Image-independent evaluation of FluoroSpot Assays using Thermo Scientific fluorescence microplate readers
The enzyme-linked immunospot (ELISpot) assay is an elegant ex vivo recreation of the ELISA assay. This assay can be used in the research fields of cancer, infectious disease, and immunology for basic and translational applications . ELISpot assays can be applied for quantification of antibody-secreting B cells, or cells secreting protein antigens (e.g., T cells secreting cytokines, or glial cells or macrophages secreting growth factors). A remarkable feature of the ELISpot assay is its sensitivity. It has been established that ELISpot has a sensitivity several orders of magnitude better than intracellular cytokine staining (ICS) or ELISA assays. With ELISpot, it is possible to detect a single cytokine-producing cell from as many as 1 million bystander cells, and such sensitivity is important since antigen-specific T cells are typically present at low frequencies in vivo [2,3].
An ex vivo solid-surface capture assay for analyte quantification
The ELISpot workflow includes a cellular stimulation step, following which the secretory response is directly quantified upon the specific capture of the secreted molecules on a solid support. Briefly, the workflow is initiated by the addition of the target cells into antibody-coated plates in the presence or absence of stimuli. Cells are incubated to allow the secretory response to take place, and, after their removal, the captured analyte (e.g., a cytokine, a chemokine, or an immunoglobulin), is revealed by the addition of a biotinylated detection antibody followed by the avidin-enzyme-substrate amplification cascade (Figure 1). This results in the production of colored spots (visible to the naked eye), where each spot has been assumed to correspond to a single secreting cell. Thus, by counting the number of colored spots, it is possible to determine the frequency of antigen-specific T cells, which is a very relevant parameter in immunological applications. In addition, each spot represents the integration of the amount of the secreted analyte and its secretion kinetics; thus, the spot sizes also provide important biological information. For instance, in certain acute HIV subjects, it was shown that the frequency of cells producing IFN-γ was comparable to healthy donors, but the spot sizes were significantly reduced. These ELISpot results pointed to a new immunodeficiency mechanism in HIV subjects: a reduction in the IFN-γ output of individual cells, instead of a reduction in the number of T cells producing IFN-γ . In ELISpot, the color of the spots depends on the enzyme–substrate pair selected for the assay. Detection of cell-secreted proteins can also be achieved using fluorescence-labeled detection reagents (including streptavidin fluorescent conjugates), and in this case the assay is known as FluoroSpot. ELISpot and FluoroSpot assays use the same protocol for coating plates with capture antibodies, cell culture, and stimulation and washing steps (Figure 1). FluoroSpot has an advantage over ELISpot in that it is more suited for the simultaneous detection of multiple analytes, also known as multiplexing.
Until now, the colored or fluorescent spots have been enumerated by analyzing images taken with dedicated ELISpot/FluoroSpot readers. The existing ELISpot/FluoroSpot instruments take pictures of the wells and count the number of fluorescent spots with spot-recognition algorithms. In general, an accurate count strongly depends on the user’s inspection of the spots, which makes it a subjective and slow process. When dedicated readers are not available, researchers can use plate scanning services from reader manufacturers. However, unlike the colored spots from ELISpot, the fluorescent spots from FluoroSpot are not visible to the naked eye. As a result, researchers performing FluoroSpot assays cannot judge beforehand if the results from their experiments are acceptable or not as they cannot readily evaluate if there is a high number of spots in their positive controls (stimulated cells). By submitting plates with failed results for reading by dedicated imagers, researchers waste valuable time and incur unnecessary costs.
In this technical note, we introduce a new image-independent procedure for the initial evaluation of FluoroSpot plates, which can be performed with Thermo Scientific fluorescence microplate readers (Varioskan LUX Multimode Microplate Reader, Fluoroskan Microplate Fluorometer or Fluoroskan FL Microplate Fluorometer and Luminometer). For the current study, FluoroSpot assays were evaluated using the Varioskan LUX multimode reader. The procedure involves a rapid measurement of the fluorescence intensity within a defined arrangement of multiple points on the bottom of every well. The limit of detection (LOD) of this FluoroSpot pre-screening is calculated here for two model cytokines (IFN-γ and IL-2). The limitations and applicability of this image-independent FluoroSpot pre-screening method are also further discussed.
