is best for preserving structure while acetone improves permeabilization. Following fixation samples are washed with PBS 2-3 times to remove alcohol and rehydrate the specimen.

Another important considerationof a fixation protocol is the buffer selection. Ideally the buffer is isotonic in nature so as to not disrupt cellular structure, while maintaining a pH as close to physiologic as possible. The most frequently used buffer is phosphate buffered saline (PBS). When using aldehydes as fixatives avoid using amine-containing buffers, such as Tris, as they will react with the fixative.


Fixation with the use of cross linking agents does not effectively diminish cell membrane structure. Therefore, formaldehyde fixation requires the cells to be permeabilized to allow antibodies access into the interior of cells. The large size and ionic nature of antibody proteins precludes them from gaining access without membrane disruption (Figure 1-2). Alcohol and acetone based fixation generally does not require this step.

Digitonin, Leucoperm and Saponin are relatively mild detergents that generate large enough pores for antibodies to pass through without completely dissolving the plasma membrane. These are suitable for antigens in the cytoplasm or the cytoplasmic face of the plasma membrane (Figure 1). Saponin acts by dissolving cholesterol present in the plasma membrane. When used at low concentrations, internal membranes remain intact so it is useful for labeling smaller molecules that exist in a soluble state within the cytoplasm. It should be prepared as a stock in DMSO, and is typically used at 0.5 to 1 mg/mL at room temperature for 10 to 30 minutes.

Triton X-100 is probably the most commonly used permeabilization agent for immunofluorescent staining. This detergent efficiently dissolves cellular membranes without disturbing protein-protein interactions (Figure 1). Triton is usually used at concentrations ranging from 0.1 to 1% for permeabilization. We typically use Triton X-100 or NP-40 at 0.1% in PBS at room temperature for 10 minutes. Besides cell membrane permeabilization, these detergents will partially dissolve the nuclear membrane making them suitable for nuclear antigen staining.

Figure 1. Antibody Accessibility with Unpermeabilized and Digitonin or Triton X-100 Permeabilized Cells. Antibodies can only access the exterior of aldehyde fixed unpermeabilized cells, while mild agents such as digitonin will permeabilize the plasma membrane allowing access to cytoplasmic interior, but not interior membrane bound organelles such as the nucleus of mitochondria. Stronger nonionic detergents, such as Triton X-100, permeabilize both the plasma membrane and interior membranes, allowing full access while still preserving cell structure.

SDS is an anionic detergent that is commonly used to denature proteins and provide them with a large negative charge for electrophoresis. It is useful as a permeabilizing agent to induce slight denaturation of fixed cells in order to reveal epitopes which may normally be masked from an antibody. It may extract small, poorly cross-linked proteins from fixed specimens, and should not be used on samples fixed by precipitation (e.g. methanol).

When using antibodies to stain cellular objects in specimens, it is necessary to &ldquoblock&rdquo the sample in order to reduce non-specific binding. Non-specific binding may occur for several reasons: unreacted fixative aldehydes may crosslink antibodies to inappropriate structures structures within the samples may &lsquotrap&rsquo antibodies or, if using polyclonal antibodies, low affinity IgG molecules may bind to inappropriate structures. These potential issues may be prevented by treating the specimens with a protein solution that will compete for non-specific binding sites prior to staining with antibodies. Commonly used blocking agents are bovine serum albumin (BSA), casein (or a solution of non-fat dry milk), gelatin, or normal serum obtained from the species of animal in which the secondary antibodies are made. Avoid using serum of the same species as the primary antibodies. Typically, the protein solutions are used at concentrations of 1 to 10% in buffer and the samples are treated after permeabilization for 10 to 30 minutes. It is also advisable to include a small amount of detergent (if the sample is to be permeabilized) to compete for hydrophobic interactions with the antibodies. In this case, low (around 0.1%) concentrations of Triton X-100 should be used, for example 3% BSA in PBS with 0.1% Triton X-100 or 5% fetal calf serum (FBS) in PBS with 0.1% Triton X-100 for 30 minutes at room temperature works well.

