SEARCH

SEARCH BY CITATION

Keywords:

  • laser scanning cytometry;
  • photobleaching;
  • polychromatic cytometry;
  • northernlights;
  • molecular and optical live cell imaging

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. LITERATURE CITED

In recent years, slide-based cytometry has become a key technology for polychromatic cytometric investigations, and many efforts have been made to increase the number of measurable fluorochromes for multiparametric analysis. Sequential photobleaching of fluorochromes next to very photostable dyes is one approach for this technology. As the ALEXA dyes are known to be photostable as compared to the conventional fluorochromes FITC, PE (Riggs et al., Am J Pathol 1958;34:1081–1097), and APC, a differentiation within a fluorochrome pair is possible. Here, we have analyzed the newly available NorthernLights™ secondary antibodies for use in slide-based cytometry and microscopy. Currently, these fluorochrome-conjugates are now available with three distinct excitation- and emission maxima (NL493, NL557, NL637). Their spectral properties are similar to the frequently used fluorochromes FITC, PE, and APC and can, therefore, be used with most common excitation sources of cytometers or microscopes. As the NorthernLights™ are bright, resistant to photobleaching, stable in alcohols and xylene and of affordable price, these dyes are promising candidates for use with most laser- and HBO/XBA-based fluorescence microscopy-like techniques. © 2010 International Society for Advancement of Cytometry

IN recent years, multiparametric analysis of biological specimens has become increasingly important in various fields of biology, in particular due to the emerging new fields of high-content and high-throughput single cell analysis for systems biology (1–9) and cytomics (10, 11). In small sample volumes or very precious samples, multiparametric analyses are in the majority of cases the only way to get requested information. Furthermore, research mainly focuses on analyses that allow to virtually measure “anything” in a cell simultaneously, for example, to unravel the complexity of the immune system (8, 9, 12, 13). This inquiry led to the development of new attempts in fluorescence analyses where various cellular constituents of interest can be determined in one single analytical run.

In multicolor analysis, cellular information is obtained by staining the sample using a multitude of different fluorescent dyes. The used dyes have to be separated from each other due to spectral characteristics. This means, to be able to identify the respective dye, that is, cell marker, distinct signals must be applied for analysis. If fluorescence signals are differentiated by physical or virtual band-pass filters, a multitude of control measurements have to be performed to determine and exclude spillover signals (14). Additionally, the hardware of the analysis system, mainly available excitation sources and band-pass filter cubes, defines applicable dyes and limits the amount of analyzable parameters. Admittedly, modern flow cytometers, for example, the LSR II (BD Biosciences), are equipped with up to four lasers and an octagon detector array, which allows for acquiring up to 18 colors. This system makes multiparametric analyses possible but also requires highly skilled operators (12).

Next to modern flow cytometry, slide-based cytometry (SBC) also allows for highly multiplexed fluorescence detection. Besides laser and detector upgrades, SBC enables another way for multiparametric analyses. By the fact that samples will not get lost by the measurement, the reanalysis of the sample in combination with a subsequent merge of the data sets may help to obtain additional parameters on a single-cell basis. This approach has been used to restain cell nuclei after immunophenotyping (8, 9, 13, 15), to combine alive and dead information (16), to analyze destaining kinetics (17), and to restain antigens with additional sets of antibodies on fixed single cells (18) or in tissue (19). But not only restaining procedures lead to multiparametric fluorescence analysis: different sensitivity of fluorescence dyes can also be used to obtain additional parameters for analysis (20). By sequential photobleaching, fluorescence dyes with a similar emission spectrum (e.g., APC and AlexaFluor 633) can be distinguished after prolonged excitation due to their different photo stability to excitation light. This means, on the one hand, for multiparametric analyses in SBC the limits of the instrument can be exhausted and on the other hand, the amount of analyzable parameters can be further expanded by reanalysis procedures without the need to modify filter settings in existing technical setups. As it is necessary for multicolor analyses to have several detectors and various excitation sources (which normally means modification of existing hardware or purchase of new and expensive cytometers), the newly available NorthernLight dyes have been evaluated for suitability in a different approach, sequential photobleaching. It was demonstrated earlier (20) that fluorescence dyes with similar emission spectra are distinguishable because of their different susceptibility to photobleaching. NL dyes were analyzed for their bleaching characteristics and hence, their applicability in multiparametric cytometry and fluorescence microscopy.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. LITERATURE CITED

