Combined Tyramide Signal Amplification and Quantum Dots for Sensitive and Photostable Immunofluorescence Detection

Abstract

Conventional immunofluorescence detection of biologically relevant proteins and antigens in tissue sections is often limited by relatively weak signals that fade rapidly on illumination. We have developed an immunohistochemical protocol that combines the sensitivity of tyramide signal amplification with the photostability of quantum dots to overcome these limitations. This simple method provides a sensitive and stable fluorescence immunohistochemical alternative to standard chromogen detection.

Immunofluorescence microscopy is a well-characterized technique for the detection of proteins and antigens in tissue sections. However, currently available organic fluorescent labels, such as fluorescein, cyanine dyes, and AlexaFluor dyes, are variably limited by relatively rapid and irreversible photobleaching under high-intensity illumination. Although specific photo-stabilizing buffers have been developed to decrease photobleaching, these buffers may decrease initial fluorescence signal intensity, are variably effective, and are incompatible with some organic fluorophores (Ono et al. 2001).

Some of the limitations in immunofluorescent labeling have been overcome by using enzyme-amplification techniques, particularly tyramide signal amplification (TSA; Perkin-Elmer Life Sciences, Boston, MA), which may dramatically increase immunofluorescence sensitivity over traditional fluorescence detection procedures (Bobrow et al. 1992; van Gijlswijk et al. 1997). TSA is based on the horseradish peroxidase (HRP)-catalyzed deposition of labeled tyramide molecules at sites of antibody binding (Bobrow et al. 1989,1992; Shindler and Roth 1996). Tyramide is converted by HRP into a highly reactive oxidized intermediate that binds rapidly and covalently to cell-associated proteins at or near the HRP-linked antibody. TSA immunofluorescence can be performed either with fluorophore-conjugated tyramide or with biotinylated tyramide that is subsequently labeled with fluorophore-conjugated streptavidin. Although TSA can increase the sensitivity of immunofluorescence detection, it does not affect the photostability of the organic fluorophore used to localize the deposited tyramide. Therefore, photobleaching may remain a significant problem with TSA fluorescence detection.

Quantum dots are newly available photostable fluorophores that also overcome some of the limitations of organic fluorescent dyes. These nanocrystals (commercially available as Qdots from Quantum Dot Corporation; Hayward, CA) consist of a semiconductor core of cadmium selenide coated with a shell of zinc sulfide. An additional polymer layer over the quantum dot enhances water solubility and enables conjugation to streptavidin or other biomolecules such as IgG (Wu et al. 2003). These particles have the optical properties of high brightness, photostability, narrow emission spectra, and an apparent large Stokes' shift (Watson et al. 2003). Slight changes in the size of the quantum dot semiconductor core change the emission spectra, resulting in a redder color as the size of the core increases. Here, we describe the combined application of TSA and quantum dot fluorophores to immunofluorescent detection of intracellular antigens in tissue sections.

Materials and Methods

Materials

Qdot-streptavidin conjugates (Qdot 605, 585, 565, 525) were purchased from Quantum Dot Corporation. Anti-synaptophysin, a mouse monoclonal antibody (MAb), was purchased from Boehringer Mannheim Biochemica (Mannheim, Germany). Anti-MAP2, a mouse MAb, was purchased from Sigma (St Louis, MO). Secondary antibodies (biotinylated and HRP-conjugated anti-mouse antibodies) and Cy3-conjugated streptavidin were purchased from Jackson Immunoresearch Laboratories (West Grove, PA). Fluorescein isothiocyanate (FITC)-conjugated streptavidin was obtained from Pharmingen (San Diego, CA). Alexa 546-conjugated streptavidin was purchased from Molecular Probes (Eugene, OR). Avidin-biotin blocking kit was purchased from Vector Laboratories (Burlingame, CA). Biotin tyramide and TSA amplification buffer were obtained from Perkin-Elmer Life Sciences.