Antibodies for immunofluorescence are divided into two categories: primary and secondary antibodies. These two groups are classified based on whether they bind to antigens or proteins directly or target another (primary) antibody that, in turn, is bound to an antigen or protein. Primary antibodies can be labeled or unlabeled. Primary antibodies that have been covalently linked with a fluorescent moiety will have specificity and bind directly to the target as well as the tag necessary for imaging. This is known as direct immunofluorescence (Figure 2). By coupling the primary antibody with a fluorophore, direct immunofluorescence is faster than the indirect version because time-consuming washing and incubation steps are omitted. Thus, direct immunofluorescence is easier to handle and therefore suitable for the rapid analysis of samples in standardized experiments, for example in clinical practice. Unlabeled primary antibodies on the other hand will bind to the target but require a secondary antibody that will specifically recognize the antibody. This situation is called indirect immunofluorescence.

Figure 2. Directly labeled antibody fluorescence on unpermeabilized SK-BR-3 Cells. Anti-EGFR antibody covalently labeled with FITC and incubated with unpermeabilized SK-BR-3 cells counter stained with Hoechst 33342. Cells were imaged with a Cytation&trade 3 Cell Imaging Multi-Mode Reader (BioTek Instruments). Green fluorescence is located only on outer cell periphery as EGFR is a plasma membrane receptor.

Primary antibody binding to the specific antigen involves the Fab domain of the antibody, which in turn leaves the Fc domain exposed (Figure 3). The secondary antibody&rsquos Fab has been developed to recognize the Fc domain of the primary antibody. Since antibody&rsquos Fc domain is constant within the same animal class, only one type of secondary antibody is required to bind to many types of primary antibodies. This reduces the cost by labeling only one type of secondary antibody, rather than labeling various types of primary antibodies.

Figure 3. Structure of IgG Protein.

Primary antibodies can be either polyclonal or monoclonal in nature. Polyclonal antibodies are a heterogeneous mixture of antibodies directed against various epitopes of the same antigen. As they are generated from different B-cell clones they are immunochemically dissimilar and can have different specificities and affinities. While a number of different animal species, including goat, swine, guinea pig and cow, are routinely used to produce polyclonal antibodies, rabbits are most frequently used. Monoclonal antibodies are a homogeneous population of immunoglobulin directed against a single epitope. These antibodies are generated from a single B-cell clone isolated from a single animal and as such have the same Fab structure. While the vast majority of monoclonal antibodies are produced from mice, increasingly, they are being produced from rabbits as well.

Binding affinity and the resultant titer towards its antigen is different for every antibody. The use of monoclonal antibodies removes much of the variability, but batch to batch concentrations can vary even with the same monoclonal due to aggregation or denaturation. As a result the antibody concentration required and the incubation time used can vary lot to lot. Optimal assay development requires that the primary antibody be titrated for best results. Too little antibody results in an artificially low signal, due to a lack of target saturation. Too much antibody can result in nonspecific binding despite have used a blocking step. Because of the reagent cost, it is paramount that only as much antibody as required be used. However, for a one-off experiment, or a small series of experiments the time, effort and expense required for a full antibody titration is often not warranted. Typically, a dilution of 1:1000 for the commercially available monoclonal antibodies is used. Because of the low amounts of protein in the sample, it is best to use a buffer containing BSA to dilute the antibody, for example, PBS 0.1% Triton X-100 supplemented with 30 mg/mL of BSA. This can be made in bulk, filter sterilized and aliquoted. Use only enough antibody dilution to cover the sample, - for a typical well in a 96-well microplate this is 20-25 &muL. Most of the commercially available antibodies have sufficient titer to bind high copy number targets in 30-60 minutes at room temperature. Results with low copy number targets often are improved by overnight incubation at 4 °C, and shaking can help as well. An incubation time of 60 minutes at room temperature works well for most experiments, but an overnight incubation can be used as a convenient end of the day stopping point for sample processing even with high target number epitopes.

Figure 4. Direct vs. Indirect Antibody Labeling.