Staining

Slide-based cytometry

Aliquots of EDTA anticoagulated blood (40 μl) were stained with 5 μl CD3-Biotin (Invitrogen, Karlsruhe Germany) and incubated for 30 min at room temperature in the dark. Afterward samples were washed three times with 1 ml of PBS followed by a subsequent centrifugation step (400g, 5 min). Supernatant was removed and 5 μl of Streptavidin coupled to FITC, PE, APC, Alexa488, Alexa532, Alexa633 (all Invitrogen), Cy2, Cy3, Cy5 (all Jackson ImmunoResearch, Newmarket, UK), NL493, NL557, or NL637 (all R&D Systems, Wiesbaden, Germany) at a concentration of 1:10–1:500 was added and incubated for 2 h at 4°C. Following incubation, erythrocytes were lysed by adding 1 ml of lysis buffer (FACS Lysis; BD Biosciences) for 10 min. After centrifugation, samples were washed twice with PBS (pH 7.4, Sigma Aldrich, St. Louis, MO, USA). DNA was stained with 5 μl Hoechst33342 (Invitrogen, 50 μg/ml) for 10 min. A well of a flat-bottom 96-well plate was preloaded with one drop of DAKO fluorescent mounting medium (DAKO, Glostrup, Denmark). The stained sample was mixed with the medium. Cells were spun down at 400g for 2 min. Samples were stored at 4°C overnight to achieve immobilization.

Fluorescence microscopy

Cells from the renal cell carcinoma cell line A498 were plated out in two well LabTec™ II Chambers Slides (NUNC, Wiesbaden, Germany)—a specific nonfluorescent glass microscope slide—in a final density of ∼100,000 cells per well. A498 cells show adherent growth characteristics. Cells were fixed before staining procedures using a mild 70% EtOH-fixation for 3 min at room temperature, followed by 10 min drying.

For an overall cellular staining a mouse anti-human cytokeratin antibody (IgG, DAKO, Hamburg, Germany) was used in a concentration of 1:50 and incubated at room temperature for 45 min. Surplus antibody was removed by a two times wash step using PBS (without CaCl2, MgCl2, pH 7.4, Gibco, Invitrogen, Heidelberg, Germany). Cells within the left chamber were then incubated with anti-mouse IgG-NL557 antibody (1:200; R&D Systems, Wiesbaden, Germany) and cells within the right chamber were incubated with anti-mouse IgG-PE (1:100; BD, Heidelberg, Germany) according to manufacturers' instructions for another 45 min. Another slide was stained for the second fluorochrome pair FITC (anti-mouse IgG- FITC, 1:100, BD, Heidelberg, Germany) and NL493 (anti-mouse IgG-NL493, 1:200; R&D Systems, Wiesbaden, Germany). Surplus secondary antibody was removed by carefully washing twice with PBS. The chambers were removed gently and remaining silicone was removed using a scalpel. To reduce photobleaching, Vectashield embedding medium containing DAPI was used and slides were covered with cover slips (Vectashield H1200, Vector Laboratories, Burlingame, CA, USA).

Data Acquisition and Measurement

Cytometric analysis

Leukocyte samples were analyzed by iCys Research Imaging Cytometer (CompuCyte Corp., Westwood, MA, USA). The 633-nm laser line excited APC, Cy5, Alexa633, and NL637; for all other streptavidin conjugates 488 nm was used (Fig. 1). The Hoechst signal served as trigger for analysis.

thumbnail image

Figure 1. Absorption and emission characteristics of tested fluorochromes.: Left side shows absorption characteristics of the dyes and the used laser excitation line (488 nm or 633 nm). On the right hand side, emission characteristics (ex/em data: Invitrogen, except for NorthernLights™: R&D) are displayed with the respective band-pass filter used for cytometric analysis as an underlay. Maximum for absorption and emission is indicated in the center. As one can see spectral characteristics of used dyes for comparison are very similar.

Download figure to PowerPoint

PMT values, offsets, and threshold were set for optimal cell/fluorescence detection. Each sample was scanned 30 times for obtaining bleaching characteristics of the respective fluorescence dye. For spillover analysis, samples were measured with following band-pass filters in descending wavelength order: 463/39 nm (filter for Hoechst), 530/30 nm (filter for FITC and analogs), 580/30 nm (filter for PE and analogs), 667/30 nm (filter for APC and analogs), 715/30 nm, and 760/40 nm.

Debris and doublets were excluded by area and lymphocytes were gated for analysis. Positively and negatively stained cells were gated. Fluorescence intensity (MFI) values for both populations were recorded for each of the 30 scan cycles.

The fluorescence integral value of the negative population was subtracted from the positive one and normalized to the value of the first scan.