Materials and methods
A double-color (human IFN-γ and IL-2) FluoroSpot kit, CTL Anti-Aggregate Wash, and serum-free cell culture medium (CTL-Test medium) were purchased from CTL ImmunoSpot (Bonn, Germany). The experiments on human cells were performed using cryopreserved cells also acquired from the ePBMC library of CTL ImmunoSpot. Immobilon FL Millipore plates precoated with PVDF membranes (Cat. No. IPFL00010) were used in the measurements. Tween-20 was obtained from Sigma-Aldrich. eBioscience Cell Stimulation Cocktail (containing a combination of phorbol 12-myristate 13-acetate (PMA) and ionomycin (Cat. No. 00-4970-93)), PBS (Cat. No. 10010015 or 10010023) and trypan blue reagents (Cat. No. T10282 or C10228, included in the Countess Cell Counting Chamber Slides) were obtained from Thermo Fisher Scientific (Eugene, OR, USA).
Assays with cryopreserved human PBMCs
The immune response was studied using cryopreserved PBMCs. Cells were thawed according to the instructions of CTL ImmunoSpot. Briefly, vials (each with 40 million cells) were placed in a 37°C water bath for 8 min and samples were resuspended in 10 mL of prewarmed CTL Anti-Aggregate Wash, followed by centrifugation at 330 x g for 10 minutes. The supernatant was discarded, and the cell pellet was washed once more with 10 mL of CTL Anti-Aggregate Wash as just described. Before performing the second centrifugation step, an aliquot of the suspension was removed for cell counting. This was carried out using the Countess II FL Automated Cell Counter, and viability was assessed by trypan blue staining using Countess Cell Counting Chamber Slides (Cat. No. C10228). After the second centrifugation step, the supernatant was decanted, the PBMC pellet was resuspended into CTL-Test medium, and the concentration adjusted. The cell stimulation cocktail mixed with CTL ImmunoSpot’s co-stimulatory anti-CD28 antibody (0.1 µg/mL) was also prepared in CTL-Test medium, and this mixture (100 µL/well) was added to the precoated plates provided in the double-color (human IFN-γ and IL-2) FluoroSpot kit from CTL ImmunoSpot. Alternatively, 100 µL of CTL-Test medium alone was added to the precoated plates as a control for the background signal. PBMCs (100 µL/well) were then pipetted into the plates, with cellular concentrations ranging from 1.5 x 103 to 8 x 105 cells/well. At each concentration, cells were either exposed to the cell stimulation cocktail or to the CTL-Test medium to control for overall assay functionality (activated cells) and controls for background reactivity (background signal), respectively. These controls are essential to include in every FluoroSpot assay. During the stimulation step, cells were kept in a Heracell 150i CO2 humidified incubator (37°C, 5% CO2) for 20 h. Plates were not stacked or moved in any manner during this period. At the end of the stimulation period, plates were retrieved from the incubator, and the FluoroSpot assay was performed as described in the section below.
Double-color (IFN-γ and IL-2) FluoroSpot assay
The instructions of the kit manufacturer (CTL ImmunoSpot) were carefully followed during performance of the FluoroSpot assay. After the stimulation period, the plates were washed twice, first with PBS, and then with 0.05% Tween–PBS (200 µL/well in each case). Following the washing steps, the anti-human IFN-γ (FITC) and IL-2 (Hapten2) detection solutions from the CTL ImmunoSpot kit were added (80 µL/well), and samples were incubated at room temperature for 2 h. Afterwards, the samples were washed 3 times with 0.05% Tween–PBS (200 µL/well), followed by the addition of the CTL ImmunoSpot kit’s tertiary solution (80 µL/well), containing anti-FITC Alexa Fluor 488 and anti-Hapten2 CTL-Red probes for visualization of IFN-γ and IL-2, respectively. Samples were then incubated at room temperature for 1 h, followed by 3 washes with Milli-Q water (200 µL/well) obtained with an Elix Essential Water Purification System. After the wash steps, the protective underdrain was removed, and the back of the plates were rinsed with sterile water. Then the plates were left to air dry overnight before the generated spots were quantified as described below.