A secondary antibody is required with indirect immunofluorescence, where the primary antibody provides no means of detection and works by binding to a primary antibody, which directly binds to the target antigen. The use of secondary antibodies to indirectly detect target antigens requires additional process steps but can also offer significant advantages over labeled primary antibodies. Secondary antibodies increase the sensitivity through the signal amplification that occurs as multiple secondary antibodies bind to a single primary antibody (Figure 4). In addition, a given secondary antibody can be used with any primary antibody of the same isotype and target species, making it a more versatile reagent than individually labeled primary antibodies. This flexibility is a significant advantage of indirect immunofluorescence. Because the vast majority of primary antibodies are produced in just a few host animal species and most are of the IgG class, it is economical to produce and supply ready-to-use secondary antibodies for many methods and detection systems. From a relatively small number of secondary antibodies, many options are available for purity level, specificity and label type for a given application. A dilution of 1:500 for some commercially available secondary antibodies (Figure 5). As with primary antibodies, the low amounts of protein in the reagent makes it imperative that a buffer containing BSA or other non-specific protein be used to dilute the antibody, for example PBS 0.1% Triton X-100 supplemented with 30 mg/mL of BSA can be used for the dilution of secondary antibodies.

Figure 5. Fixed and three color fixed and stained HeLa cells. Mitochondria identified by a mouse anti-mitofilin primary antibody followed with a Rabbit antimouse IgG monoclonal antibody labeled with Texas Red (Red). Nuclei and actin filaments are identified by DAPI (blue) and AlexaFluor®-488-phalloidin (green) counterstaining respectively. Scale bar indicates 30 &mum.

It is important that the secondary antibody employed in the experiment is raised against the host species used to generate the primary antibody. Polyclonal primary antibodies are usually raised in rabbit, goat, sheep or donkey and are generally IgG isotypes. The secondary antibody therefore, will typically be an anti-IgG antibody. Monoclonal primary antibodies are commonly raised in mouse, rabbit and rat. For example, if the primary monoclonal antibody is a mouse IgG, you will need an anti-mouse IgG.

The use of pre-absorbed secondary antibodies reduces the risk of cross reactivity with other antibodies and should be used with multi-color experiments when several primary antibodies and their corresponding secondary antibodies are used simultaneously.

Pre-adsorption is an additional processing step where the secondary antibodies are passed through a column matrix containing immobilized serum proteins from potentially cross reactive agents. For example, only antibodies specific to rabbit IgG will pass through a pre-absorbing column containing immobilized mouse, human and horse IgG whereas, antibodies cross reacting will bind and stay adsorbed to the matrix.

Between each process step, there is usually a wash step to remove unbound excess reagents. The buffer used for these steps should take into account the buffer used for the process steps. As described previously, it is best to use a buffer system that is isotonic to preserve cellular structure and close to physiological pH. Washing, as the name implies is a series of aspiration and buffer reagent additions performed in sequence. Depending on the volume used, 2-3 cycles is sufficient for washing, with the final step being an aspiration to remove the fluid prior to the addition of the next processing reagent. PBS is commonly used, and a surfactant such as 0.1% Triton X-100 or 0.05% Tween&trade 20 can also be added. Note that it is important to make these steps in a timely fashion to prevent the sample from drying out.

Counter Stains

Counter stains are used for two different purposes reduction of background fluorescence or the identification of cellular organelles. Blue dyes, such as Evans blue can be used to reduce non-specific background fluorescence with fluorescein labeled samples. Because these dyes can mask weakly fluorescent desired binding they are not recommended for routine use. More importantly is the use of counter stains to identify cellular organelles and provide information regarding signal localization.

Figure 6. NIH 3T3 cell nuclei stained with DAPI.

The nucleus is the predominately targeted cellular organelle and can be identified through the use of stains that bind DNA (Figure 6). DAPI (diamidino-2-phenylindole) is a nuclear counter stain useful with multicolor fluorescent experiments. Its blue fluorescence differs significantly from fluorescein (green) or Texas red fluorescent probes used for other structures. It fluoresces brightly upon selectively binding to the minor groove of double stranded DNA. Its selectivity for DNA allows efficient staining of nuclei with little background from the cytoplasm. Other nuclear stains include Hoechst 33342, which is cell permanent and can be used with live as well as fixed cells, and propidium iodide, long used as a nuclear marker for flow cytometry which fluoresces in the red range. Far red dyes such as Draq5 can also be used a nuclear counter stain.