  • equation image

These normalized fluorescence intensities were displayed over scan cycles to obtain bleaching characteristics for each fluorescence dye. For spillover analysis, positively and negatively stained cell populations were traced in the main channel for the respective dye, and their fluorescence intensities were recorded in all other channels. Spillover signals were calculated as follows (SC: Spillover channel, MC: main detection channel for the respective dye, +: integral fluorescence intensity of positive cells, −: negative cells):

  • equation image
Fluorescence imaging

iCys Research Imaging Cytometer (CompuCyte Corp.) used for microscopic fluorescence imaging is additionally equipped with a high resolution black-and-white CCD camera (F-View II extended, Olympus, Hamburg, Germany) in combination with high quality hard coated fluorescence filter sets (AHF Filtertechnik, Tübingen, Germany) and the combination of a HBO- and XBO excitation source. The F-View II camera is connected to a second autonomous PC-system and is, therefore, not interacting with the iCys management. Once the cells are analyzed by the iCys system (20× objective) and the corresponding histograms are recorded for both competitive channels (PE vs. NL557 and FITC vs. NL493), the high resolution F-View camera was used to image representative cells from the scanned area (100× objective) via the second PC-system. Images were taken as black and white pictures using specific filter sets (FITC/NL493: HC F36-501 ex 482 nm ± 35 em 536 nm ± 40 nm, PE/NL557: HC F36-542 ex 531 nm ± 40 em 593 nm ± 40 nm, both filters by AHF Analysentechnik, Tübingen, Germany) for both channels (cytokeratin [FITC/NL493] and DAPI or Cytokeratin [PE/NL557] and DAPI. The pictures were then merged and analyzed by CELL-F® (Olympus, Hamburg/Germany). A horizontal intensity profile enables a precise measurement of the gray values of the respective black-and-white-channels. The process iCys-scanning/F-View imaging was repeated five times. To compare photobleaching effects within one competitive fluorescence pair (PE vs. NL557 and FITC vs. NL493), the mean grey values of three representative images were compared to each other.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. LITERATURE CITED

Methods to differentiate fluorescence dyes with a comparable excitation and emission spectra were described earlier (20). By prolonged excitation of a sample stained with AlexaFluor dyes (e.g., Alexa633) and their conventional analogs (e.g., APC), a discrimination of the cell populations stained with these dyes was possible.

Aim of the repetitive scans on leukocytes was to evaluate further dye combinations for their use in multiparametric analyses. To this end, leukocytes were stained with CD3-Biotin followed by Streptavidin coupled either to AlexaFluor dyes, Cyanine dyes, Northern Lights, or their conventional analogs and checked for photobleaching by 30 repetitive scan cycles. Additionally, NorthernLights were checked for their bleaching characteristics and capabilities in microscopy. Focusing in microscopy often is done on the fluorescence of nuclear staining. Hence, used fluorescence dyes must resist excitation light during focusing procedures, localization of cells, and selection of an appropriate area. Therefore, aim of this experiment was to test bleaching characteristics of fluorescence of additional secondary antibodies (next to the DNA dye) by exciting the sample with UV light.

Bleaching

As expected, fluorescence intensity of APC and PE decreased very rapidly (loss of 50% fluorescence intensity of the initial value after 3–4 scans, Fig. 2 B, C). The positive cell population could hardly be distinguished from the negative population after a few scan cycles. Surprisingly, FITC showed a relatively good photo stability (Fig. 2A). After 30 scan cycles still 50% of the initial brightness remained. Despite a continuous loss of fluorescence intensity, the level of FITC was always higher than that of Alexa488 and Cy2. This is rather surprising as our previous bleaching experiments (21) as well as literature (22) reports the contrary. Nevertheless, a higher resistance to repeated laser excitation showed Alexa532 which was with about 1% intensity loss per scan quite stable. After an initial drop by about 10%, further bleaching of Cy3 was only marginal and reached with 83% after 30 scans higher fluorescence intensity levels than Alexa532 (73%). The new NL557 was almost nonbleachable and is, therefore, more photo stable than the other tested orange/red emitting dyes. Alexa532 showed similar bleaching characteristics than Cy5 (loss of 1.5% per scan), whereas Alexa633 was comparable to Cy3. Anyway, Alexa633 showed a better long-term stability (remaining 62% at scan 30) than Cy5 (51%). NL637, however, exhibited a completely deviant behavior to repeated laser excitation. One can hardly speak about bleaching characteristics. With every scan cycle, NL637 became brighter until a maximum of 208% (at scan 22) was reached. Not until then did the fluorescence intensity start to drop down. Similar characteristics also showed NL493, although with 112%, not such a strong rise could be observed.

thumbnail image

Figure 2. Bleaching characteristics: Background-corrected integral fluorescence intensities (normalized to 100% at scan 1) of each used dye are displayed for each of the 30 scan-cycles. Bleaching experiments were repeated 3–5 times per dye and showed high reproducibility; measurement means are displayed. Whereas the fluorescence intensity of almost all fluorescence dyes is decreasing by repeated excitation with the respective laser line, NL493 and NL637 get brighter.