Immunohistochemical Methods

Brains were obtained from either adult Sprague-Dawley rats (Harlan; Indianapolis, IN) or C57BL6/J mice (Jackson Laboratory; Bar Harbor, ME). Embryos were obtained from gestational day 14 C57BL6/J mice. Whole mouse embryos and adult mouse brain tissues were immersion-fixed overnight at 4C in Bouin's fixative. Adult rat brain was immersion-fixed in either Bouin's fixative or 4% paraformaldehyde to compare the effects of the two fixatives on Qdot immunofluorescence detection. Brains used for frozen sections were cryoprotected in 30% sucrose in PBS (10 mM Na₂HPO₄, 1.4 mM KH₂PO₄, 140 mM NaCl, 2.7 mM KCl, pH 7.2), frozen in liquid 1,1,1,2-tetra-fluoroethane on dry ice, and then sectioned by cryostat into 10-μm sections. Four-μm-thick paraffin sections were prepared on a microtome.

After rehydration of frozen sections in PBS or deparaffinization and rehydration of paraffin-embedded sections in serial solutions of Citrisolv (Fisher Scientific; Pittsburgh, PA), isopropyl alcohol, and water, selected tissue sections underwent antigen retrieval by steaming in 10 mM citric acid, pH 6.0, for 20 min and cooling for 20 min. The remaining steps were performed at room temperature unless indicated otherwise. Endogenous peroxidase activity was inhibited by incubating paraffin sections in 3% H₂O₂ in PBS for 5 min or frozen sections in 0.3% H₂O₂ in PBS for 15 min, followed by three 5-min washes in PBS. Sections were then incubated for 30 min in PBS-blocking buffer (PBS-BB; 1% bovine serum albumin, 0.2% powdered skim milk, 0.3% Triton X-100 in PBS) followed by incubation overnight at 4C with primary antibody diluted in PBS-BB or in PBS-BB alone. Slides were then rinsed three times for 5 min in PBS before application of the secondary antibody in PBS-BB for 1 hr (biotinylated and HRP-conjugated anti-mouse antibodies were diluted 1:500 and 1:2000, respectively).

In experiments employing biotinylated secondary antibody, slides were washed three times for 5 min in PBS and incubated with Qdot-streptavidin conjugates for 30 min at the indicated concentration. Unless specified, experiments were performed using Qdot 605-streptavidin conjugate that was suspended in the manufacturer-supplied buffer and preincubated for 30 min at RT before application to the sections. In specific experiments, TSA was performed using HRP-conjugated donkey anti-mouse IgG and biotinylated tyramide diluted 5 ug/ml in TSA amplification buffer. After a 10-min TSA reaction, slides were washed three times for 5 min each in PBS and then incubated with fluorophore-conjugated streptavidin, diluted in PBS-BB, for 30 min at RT. Cy3-, FITC-, and Alexa 546-streptavidin conjugates were used at a concentration of 1 μg/ml. Qdot 605- and Qdot 585-streptavidin conjugates were used at a 20-nM concentration and were diluted in manufacturer-supplied buffer. Slides were then washed three times for 5 min in PBS and coverslipped in 90% glycerol-10% PBS.

Imaging Methods

Immunofluorescence samples were examined with a Zeiss-Axioskop microscope equipped with epifluorescence. Sections labeled with Cy3, Alexa 546, Qdot 605, Qdot 585, or Qdot 565 were viewed with a Chroma filter set consisting of an excitation filter of 545/30 nm, dichroic of 575 nm longpass, and an emission filter of 610/75 nm. Sections labeled with FITC or Qdot 525 were viewed with a Chroma filter set consisting of an excitation filter of 470/70 nm, dichroic of 500 nm longpass, and an emission filter of 535/40 nm. Digital images were captured with a Zeiss Axiocam and Axiovision software using identical acquisition parameters per experiment. Multiple X40 fields were captured from each tissue section. Care was taken to ensure that images were obtained from consistent regions (i.e., the cerebral cortex of adult brains or spinal cord of embryos) and that the peak pixel intensity for a specific experiment did not exceed the range of the arbitrary optical density units. Images were saved as TIF files and immunofluorescence was quantified using Scion Image freeware (Scion; Frederick, MD) by measuring the average pixel intensity for the entire image. Values are the mean ± SEM of three or four fields per tissue section.

Statistics

In the photostability study, the normalized pixel intensities at time 0 were compared to later time points using paired Student's t-test comparisons with Bonferroni correction for multiple comparisons.