Fluorescently labeled phalloidin can be used as a counter stain for actin (Figure 7). This cyclic peptide isolated from the death cap mushroom (Amanita phalloides) has a very high affinity towards F-actin filaments. Because of its relatively small size and chemical stability, fluorescent derivatives are enormously useful in identifying actin filaments with high density labeling of the cytoskeleton which provides a means for the assessment of cellular morphology. As they are commercially available in a multitude of different colors they can be multiplexed with several other cellular probes.

Fluorescently tagged wheat germ agglutinin (WGA) can be used to stain for cell membranes. WGA is a carbohydratebinding protein that selectively recognizes sialic acid and N-acetylglucosaminyl sugar residues which are predominantly found on plasma membrane proteins.

Unlike antibodies, counter stains usually do not have cross reactivity with one another and often can be combined into a single process step - the only limitation would be buffer compatibility. For example, DAPI (5-10 &mug/mL) and phalloidin (20 nM) counter stains can be combined in a mixture of PBS 0.1% Triton X-100 from concentrated stock solution, freshly made on the day of the experiment. A 10-minute incubation at room temperature is sufficient to adequately stain fixed tissue culture cells.

Figure 7. Fixed NIH 3T3 cells expressing GFP counterstained with DAPI and Texas Red labeled phalloidin. NIH3T3 cells cultured in 96-well plates were formaldehyde fixed then counter stained with DAPI and Texas red labeled phalloidin. Cells were imaged with a Cytation&trade 3 using configured with DAPI, GFP and Texas Red light cubes.

Fixation and staining of slides or microplates is a multistep process with a number of reagent additions, wash steps and incubations (Figure 8). The processing of slides or coverslips is generally a low throughput operation and can usually be accomplished with a manual operation. Both fixtures can be treated using small dishes or in the case of slides dedicated Coplin jars can be used to store reagents or as vessels for wash media. Microplates, particularly 96- and 384- well plates with cells are much more amenable to automation. Because microplates have standardized well size, spacing and footprint, automation can be used to provide reagent addition and washing (aspiration and dispense). The small size of coverslips also allows them to be placed in the wells of low-density plates (6- and 12-well) for processing prior to adhering them to a slide.

Figure 8. Automated Workflow for Cell Seeding, Fixation, Permeabilization and Three Color Staining Process. An EL406&trade Combination Washer Dispenser (BioTek Instruments) was used to carry out the process steps for cell fixation, permeabilization and staining with three colors (DAPI, Alexa-Fluor® 488 phalloidin stain, and Texas red labeled secondary antibody) are indicated in red.

Storage of fixed and stained specimens can vary depending on the sample. Specimens on slides (or coverslips) can be treated with a mounting medium and sealed for long term storage. Mounting medium helps preserve the sample and raises the refractive index, which will improve performance with oil objectives. Mounting medium often has added agents to scavenge free radicals and reduce photobleaching. Prolong Gold (Life Technologies) and Fluoromount&ndashG (SouthernBiotech) are examples of commercially available mounting media. It is important to rinse the slide/coverslip in distilled water prior to sealing as the salt from the PBS wash will precipitate out as crystals as it dries. After adhering the coverslip to the slide, allow the slide to dry overnight prior to sealing. Seal the edges of each coverslip with regular transparent nail polish and allow drying for 3 minutes cells are now ready for imaging. Slides should be stored in a light tight container to prevent photobleaching. This will provide semi-permanent preparations that will last months if kept in the dark.

Microplates storage is more challenging. The small well size and well depth makes it almost impossible to seal the sample through the use of a coverslip. Microplates are better stored wet at 4 °C - PBS with 0.1% sodium azide can be used as a preservative. Plates are sealed with an adhesive plate seal and the plates wrapped in aluminum foil. While they will lose resolution over time, clear images can be obtained after one month in storage.

Final Comments

There is no &ldquobest&rdquo procedure for immunostaining and counter staining fixed cells on slides or microplates. There are numerous methods to fix, permeabilize, and stain cells, each with strengths and weaknesses. Immunostaining intracellularly involves detecting small molecules on interior cell structures often using much larger molecules, such as antibodies. Structures need to be stabilized, while at the same time holes need to be opened in membranes large enough to allow antibodies access to the interior. Different antibodies have different affinities and will often recognize different epitopes on the same protein, while different fixative methods will expose or mask different epitopes within the cell. While there is no one-size-fit-all technique for immunostaining, there are a number of different techniques that can be tested in order to optimize results.