Download figure to PowerPoint

Spillover

For multicolor measurements, spillover signals play an important role. Possible spectral interferences have to be known if combining several fluorescence dyes. For an unequivocal identification of each single dye in analysis, a spillover matrix must be created. We performed these matrices in this study for the tested dyes and the band-pass filters used in SBC (23).

All green emitting dyes showed a substantial degree of spillover into the next (orange) channel (Fig. 3). Alexa488 showed thereby with 30% twice as much spillover as FITC, Cy2, and NL493 (Table 1). Further interferences into other channels could not be detected, for none of these dyes. Within the orange emitting fluorescence dyes, PE was the only dye that exhibited a spillover signal in the third channel (11%). The spillover of the other dyes was below 5%. Disturbing signals in the green channel could only be observed for Alexa532. All tested red excited dyes partially generated very prominent signals in the used red emission filters. Cy5 was thereby the dye with the highest spillover signals.

thumbnail image

Figure 3. Spillover signals: Leukocytes were identified by their DNA signal. CD3+ cells were distinguished in the primary channel for the respective fluorescence dye. Spillover signals, that is, fluorescence intensities resulting from stained cells (grey), were analyzed in all other channels (bandpass filters for acquisition see above). Almost all fluorescence dyes showed spillover signals into other channels. Despite the similar emission spectra, there are differences between the dyes. The new NorthernLights™ showed partially lower spillover than the established fluorescence dyes and are, therefore, good alternatives for multiparametric analyses.

Download figure to PowerPoint

Table 1. Spillover matrix
 488 nm excitation633 nm excitation
530/30 nm (FITC)580/30 nm (PE)667/30 nm (PE-Cy5)715/30 nm (PE-Cy5.5)760/40 nm (PE-Cy7)667/30 nm (APC)715/30 nm (APC-Cy5.5)760/40 nm (APC-Cy7)
  1. Positively and negatively stained cell populations were traced in the main channel for the respective dye. Their fluorescence intensities were recorded in all other channels. Spillover signals were calculated dividing the fluorescence intensity (background corrected) of the dye in the main channel by its intensity in the spillover channel multiplied by 100.

FITC100.0013.840.04−0.16−0.46−0.40−0.31−0.50
Cy2100.0015.05−0.34−0.50−0.54−0.14−0.57−0.78
Alexa488100.0030.071.060.02−0.22−0.04−0.15−0.31
NL493100.0013.580.52−0.35−0.70−1.12−1.01−0.90
PE1.87100.0010.702.211.620.25−0.02−0.26
Cy3−8.33100.003.660.61−0.26−0.69−0.49−0.92
Alexa53210.19100.002.08−0.05−1.27−0.84−0.69−1.33
NL557−0.52100.004.990.860.27−0.48−0.60−0.59
APC−0.58−1.542.36−0.06−0.72100.0017.7410.07
Cy5−1.52−1.210.240.05−0.16100.0066.4124.54
Alexa633−2.00−1.500.35−0.12−0.38100.0054.3712.92
NL637−1.77−1.382.772.520.22100.0040.349.23

Microscopy

Both tested NorthernLights™ labeled secondary antibodies (NL493 and NL557) show nearly no photobleaching and very bright signals compared to FITC and PE (Fig. 4).

thumbnail image

Figure 4. Photobleaching of NL493 and FITC, and NL557 and PE: A: stained cells before first bleaching step, B: after 6 min and C: after 20 min UV-light bleaching (405 nm). DNA staining by DAPI is blue, FITC and NL493 green, and PE and NL557 red. Whereas FITC and PE lost most of their fluorescence (although excitation light is not optimal for excitation) the NorthernLights™ analogs keep most of their intensity.

Download figure to PowerPoint

Switching to the F-View II camera system after scanning with the iCys system offers to record sharp pictures from one focus level instead imaging pictures over a wide focus using the LSC technique. Remarkably, the staining behavior of the used second antibodies is different, comparing the HBO-fluorescence based F-View II pictures. Although the NorthernLights ™ show a precise staining of the cytokeratin in the form of clearly visible fiber-like structures, the PE labeled antibody did show a patchier staining with no fiber-like structures (Fig. 4).