Results

Optimization of Qdot Concentration

In preliminary experiments, we determined that use of a primary antibody with strong, diffuse uniform staining enabled optimal quantitation of immunofluorescence. Synaptophysin is an integral membrane component of presynaptic vesicle membranes of differentiated neurons and is widely distributed throughout the nervous system (Wiedenmann and Franke 1985). Use of an anti-synaptophysin MAb at a concentration of 100 ng/ml and conventional organic fluorophore detection resulted in diffuse, relatively homogeneous synaptic labeling throughout the adult rat cortex, making synaptophysin immunoreactivity an ideal tool for our experiments. Other mouse MAbs used at an identical concentration, or omission of primary antibody, failed to show the diffuse synaptic staining observed with anti-synaptophysin antibodies (data not shown). To determine the optimal Qdot 605 concentration for immunofluorescence detection, various concentrations of Qdot 605-streptavidin conjugate were added to paraformaldehyde-fixed rat brain sections incubated with either synaptophysin antibodies (final antibody concentration 100 ng/ml) or PBS-BB followed by biotin-conjugated donkey anti-mouse IgG. Differences in signal intensity between the synaptophysin immunolabeling and blocking buffer could be appreciated at 2 nM Qdot 605, the lowest Qdot concentration used. Although there was no significant increase in background staining from Qdot concentrations of 2–20 nM, there was a marked increase in synaptophysin immunofluorescence with increasing Qdot 605 concentration up to 20 nM (Figure 1). An optimal signal-to-noise ratio was achieved at approximately 20 nM Qdot 605, and this concentration was used in subsequent experiments. Similar results were obtained with both frozen and paraffin-embedded tissue sections, and there was no appreciable difference in Qdot 605 detection of synaptophysin between Bouin's- and paraformaldehyde-fixed brain tissue.

As the Qdot 605 concentration increased, we observed a diffuse “haze” in both antibody- and PBS-BB-incubated sections. This nonspecific signal was observed in both paraformaldehyde- and Bouin's-fixed tissue and in both frozen and paraffin-embedded sections. Interestingly, the haze faded markedly with illumination to barely detectable levels. To limit this confounding “nonspecific” effect, all fields were exposed to 4 min of illumination before capture of images by digital camera. No further reduction of background was seen with additional illumination. Furthermore, no loss of specific Qdot 605 labeling occurred with prolonged light exposure (see below).

In sections incubated with higher concentrations of Qdot 605 (40 nM), Qdot-specific, antibody-independent labeling of neuronal cell bodies was observed. This labeling was present even in the absence of secondary antibody, as rehydrated brain sections treated only with 40 nM Qdot 605 showed similar neuronal labeling (Figure 1C). This Qdot-specific, antibody-independent labeling was also observed to a lesser degree with 40-nM concentrations of Qdots 585, 565, and 525 (data not shown). In contrast to the diffuse background staining, which faded after several minutes of illumination, this Qdot-specific neuronal staining did not diminish with prolonged light exposure. The neuronal labeling did not appear to be related to endogenous biotin in the tissue sections because preblocking with avidin and then biotin (Avidin-Biotin Blocking Kit; Vector Laboratories) did not significantly inhibit the Qdot neuronal staining, nor was it observed with organic fluorescent dyes (data not shown). This neuronal staining was observed in both paraffin sections of Bouin's-fixed sections of adult mouse brain and in frozen sections of both paraformaldehyde- and Bouin's-fixed adult rat brain. The Qdot concentration-dependent neuronal staining was more pronounced in tissue sections subjected to antigen retrieval than in standard sections.