Cancer is a complex disease caused by multiple types of interactions. To simplify and normalize the assessment of drug effects, spheroid microenvironments have been utilized. Research models that involve agent measurement with the examination of clonogenic survival by monitoring culture process with image analysis have been developed for spheroid-based screening. Meanwhile, computer simulations using various models have enabled better predictions for phenomena in cancer. However, user-based parameters that are specific to a researcher’s own experimental conditions must be inputted. In order to bridge the gap between experimental and simulated conditions, we have developed an in silico analysis method with virtual three-dimensional embodiment computed using the researcher’s own samples. The present work focused on HeLa spheroid growth in soft agar culture, with spheroids being modeled in silico based on time-lapse images capturing spheroid growth. The spheroids in silico were optimized by adjusting the growth curves to those obtained from time-lapse images of spheroids and were then assigned virtual inner proliferative activity by using generations assigned to each cellular particle. The ratio and distribution of the virtual inner proliferative activities were confirmed to be similar to the proliferation zone ratio and histochemical profiles of HeLa spheroids, which were also consistent with those identified in an earlier study. We validated that time-lapse images of HeLa spheroids provided virtual inner proliferative activity for spheroids in vitro. The present work has achieved the first step toward an in silico analysis method using computational simulation based on a researcher’s own samples, helping to bridge the gap between experiment and simulation.

Citation: Minamikawa-Tachino R, Ogura K, Ito A, Nagayama K (2020) Time-lapse imaging of HeLa spheroids in soft agar culture provides virtual inner proliferative activity. PLoS ONE 15(4): e0231774.

Editor: Donald Gullberg, University of Bergen, NORWAY

Received: November 11, 2019 Accepted: March 31, 2020 Published: April 17, 2020

Copyright: © 2020 Minamikawa-Tachino et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the manuscript and its Supporting Information files.

Funding: This work was supported by JSPS KAKENHI Grant Number JP24650156.

Competing interests: The authors have no conflicts of interests to declare.


While current in vitro models do not fully recapitulate all aspects of the in vivo microenvironment, there is a strong body of literature indicating the physiologically relevant cues that guide tissue function it is believed inclusion of such cues in an in vitro model will improve its ability to predict drug effects on humans. For instance, fluid flow in vivo influences tissue function by providing perfusion at lower flow rates 20–23 and by additionally generating a flow-induced shear stress (FSS) 24–27 at higher flow rates, highlighting the importance of including the capability to generate a range of flow rates 9 in OOC models. In addition, the spatial organization and chemical environment of cells influence tissue function examples include heterotypic cell organization, 28 extra-cellular matrix (ECM) composition, 29 and spatial control of media composition and biochemical gradients to create distinct in vivo phenotypic zones. 30,31 Additionally, maintaining near physiologic concentrations of autocrine and paracrine signals requires a low media-to-cell ratio within in vitro culture systems. 32 OOCs that allow for integration of a diverse range of cell types, including those from various organs and with different qualification status, can increase the range of tissue model applications and cell sources. Incorporating these cues and features in an OOC enables selection of conditions to drive tissues toward physiologically accurate and, therefore, clinically predictive models.

The next generation of OOCs should improve upon standard in vitro models by integrating both new and established sensing modalities while maintaining compatibility with state-of-the-art data collection tools. Integrated sensors in OOCs can obtain real-time data over the course of an experiment 17,18,33–35 without disturbing the microenvironment. While large data sets can be generated in vitro via high-throughput screening tools such as high-content screening (HCS) and RNA sequencing (RNA-seq), current OOC systems only interrogate one or a few tissue culture conditions 32 in each experiment. We aimed to produce a next generation OOC system with on-line integrated sensing capabilities that maintains downstream compatibility with HCS and RNA-seq while enabling high-throughput interrogation of many tissue culture conditions simultaneously (or allows for a large number of replicates for a few culture conditions). Such a system will enable rapid parallel screening of many drug candidates and the generation of highly robust and complete data sets, including the ability to measure dynamic tissue behavior with previously inaccessible temporal resolution.