Photobleaching levels were measured over time using a UV-light source for excitation (150W HBO-source, 405 nm filter). Gray value intensities were measured using the Horizontal Intensity Profile Measurement Function of the DOKU®-Software. Although the PE gray value intensity was bleached out to zero using 6 min bleaching by UV-light, the corresponding NL557 dye gray value after 6 min bleaching still remains at 2500 (Fig. 5). Remarkably, after 40 min bleaching, the NL557 signal is comparable to the initial PE signal before any bleaching. NL493 and FITC showed similar results.

thumbnail image

Figure 5. Photobleaching by UV excitation: All displayed dyes were bleached by 405 nm UV-light excitation (150W HBOlamp). Fluorescence intensity was determined at different time points. Both NorthernLight™ dyes (NL557 and NL493) show about 35–50% higher gray value intensities compared to the corresponding FITC and PE values.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. LITERATURE CITED

We can demonstrate here that the new NorthernLights™ are applicable for SBC and microscopy. Secondary antibodies are available at present in three different spectral characteristics (NL493, NL557, and NL637, Fig. 6), which makes them feasible for most excitation sources in cytometers or microscope-based systems without any technical modifications. NorthernLights™ are available as Streptavidin conjugates or as secondary antibody conjugates directed to IgG of different species. Hence, they are ideal partners for staining most primary IgG-antibodies. Brightness is comparable to the respective dye analogs. Therefore, these new dyes are definitely alternatives to the commonly used fluorochromes in fluorescence analyses. Because of their excellent photo stability, they are actually better suited for microscopy than AlexaFluor dyes.

thumbnail image

Figure 6. Chemical structures of NorthernLights™: Chemical structures of NorthernLight molecules are kindly provided by R&D systems.

Download figure to PowerPoint

The fact that NorthernLights™ gets even brighter during long-term excitation makes them interesting for hyperchromatic cytometry (21). It could be demonstrated earlier that different bleaching characteristics of fluorescence dyes can be used to increase the amount of measurable parameters in multiparametric analyses (20). In this study, APC and Alexa633 were used for staining of leukocytes in one sample. This dye combination, that is normally not distinguishable, could be discriminated by bleaching procedures. Because of the loss of fluorescence intensity in APC and the stability of Alexa633, the stained cell populations could be distinguished after photobleaching. Given that NL637 is not only slowly-bleachable but becomes even brighter by bleaching procedures. This dye could be combined with APC and Alexa633 for a 3 “color” approach. As a possible cause for this fluorescence behavior, two possibilities are considered in our view. On the one hand, the chemical structures (see Fig. 6) of the NorthernLights™ could be changed by laser light, resulting in new dyes with brighter emission characteristics. Nevertheless, on the other hand, a quenching of these dyes is more plausibly. Overlabeling can cause quenching as a result of fluorescent emissions from one dye molecule being absorbed by neighboring dye molecules. The F/P ratios are kindly given by the quality control of the producer as following: NL557 1:2.3, NL637 1:1.52, and NL493 1:1.23.

Although photo stable dyes (e.g., Alexa633) are located on a diagonal if displaying scans before and after bleaching (due to nearly the same fluorescence intensity before and after bleaching), APC stained cells will drop down to the prebleach axis and NL637+ cells will move to the postbleach axis. Therefore, the discrimination of these dyes is possible only on the basis of their different photo stability characteristics.

As NL637 and NL493 showed the same characteristics in bleaching, although not with the same intensity, it seems to be obvious that NL557 also exhibits similar features. The most likely reason for this altered characteristic may be the nonoptimal excitation at 488nm. If optimally excited with green light, NL557 also gets brighter during long-term excitation.

Also for multicolor analyses, these new dyes may be good alternatives for staining. The relatively low spillover into other detection filters makes the NorthernLights™ appropriate dyes for multiparametric fluorescence analyses.

Excellent bleaching characteristics make these dyes promising candidates for use in laser-based in-vivo and in-vitro microscopy, for example, for the detection of single cells in intra-vital microscopy (24) or for multicolor immunofluorescence spectral imaging microscopy (25). These new dyes are definite alternatives to the widely used ALEXA dyes (22). It can be demonstrated in this study that NL dyes get brighter (twice as bright as in the beginning) before they start to fade. This is contradictory to the common and often bemoaned behavior of fluorescence dyes at prolonged excitation. Excitation and emission characteristics combined with unique bleaching behavior will make these dyes very interesting for scientists working in the field of microscopy and/or cytometry.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. LITERATURE CITED