Image-independent evaluation of FluoroSpot assay: FluoroSpot pre-screening
Plates were evaluated using multipoint fluorescence intensity detection with top optics on the Varioskan LUX multimode reader, operated with SkanIt Software 6.0. The measurement protocol was set up by adding a "Fluorescence" measurement step with "Top" optics. Because a double-color FluoroSpot assay was used, two fluorescent labels were measured. To detect IFN-γ (with the anti-FITC Alexa Fluor 488 antibody), fluorescence signal was quantified with λex = 480 nm and λem = 520 nm, and to detect IL-2 (with the anti-Hapten2 CTL-Red antibody), fluorescence signal was quantified with λex = 480 nm and λem = 690 nm. For multipoint measurements, the option to perform "FluoroSpot pre-screening" was selected from the "Advanced parameters" dropdown menu as indicated in Figure 2. It is recommended that users select 12 nm of excitation bandwidth, whenever possible. This allows the light beam diameter to be 1.5 times larger than the one that is used when selecting 5 nm of excitation bandwidth.
Figure 2. Multipoint fluorescence intensity detection with top optics (FluoroSpot pre-screening) using SkanIt 6.0 software.
Following fluorescent measurement, the "top" fluorescence signal was mathematically integrated using SkanIt 6.0 software. For that purpose, a "Multipoint Reduction" step was added, and the "Sum" of the fluorescence points measured across every well was selected as the "Calculation type" (Figure 3).
Figure 3. Selection of calculation type when performing multipoint reduction in FluoroSpot evaluations.
Image-based evaluation of FluoroSpot assay
To image fluorescent spots, plates were processed using the ImmunoSpot S6 Ultra-V analyzer. Counting was performed using the ImmunoSpot software (version 184.108.40.206). The ImmunoSpot software analyzes digital images by measuring color density to distinguish spots from the background . Calculation of spot-size distributions is also a built-in function of the ImmunoSpot software. Data are recorded as spot forming units (SFU) per well. Spot size histograms can be created when needed, using the embedded features of the software.
Results and discussion
ELISpot and FluoroSpot assays measure cell's secreted cytokines
ELISpot and FluoroSpot assays are utilized for quantification of the secretory activity of individual immune cells. Flow cytometry-based ICS assays can also be used. However, ICS assays are limited in that they only provide information about the phenotypes of the cells, which may not necessarily relate to their biological functions. ELISpot assays, by contrast, do not allow for cell sorting based on cell surface or intracellular markers, but they do allow for the measurement of a cell’s secreted cytokines.
ELISpot and FluoroSpot assays offer sensitive detection and broad linear range
The detection limit is another important advantage of ELISpot assays. In ELISpot assays, it has been reported that 1 spot (corresponding to one secreting cell) can be identified among 1 million antigen presenting cells (APCs) for a detection limit of 0.0001% . As a general guideline, individual colored or fluorescent spots can be discretely discerned from 1 up to as many as ~500 spots per well in 96-well plates. Over the count of 500, the spots become too confluent to be reliably distinguished from the surrounding signal in the limited area of one well (~32 mm2). Other factors can also interfere with the spot-counting, such as the so-called ELISA effect.
The ELISA effect refers to an increase in the overall color of the membrane in a well when the number of cytokine-producing cells is high. The cytokine that is not captured around the secreting cell escapes in the supernatant and binds to the capture antibodies in their vicinity, thus changing the color of the background of the well.
A theoretical linear range is expected between 1 and 500 IFN-γ positive T cells in ELISpot/FluoroSpot assays in 96-well plates. Of note, nonspecific spots can be present in negative controls (for example, as result of cytokine generation by APCs). These nonspecific spots are typically accounted for by gating them out, using the image processing software of ELISpot/FluoroSpot readers.