Here we present an advanced OOC platform technology that combines programmable flow control, integrated real-time sensors, and an ability to include heterotypic cell-type complexity within a high-throughput layout that readily interfaces with industry-standard infrastructure and state-of-the-art data collection tools. The platform comprises 96 individual OOC devices in one cell culture plate, up to 192 individual and actively-controlled micropumps contained within the plate lid, and 384 electrical contacts to make electrical measurements within each of the 96 devices. 36 Each OOC device contains two microchannels separated by a semi-permeable scaffold capable of supporting a range of cell types in mono- or co-culture to generate various tissue-relevant models. Fluid flow in each individual channel is controlled by a separate micropump with low fluid circuit dead volume to minimize dilution of cell-secreted factors. The platform employs optically clear thermoplastic materials chosen to minimize drug sorption and enable histological analysis using standard microscopy techniques. The format is compatible with existing life science tools, including liquid and plate handling tools and assays, such as multiplex immunoassays, HCS and RNA-seq. In this manuscript, we demonstrate the versatility and broad capabilities of the platform by presenting a series of proof-of-concept experiments comprised of data from human kidney, vascular, liver, and intestine tissue models with relevant physiological readouts.

The models demonstrate distinct organ/tissue functionality, which is enabled by the novel features of the platform technology. Specifically, we demonstrate (a) two physiologically relevant flow regimes: perfusion flow that enhances hepatic tissue function and high-shear stress flow that aligns endothelial monolayers (b) integrated electrical sensors that quantify barrier function of primary gut colon tissue in real-time (c) optical access to the tissues that enables direct quantification of renal active transport and oxygen consumption and (d) a format compatible with HCS and RNA-seq. These features, combined with the throughput capabilities, result in a platform that can both generate and interrogate microscale tissues and is expected to increase predictive accuracy of in vitro drug screening.


Background Information

Mitophagy is a type of selective autophagy event that specifically removes damaged and dysfunctional mitochondria caused by accumulation of mitochondrial DNA (mtDNA) mutations or misfolded proteins, loss of mitochondrial membrane potential, and production of reactive oxygen species (ROS). Seminal work by Narendra et al. ( 2010 ) discovered that PINK1 kinase, which is normally degraded via the N-end rule under normal conditions (Yamano & Youle, 2013 ), is stabilized and activated on damaged mitochondria. PINK1 activation triggers phosphorylation of both ubiquitin and Parkin, a E3-ligase, which promotes Parkin to translocate to damaged mitochondria and ubiquitinate many mitochondrial outer membrane proteins. Ubiquitin chains on mitochondrial proteins are bound by autophagy adaptors such as NDP52, OPTN, and TAX1BP1 to mitochondria, which in turn bind and recruit the general autophagy machinery to mitochondria to initiate the mitophagy.

Initially, the mitochondrial outer membrane protein Tom20 was used to monitor the clearance of mitochondria by confocal imaging or Western blot (Narendra, Tanaka, Suen, & Youle, 2008 ). Later on, it was revealed that many mitochondrial outer membrane proteins that are ubiquitinated by Parkin undergo ubiquitin/proteasome system (UPS)-mediated degradation rather than autophagic degradation (Chan et al., 2011 Sarraf et al., 2013 Tanaka et al., 2010 ), making them unsuitable for monitoring mitophagy.

One of the widely used mitophagy reporters, mito-QC, which mimics the GFP-RFP-LC3 for autophagy flux, was created by fusing mCherry-GFP in tandem with the C-terminal tail (101-152) of the outer membrane protein Fis1 (Allen, Toth, James, & Ganley, 2013 ). Under normal conditions, mito-QC fluoresces both red and green to appear yellow together. Upon induction of mitophagy, mitochondria are engulfed by autophagsosomes, which then fuse with lysosomes to form autolysosomes. In this acidic environment, the GFP signal gets quenched but mCherry signal remains stable to make mito-QC appear to be red-only. The percentage of red-only mito-QC signal can be used to measure mitophagy. As mentioned above, mito-QC itself can be degraded at least partially via the UPS and may result in an underestimation of the mitophagy events. In addition, factors involved in the UPS pathway may inadvertently influence mito-QC readings.