Figure 1

Increased fluorescence intensity with increasing concentrations of QDot 605-streptavidin conjugate. Paraformaldehyde-fixed rat brain frozen sections were incubated with mouse anti-synaptophysin (10 ng/ml diluted in PBS-BB) or PBS-BB alone followed by immunodetection with biotinylated donkey anti-mouse IgG and increasing concentrations of Qdot 605. Digital images were obtained from individual fields within the cerebral cortex. Qdot 605 detection (20 nM) of synaptophysin immunoreactivity revealed strong diffuse labeling of synaptic structures throughout the rat cerebral cortex (A). Bars = 50 μm. Omission of primary antibody showed minimal signal at this concentration of Qdot 605 (B). Higher concentrations of Qdot 605 (40 nM) produced intense neuronal cell body staining in the absence of primary and secondary antibodies (C). Bars = 50 μm. The effect of increasing concentrations of Qdot 605-streptavidin conjugate on synap-tophysin-specific and nonspecific signal is graphically depicted in D (synaptophysin antibody in solid lines; PBS-BB control in dashed lines). Values represent the mean of four fields from each Qdot 605 concentration ± SEM.

Increased Sensitivity with Tyramide Signal Amplification

We next determined whether a TSA reaction with biotinylated tyramide deposition would improve the sensitivity of Qdot detection. Bouin's-fixed frozen rat brain sections were treated with increasing concentrations of mouse anti-synaptophysin antibody (0.1–100 ng/ml) followed by incubation with either biotin-conjugated donkey anti-mouse IgG or HRP-conjugated donkey anti-mouse IgG, followed by indirect TSA with biotin tyramide. Both sets of slides were then incubated with Qdot 605-streptavidin conjugate (20 nM) for 30 min and were coverslipped in 90% glycerol-10% PBS.

TSA produced a dramatic increase in Qdot 605 detection of synaptophysin immunoreactivity compared to that achieved with biotinylated secondary antibody detection (Figure 2). Signal-to-noise ratio was dramatically improved and the antibody concentration required to detect specific synaptophysin signal was markedly reduced. Similar results were observed with other mouse monoclonal and rabbit polyclonal antibodies (results not shown). Even with the increased sensitivity afforded by TSA detection, minimal changes in background staining were observed.

Photostability of Quantum Dots Compared with Other Fluorophores

One of the reported advantages of Qdot detection is photostability. To confirm this, Bouin's-fixed, paraffin-embedded 14-day gestation mouse embryo sections were antigen-retrieved and incubated with MAP2 antibodies followed by TSA and detection of biotin-tyramide with the streptavidin conjugates of Cy3 (1 μg/ml), FITC (1 μg/ml), Alexa 546 (1 μg/ml), Qdot 605 (20 nM), or Qdot 585 (20 nM). MAP2 is present at high levels in neuronal cell bodies, axons, and dendrites during early nervous system development (Papandrikopoulou et al. 1989). Therefore, MAP2 immunoreactivity is a useful tool for comparison of different fluorophores. Diffuse MAP2 staining was detected in brain and spinal cord of embryos with all fluorophores. However, continuous illumination resulted in rapid loss of signal in the slides incubated with the conventional organic fluorophores. After 5 min of illumination, less than 15% of the initial pixel intensity remained for sections stained with FITC (ρ<0.01 for all time points) or Alexa 546 (ρ<0.05 for all time points). Approximately 50% of the initial Cy3 pixel intensity remained after 5 min illumination (ρ<0.05) and this was reduced to less than 20% after 20 min of direct illumination (ρ<0.01). In contrast, there was no loss of pixel intensity in brain sections treated with Qdot 605 or Qdot 585 after 20-min illumination (Figure 3). No loss of Qdot signal was observed even after 1 hr of direct light exposure (data not shown). In fact, a slight increase in Qdot pixel intensity (>5% for Qdot 605; >15% for Qdot 585) was observed after 20 min of continuous illumination, with no further increase in pixel intensity after 30 min.

Figure 2

Comparison of biotinylated secondary antibody vs TSA. Bouin's-fixed frozen adult rat brain sections underwent antigen retrieval followed by incubation with PBS-BB or increasing concentrations of mouse anti-synaptophysin antibody. Sections were then incubated with either donkey anti-mouse IgG conjugated to HRP, then biotin tyramide (solid line) or biotinylated donkey anti-mouse IgG (dashed line). Both sets of sections were then incubated with 20 nM Qdot 605-streptavidin conjugate. Digital images were obtained from individual fields within cerebral cortex using a X40 objective. Values represent the mean ± SEM of four fields from each concentration of anti-synaptophysin.