FluoroSpot assays can be pre-screened using Thermo Scientific fluorescence microplate readers
We hypothesized here that the fluorescence signal generated by the FluoroSpot assay could be alternatively evaluated using multipoint fluorescence detection using a microplate reader with top-reading capabilities, such as the Varioskan LUX multimode reader used in this study. Because of the high-quality optics used in Thermo Scientific fluorescence microplate readers, we reasoned that it would be at least theoretically possible to distinguish the fluorescence signal generated by a low number of cytokine producing T cells from the background signal of a FluoroSpot assay. To evaluate this hypothesis, we performed an actual comparison of the signal produced by IFN-γ– and IL-2–secreting T cells detected with the Varioskan LUX microplate reader and a reference spot-counting instrument, using human PBMCs. To estimate the sensitivity of microplate reading, we first measured the LOD values for both cytokines in this cellular system (Table 1). The LOD value is defined as the lowest amount of analyte that can be statistically separated from the background, and the value is typically calculated as the average of the blank values plus 3*SDs (or three times the standard deviations) of the blanks, assuming they are binomially distributed. Assay blanks in FluoroSpot can produce a background fluorescence signal that originates from the nonspecific binding of fluorescent-labeled components to the FluoroSpot membranes or even from autofluorescent components in the PVDF membranes. Such nonspecific binding may not result in background spots, but it does produce a signal that is measurable with a fluorescence microplate reader, such as the Varioskan LUX multimode reader.
|LOD value (IFN-γ)
|LOD value (IL-2)
Table 1. LODs measured for IFN-γ and IL-2 secretion when using multipoint fluorescence intensity detection with top optics, or FluoroSpot pre-screening, with the Varioskan LUX multimode reader.
The calculated LOD values, expressed as IFN-γ or IL-2 fluorescent spots, are summarized in Table 1. Using the Varioskan LUX multimode reader, it is possible to detect the fluorescence generated from 201 IFN-γ positive or 163 IL-2 positive cells under the experimental conditions described here. Based on this, it is likely that the number of spots generated in the controls for overall assay functionality (activated cells), which is usually expected to be high, can be detected as significantly distinct from the fluorescence generated in the controls for background reactivity (background signal, cells exposed only by culture media) with a fluorescence microplate reader. Conversely, if fluorescence values measured with the Varioskan LUX reader do not statistically differ from the fluorescence measured in blank samples, it can be assumed that no more than 201 or 163 spots (for IFN-γ or IL-2 positive cells, respectively) have been produced in the positive controls or other measured samples. Thus, for FluoroSpot users who are familiar with their assays and are aware of the typical behavior of their positive controls for either IFN-γ or IL-2 secretion, it is possible to quickly know if the positive controls have exceeded the number of spots associated with the LOD of the fluorescence plate reader. It is recommended for other cellular systems or experimental conditions that the LOD is measured with the Varioskan LUX or Fluoroskan plate reader and compared to spot-counting values. This would help save time and resources in the long run, especially for research laboratories that do not have a spot-counting instrument at their permanent disposal, but do have a Thermo Scientific fluorescence microplate reader. We also attempted to quantify the linear ranges of the signal measured with Varioskan LUX in comparison to a spot-counting instrument. An example of the results obtained when human PBMCs were stimulated to produce IFN-γ is presented in Figure 4.
When stimulating 6,250 PBMCs/well, a total of 22 IFN-γ–positive cells (spots) can be detected and separated from the background samples with the spot-counting instrument, resulting in an acceptably high screening window coefficient (Z’) value of 0.66. Of note, detection of 10 IFN-γ–positive cells is also possible with a lower number of cells (3,125 PBMCs/well) but it does not result in a high-quality screening assay (Z’ value of 0.21) due to the smaller dynamic signal window. Because the calculated LOD of the Varioskan LUX reader is roughly 9 times higher than 22 (that is, around 200 spots), it is possible to detect a similar frequency of IFN-γ–positive spots (0.35%) from a cellular amount that is roughly 9 times higher than 6,250 cells (that is, 56,250 PBMCs/well). Indeed, with 50,000 PBMCs/well the measured fluorescence signal in Varioskan LUX is significantly different than in the background wells, resulting in an acceptably high Z’ value of 0.65. Thus, in this case, concentrations over 50,000 PBMCs/well would be sufficient to be able to detect a significant fluorescence signal on the Varioskan LUX microplate reader. In practice, the LOD value can change slightly from instrument to instrument, and will also depend on the quality of the performed assay.