mKeima belongs to a class of fluorescent proteins with large Stokes shifts (the difference in wavelength between the peak excitation and emission spectra). It was originally identified from the stony coral Montipora sp. with degenerate primers and semi-random mutagenesis (Kogure et al., 2006 ). The first identified clone was a violet-colored GFP-like chromoprotein that did not fluoresce. However, a red fluorescent protein was successfully created with five substitutions (H94N, N142S, N157D, K202R and F206S) and the addition of a valine residue at the second amino-acid position. Subsequently, four more mutations (S61F, I92T, F158Y, and S213A) were introduced to generate the Keima protein, a shogi (Japanese chess) piece that hops in the manner of the knight in chess, owing to the large Stokes shift. Keima shows a bimodal excitation spectrum peaking at 440 and 580 nm and emission spectrum peaking at 606 nm. This Keima is homotetrameric, which was further converted into a dimeric dKeima with introduction of V123T and V191I mutations. Additional mutations (L60Q, F61L, V79F, T92S, T123E, Y188R, and Y190E) led to the monomeric mKeima protein, which now shows maximal emission at 620 nm. The two excitation peaks somehow correspond to the neutral and acidic pH. Violot, Carpentier, Blanchoin, and Bourgeois ( 2009 ) used X-ray crystallography to discover that the large Stokes shift of mKeima is due to the reverse pH-dependence of chromophore protonation.

Katayama, Kogure, Mizushima, Yoshimori, and Miiyawaki ( 2011 ) first reported the use of mKeima and mito-mKeima for autophagy and mitophagy measurements. They showed that mKeima has a pKa value of 6.5 and excitation peak ratio (586/440) of more than 6. Unlike mito-QC, mito-mKeima can be used together with GFP/YFP tagged proteins. By measuring the proportion of the high ratio (550/438) signal area to the total mitochondria area as an index of mitophagy (IM) in 30 transfected cells, IM of 60% can be detected in MEF cells treated with CCCP + oligomycin. There is an obvious limitation with this imaging-based method as only 30 cells are examined. To improve the speed and sensitivity, we explored the possibility of using flow cytometry (FACS) to quantitatively and quickly measure mitophagy with mito-mKeima (Lazarou et al., 2015 ). We successfully detected as low as 5% mitophagy events when PINK1 was artificially tethered on mitochondria via chemically-induced dimerization, which is crucial to the newly proposed model: PINK1-generated phosphor-ubiquitin itself serves as the mitophagy signal and Parkin is a mere amplifier of this signal. By using FACS to detect mitophagy, over 50,000-100,000 events can be collected and analyzed, which increases the detection threshold of mitophagic events for better data representation.

Critical Parameters

For viral packaging, the most critical factor is the health and behavior of 293T cells, as transfection efficiency is rarely an issue for them. 293T cells must be split regularly, preferably at 1:5 ratio and with increased attention on how they attach and spread. Right after thawing, especially when seeded at a low density, 293T cells tend to form clumps.

We tried several different transfection reagents and did not observe any difference in virus titer when similar transfection efficiency was obtained. Occasionally, we encountered an issue of low virus titer and traced the problem back to the helper plasmids. Some of the helper plasmids have the potential to rearrange, causing small deletions, which can be detected by agarose gel electrophoresis even without digestion. We solved the problem by re-transforming the correct plasmids in NEB or Thermofisher's Stable competent cells. Therefore, it is very important to retain original helper plasmid stocks either in bacteria or in solution as backup for future amplification.

In terms of transduction or infection, forgetting to add polybrene is the most likely cause of low transduction rate, followed closely by over-seeding too many cells for infection. Freeze-thaw cycles of viruses can lead to 5-fold reduction in virus titer. We found virus can be stored for at least 1 month at 4°C without losing significant titer. Otherwise, virus should be divided into aliquots and kept for long-term storage at −80°C. When thawing the virus, the frozen virus should be incubated for a few minutes at 37°C and used quickly. Furthermore, the addition of too much virus with a high titer could be toxic. Sometimes, culture medium for target cells differ from medium for 293T cells. In this case, virus volume should be at least less than half of the final medium volume for infection.