Loss of Qdot Signal After Prolonged PBS Incubation

One of the initial limitations of quantum dots in biological applications was the relative water-insolubility of the nanocrystal. This was overcome by coating the core with a polymer that improved water solubility and hence compatibility with both fixed and living cells (Bruchez et al. 1998; Chan and Nie 1998; Wu et al. 2003). However, special measures are still required for optimal Qdot labeling, including dilution in a manufacturer-supplied buffer and coverslipping in 90% glycerol. We typically perform three 5-min PBS washes after treatment with any fluorophore but had observed loss of Qdot 605 fluorescence in sections that underwent longer PBS washes. To examine this phenomenon systematically, Bouin's- and paraformal-dehyde-fixed frozen rat brain sections were antigen-retrieved and then incubated with mouse anti-synaptophysin antibody (10 ng/ml) overnight at 4C. Sections then underwent TSA using HRP-conjugated donkey anti-mouse IgG and biotin-tyramide. The slides were then incubated with Qdot 605-streptavidin conjugate (20 nM) for 30 min and washed in PBS three times for 5 min. One slide was immediately coverslipped in 90% glycerol and the remaining slides were left in PBS for 1–24 hr before coverslipping. There was a greater than 50% reduction in synaptophysin intensity after 2 hr in PBS and an almost complete loss of signal after 24 hr in PBS (Figure 4). For comparison, tissue sections that had undergone immunodetection with TSA and streptavidin-Cy3 were also subjected to a prolonged PBS wash but showed no loss of fluorescence >24 hr later (data not shown).

Figure 3

Photostability of QDots compared to other fluorophores. Bouin's-fixed paraffin-embedded E14 mouse embryo sections underwent antigen retrieval before serial incubations with mouse anti-MAP2, donkey anti-mouse IgG conjugated to HRP and biotin-tyramide. Sections were then treated with streptavidin conjugates of Qdot 585, Qdot 605, Cy3, Alexa 546, or FITC. Qdots were used at 20 nM concentration diluted in manufacturer-supplied buffer. Remaining fluorophores were diluted 1 μg/ml in PBS-BB. Individual MAP2-positive spinal cord fields were exposed to continuous illumination and digital images were captured at 0, 5, 10, and 20 min. MAP2 staining detected with streptavidin conjugates of Qdot 605 and Cy3 is shown after 0, 5, and 20 min of continuous illumination. Bars = 50 μm. These data are represented graphically for all conjugates in the bottom panel. The average pixel intensity at each time point was normalized to the time 0 pixel intensity for that field. Values represent the mean ± SEM normalized values of three fields from each time point.

Discussion

Quantum dots are high-intensity photostable fluorophores that have several distinct advantages over standard organic fluorescent dyes. Analogously, TSA immunodetection provides increased sensitivity over standard immunohistochemical (IHC) procedures. Our study demonstrates that the combined application of TSA and Qdots results in a dramatic increase in the sensitivity and photostability of IHC fluorescence detection.

The semiconductor nanocrystals that comprise quantum dots have been studied for several decades, but the hurdles of achieving water solubility and conjugation to biomolecules have been overcome only recently (Watson et al. 2003). Initial reports of quantum dots in biological systems used cadmium nanocrystals coated in a silica shell and conjugated to biotin (Bruchez et al. 1998) to label F-actin filaments in mouse 3T3 fibroblasts. In an alternative approach, Chan and Nie (1998) coated the nanocrystals with mercaptoacetic acid to improve water solubility and enable conjugation of transferrin or IgG. The Qdot-streptavidin conjugates used in this study are composed of a cadmium selenium core surrounded by zinc sulfide and a modified polyacrylic acid coat. Qdots have previously been used in the IHC labeling of cell surface receptors, nuclear antigens, and cytoskeletal proteins (Wu et al. 2003).