As discussed earlier, the spot-counting method shows a narrower linear range because of the intrinsic limitation of the imager to resolve spots when they become crowded in the wells. This drawback is not observed when measuring fluorescence signal using a microplate reader. The fact that the signal measured with the Varioskan LUX reader stays linear even with a very high number of IFN-γ–secreting cells is not surprising, since even high concentrations of strongly emitting fluorophores do not typically cause saturation of the instrument’s detector. However, it must be noted that a high fluorescence signal measured with the Varioskan LUX microplate reader is an indication that there are a large number of spots in one well, which in some cases could be too numerous to be effectively resolved and counted with an imager. We recommend calculating the signal-to-background (S/B) ratio by dividing the RFU measured in positive control wells (activated cells) by the RFU measured in blank (background signal) samples. If the S/B ratio is higher than 2.5, there is a high chance that the number of generated spots will not be measurable with a spot-counting instrument. Thus, the user can assume that those wells will be fully saturated with spots that cannot be resolved due to the presence of too many productive cells. On the other hand, if the S/B ratio is lower than or close to 2.5, it can be concluded that the number of spots will be within the measurable signal range of an image-based FluoroSpot reader and it is reasonable to send the plates for counting.
Key things to remember when using Thermo Scientific fluorescence microplate readers for FluoroSpot pre-screening
This method does not:
- Replace FluoroSpot counting instruments
- Count the number of spots or spot forming units (SFU) per well
- Perform imaging of the wells
- Allow quantification of the fluorescence signal generated by a number of spots that falls under the LOD
- Identify cells that produce more than one cytokine
This method does:
- Identify samples with low secretory activity that generate less than a certain number of spots per well (falling below the LOD)
- Identify wells with high number of secreting cells (wells with high fluorescence signals when compared to blanks, i.e., signal-to-background values over 2.5)
- Allow evaluation of the FluoroSpot plates that is independent from user’s judgment
In summary, we have presented the benefits and limitations of utilizing a microplate reader for a first and rough evaluation of FluoroSpot results, or a FluoroSpot pre-screening. This strategy provides an image-independent method for initial FluoroSpot evaluation, but it is not a replacement for quantifying the number of spots with a suitable spot-counting instrument. This is because the number of spots is a biologically significant parameter, as 1 spot is equivalent to 1 actively secreting cell. Such information cannot be obtained by merely measuring the fluorescent signal of the spots. However, by quantifying the fluorescence with a multipoint measurement, or FluoroSpot pre-screening, as described here, the user may discriminate between failed and successful experiments by comparing the results of the positive control cells with the blank samples. This method is totally independent of the user’s judgment and can be readily applied for any type of fluorophore, given the free selection of excitation and emission wavelengths offered by the scanning fluorometer in the Varioskan LUX multimode reader and the broad availability of affordable filters for the Fluoroskan and Fluoroskan FL readers. When evaluating FluoroSpot plates with the Fluoroskan and Fluoroskan FL readers, suitable filters may need to be separately ordered.
- Janetzki S, Price L, Schroeder H et al. (2015) Guidelines for the automated evaluation of ELISpot assays. Nat Protoc 10(7):1098-115.
- Tanguay S, Killion JJ (1994) Direct comparison of ELISPOT and ELISA-based assays for detection of individual cytokine-secreting cells. Lymphokine Cytokine Res 13(4):259-63.
- Lehmann PV, Zhang W (2012) Unique strengths of ELISPOT for T cell diagnostics. In: Kalyuzhny, AE (editor) Handbook of ELISPOT. Methods in Molecular Biology (Methods and Protocols), vol 792. Totowa, NJ: Humana Press. pp 3-23.
- Helms T, Boehm BO, Asaad RJ et al. (2000) Direct visualization of cytokine-producing recall antigen-specific CD4 memory T cells in healthy individuals and HIV patients. J Immunol 164(7):3723-32.
- Lehmann PV (2005) Image analysis and data management of ELISpot assay results. Methods Mol Biol 302:117–132.
- Hesse MD, Karulin AY, Boehm BO et al. (2001) A T cell clone’s avidity is a function of its activation state. J Immunol 167(3):1353-61.
CTL Anti-Aggregate Wash, CTL-Test, CTL-Red, and ePBMC are trademarks of CTL ImmunoSpot (Bonn, Germany). Milli-Q and Elix are registered trademarks of Millipore Corp. Tween is a registered trademark of ICI Americas Inc.
For Research Use Only. Not for use in diagnostic procedures.