For mitophagy assays, the most crucial element is to have proper controls. With FACS analysis, DAPI, or DRAQ7 (far-red) should be used to gate out the dead cells. When gating either the mKeima or YFP channel, compensation might be needed to avoid interference between each channel. This is often the issue when either mKeima or YFP signal is too high. Reducing laser power could help to limit the interference but sometimes still does not solve the problem. Therefore, it is always better to do a pilot experiment to find out the optimal virus titer for infection. It is better to begin with a weak signal, as re-infection can be done to enhance the signal. For the final gating of mKeima ratio (pH4, 561 nm/pH7 488 nm), always re-normalize the untreated control of each sample set to around 1%. This is because different cell lines may have different basal mitophagy level to begin with. In addition, mKeima and YFP intensity are not identical in different cell lines even after sorting to obtain similar intensity level. Therefore, applying the same mKeima ratio gating to different cell lines is not practical and also not advised.


For possible problems and their solutions, see Table 1.

Problem Solution
Low virus titer Make sure 293T cells never grow confluent, split when they reach 80%-90% confluence. Sometimes, splitting cells several times after thawing may help, too.
Seed cells so that they are 70%-90% confluent the day before transfection. Make sure cells adhere well and spread nicely. Do not attempt transfection if cells are clumpy.
293T cells do not attach well, so be gentle when moving plates in and out the incubator. After transfection mix is added to the cells, swirl the plates very gently a couple of times to mix so that cells do not get dispersed and come off the plates.
Check the size of helper plasmids. Some helper plasmids tend to undergo recombination and small deletions might occur, which can be detected on agarose gels. Transform the helper plasmids in NEB's Stable cells (NEB, cat. no. C3040I) or similar competent cells defective in recombination for mini or midi prep.
Low transduction rate Seed cells so that they are less than 20% confluent on the day of infection. 5%-10% confluent could be used if virus titer is not high.
Make sure polybrene is added during infection. 4-8 µg/ml works well for most cell lines. Some cell lines are sensitive and lower polybrene concentration has to be used if cells die quickly after infection.
Use more virus.
mito-mKeima or YFP-Parkin signal is not even Retrovirus seems to produce more even gene expression than lentivirus. If an even signal is needed, FACS sort the cells
Mitophagy in wild type cells is not robust Different cell lines may require different concentrations of oligomycin, antimycin A for robust mitophagy response. Treatment time can also be a factor. Make sure fresh chemicals are used since they may lose activity over time or not properly stored.
Parkin expression is too low. While robust mitophagy does not require maximal Parkin expression, detectable Parkin expression level is indeed needed. It is ideal to use YFP-Parkin so that Parkin expression level can be readily monitored. Untagged or HA-Parkin can be used when YFP-Parkin is not an option.

Understanding Results

Anticipated results are shown in Figure 5. In untreated wild-type HeLa cells, mitophagy (shown as the upper box) is gated to around 1%. With OAQ treatment for 24 hr, 86.1% mitophagy is detected. In WIPI2 KO cells, mitophagy is gated to around 1% in the untreated cells, only 1.68% mitophagy is observed after 24 hr of OAQ treatment, indicating that WIPI2 KO completely blocks Parkin-mediated mitophagy. In ATG16L1 KO cells, 21.5% mitophagy is seen with OAQ treatment for 24 hr, comparing to 1.18% in untreated cells, suggesting that ATG16L1 strongly inhibits Parkin-mediated mitophagy. Different cell lines can have different cell size and mKeima signal, which reflects the different distribution patterns of cell population. When the triangle box is drawn to delineate the upper portion (mitophagy), make sure the diagonal line is parallel to the red/orange area (peak of the cell population) in the untreated control cells. See the video for details (Video 2).

Time Considerations

It takes 3-4 days to package virus from plating 293T cells for transfection to harvest virus on Day 2 and Day 3. For transduction/infection, 2-3 days are also required from seeding target cells to see optimal gene expression. For FACS analysis or confocal imaging, depending on the treatment time, 2-3 days should be planned.


I thank all members of Youle lab for critical comments, Dragan Maric for guidance on FACS analysis from NINDS Flow Cytometry Core Facilities. This work was supported by the Intramural Research Program of the NINDS, NIH.

Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Watch the video: MSU Cell Biology BIO320 - Cell Culture: Cell Seeding onto 96-Well Plate Video 9 (May 2022).