Photostability is one of the main advantages of Qdot detection. We found no loss of immunofluorescence intensity in Qdot-labeled samples even after 1 hr of continuous illumination, whereas there was nearly complete loss of FITC and Alexa 546 immunofluorescence after 5 min of illumination, and loss of Cy3 labeling within 20 min. On the basis of the manufacturer's recommendation, all Qdot-labeled slides were coverslipped with 90% glycerol/10% PBS. In preliminary experiments we found that this high concentration of glycerol enabled optimal fluorescence detection compared to 50% glycerol/50% PBS. Although antifade mounting media would have undoubtedly improved the stability of the organic dye fluorophores tested, even with antifade media Wu et al. (2003) demonstrated a loss of 50% of Alexa 488 signal after 3 min of illumination. Interestingly, we observed a slight increase in specific Qdot immunofluorescence intensity with initial illumination, which appeared to stabilize over 20–30 min of illumination. The explanation for this phenomenon is unknown.

Figure 4

Loss of QDot signal intensity with prolonged PBS wash. Bouin's-fixed frozen sections of rat brain underwent antigen retrieval followed by incubation with PBS-BB or mouse anti-synapto-physin (10 ng/ml). Sections were serially incubated with donkey anti-mouse IgG conjugated to HRP, followed by biotin-tyramide and then 20 nM Qdot 605-streptavidin conjugate. Slides were washed in PBS three times for 5 min between steps and after the Qdot incubation. Slides either were removed immediately from the final PBS wash and coverslipped in 90% glycerol-10% PBS or were kept in PBS for an additional 1, 2, 3, 8, or 24 hr before coverslipping in 90% glycerol-PBS. Digital images were obtained from individual fields within the cerebral cortex using a X40 objective. The values represent the mean ± SEM of three fields from each time point.

We found linear increases in specific immunostaining with increasing Qdot 605 concentration up to 20 nM, whereas nonspecific background staining did not significantly increase until >20 nM Qdot 605 concentrations were used. During the course of these studies we observed several unusual features of Qdot detection. First, neuronal cell body staining was noted at high concentrations of Qdot 605, which was Qdot-specific and antibody-independent. This neuronal staining was most prominent with Qdot 605-streptavidin conjugate but was also observed in brain sections incubated with 40 nM Qdot 585-, 565-, or 525-streptavidin conjugates. Even though we routinely use Qdot conjugates at concentrations that do not exhibit this effect, i.e., ≤20 nM, we now include a control slide in each experiment that is incubated only with the Qdot conjugate to monitor any Qdot-dependent labeling. Nonspecific staining has been described to be a limitation of quantum dots (Mitchell 2001), and nonspecific staining of mouse kidney nuclei has been demonstrated at higher concentrations of Qdot 605 (Wu 2003).

Second, we often observed a diffuse background “haze” in our Qdot-labeled sections that faded on illumination. Pre-illumination of the field of interest for several minutes dramatically decreased the haze and improved the signal-to-noise ratio. This diffuse photo-bleachable background was much less prominent at lower concentrations of Qdot 605, indicating that this background labeling is also a Qdot concentrationdependent effect.

Although streptavidin-conjugated quantum dots have the distinct advantage of photostability, we did observe loss of signal intensity after prolonged incubation of Qdot-labeled sections in aqueous buffer. This is unlikely to be caused by simple diffusion of streptavidin-conjugated label from the tissue because no loss of signal was observed for streptavidin-Cy3-labeled tissue sections after >24 hr in PBS. Despite modifications to the quantum dot nanocrystal to improve water solubility (Chan et al. 2002; Sutherland 2002), we speculate that Qdots, as currently formulated, may lose stability with prolonged exposure to an aqueous environment.

We have also occasionally observed loss of Qdot 605 signal in frozen brain sections coverslipped with 90% glycerol and stored for several weeks, regardless of whether slides were stored at 4C or at RT. Care was taken to ensure that the tissue sections did not dry because preliminary studies demonstrated dramatically decreased Qdot 605 signal if tissue sections had dried after Qdot labeling. No loss of Qdot 605 signal has been observed in immunostained paraffin-embedded tissue sections coverslipped in 90% glycerol and stored for up to 2 months at 4C. Further studies will be required to determine the stability of Qdots under various slide storage conditions.

This is the first description of the combined application of quantum dots and TSA for IHC detection. The increased sensitivity of the TSA reaction enables use of the primary antibody within a linear range of concentrations and, now that a photostable detection method is available, it may be possible to accurately quantitate antigen levels in tissue sections using the combination of TSA, Qdot detection, and immunofluorescence microscopy.