Context.—Renal tissue emits intense autofluorescence, making it difficult to differentiate specific immunofluorescence signals and thus limiting its application to clinical biopsy material.
Objective.—To identify and minimize autofluorescence of renal tissue and demonstrate a simple, efficient method to reduce autofluorescence using Sudan black B.
Design.—In this study, the sources and features of autofluorescence emitted from kidney tissue were examined. Broad autofluorescence was visualized in both frozen and paraffin kidney sections of normal mice and mice with Adriamycin-induced nephropathy using confocal laser scanning microscopy. Autofluorescence appeared in commonly used 4′,6-diamidino-2-phenylindole, fluorescein isothiocyanate, and Texas Red channels but not in far-red channel, and emitted extensively from red cells, injured tubulointersitial cells, and protein casts in diseased kidney. To eliminate autofluorescence, Sudan black B was used on formaldehyde-fixed paraffin sections and frozen sections of mouse kidney. The effects of Sudan black B in various concentrations were tested on kidney tissue.
Results.—The 0.1% Sudan black B effectively blocked autofluorescence from both paraffin and frozen sections without adversely affecting specific fluorescence signals. Interestingly, the solvent for Sudan black B, 70% ethanol, was also shown to reduce autofluorescence on frozen sections, but not on paraffin sections.
Conclusions.—This study demonstrates a simple, efficient, and cost-effective method to reduce autofluorescence using Sudan black B, and also provides a comprehensive approach to identify and minimize autofluorescence of renal tissue.
Biological autofluorescence (AF), arising from endogenous fluorophores, is an intrinsic property of cells and tissues. The AF properties of some tissues can serve as a useful diagnostic indicator in certain disease situations characterized by accumulation of lipofuscin granules or collagen fibers.1–4 Kidney tissue induces high levels of AF, and the intensity of AF has been used to assess severity of tubular necrosis5 and the degree of damage in ischemic kidney.6–8
However, AF is an obstacle to immunofluorescence (IF) analysis that can either mask or interfere with specific fluorescent signals,9,10 especially those with weak IF labeling, including direct IF and multiple-color fluorescence staining. In addition to intrinsic fluorescence of kidney is AF that arises from the tissue-processing techniques, including fixation agents such as glutaraldehyde11 and embedding material such as paraffin.9,11 In general, AF interferes with routine specific fluorescent labeling by red, green, and blue fluorescent probes, making it very difficult to visualize or colocalize multiple proteins of interest.
As one of the most important advances achieved in optical microscopy, confocal laser scanning microscopy (CLSM) offers several advantages over conventional fluorescence microscopy, including control of depth of field, elimination of image-degrading out-of-focus information, and collection of serial optical sections from thick specimens.12 However, the need for minimal tissue AF is more critical to the success of colocalization studies using CLSM than to those using standard fluorescence microscopy, because fluorescence signals are detected using colorblind photomultiplier tubes with CLSM but color-sensitive charge-coupled device cameras are used with standard fluorescence microscopy.
Among various reagents used to eliminate AF, Sudan black B (SBB) has been proven effective in a range of tissues, including brain,13,14 liver,15 myocardium, and cartilage.11 AF blocking efficacy varies depending on the tissue and the specimen-processing techniques,16 and to date there is no general formula for AF reduction.
Renal tissue emits intense AF, making it difficult to differentiate specific IF signals and thus limiting its application to clinical biopsy material. To the best of our knowledge, whether SBB can effectively reduce AF without diminishing specific IF labeling in both frozen and paraffin-embedded kidney tissue was until now unclear. We herein show reduced AF with SBB in normal and diseased kidneys, enabling detection of specific IF signals.
MATERIALS AND METHODS
Animal and Tissue Sections
Male BALB/c mice (4–6 weeks of age, Centre of Animal Research, Westmead Hospital, Sydney, Australia) were maintained on standard diet and water. The mice were housed in an environmentally controlled room at 23°C under an 8 am–6 pm light/dark cycle. All of the experiments were carried out in accordance with Australian Guide for the Care and Use of Laboratory Animals.
Adriamycin-induced nephropathy was established by Adriamycin injection (10 mg/kg, Pharmacia & Upjohn Pty Ltd., Melbourne, Australia) via tail vein. Four weeks after injection, mice with Adriamycin-induced nephropathy and normal mice were sacrificed and kidney tissues were removed. Some tissues were stored in Tissue-Tek and cut into 5-µm frozen sections. Other tissues were fixed in 4% formaldehyde, and then embedded with paraffin.
Confocal Laser Scanning Microscopy
An Olympus FV 1000 confocal laser scanning microscope with partial spectral capabilities (Olympus, Tokyo, Japan) was used for fluorescence imaging. The optimum settings of laser power and photomultiplier tube voltage were determined in a preliminary experiment based on the AF levels of untreated sections of either normal or diseased mouse kidney and were subsequently used for all SBB-treated sections. Differential interference contrast images were taken simultaneously at excitation of 473 nm. The microscopy conditions are summarized in Table 1. The acquired images were analyzed with FV ASW 1.7b software (Olympus) and Adobe Photoshop 7.0 (Adobe Systems, San Jose, California).
Sections of mouse kidney were stained with hematoxylin for 10 minutes. Subsequently, they were washed under running tap water for 5 minutes, dried on a hot plate, and stained with 0.5% eosin in 96% ethanol for 5 minutes. The sections were rapidly rinsed in 95% ethanol and dehydrated in 2 changes of absolute ethanol for 5 minutes each. Slides were dehydrated, cleared in xylene, and mounted in resinous medium.
Blocking of AF With SBB
Sudan black B (Merck KGaA, Darmstadt, Germany) solutions were made up in 70% ethanol at different concentrations (0%, 0.01%, 0.1%, 0.3%, 0.5%, and 1.0%) in the dark and were filtered. To block AF, unstained and IF-stained sections were incubated in SBB for 25 minutes and were then rinsed with water for 10 minutes. Finally, the sections were mounted with Dako Cytomation fluorescent mounting medium (Dako, Carpinteria, California) and coverslipped as usual for CLSM observation.
Frozen sections were fixed in acetone at −20°C, and paraffin sections were immersed in Histoclear (Emgrid Australia Pty Ltd, The Patch, Victoria, Australia) and dehydrated. For single-color IF labeling, the sections were incubated with primary anti-CD4 antibodies (R&D Systems, Gymea, Australia) directly conjugated with Alexa Fluor 633 or anti-F4/80 conjugated with fluorescein isothiocyanate (FITC; Invitrogen, Mulgrave, Victoria, Australia) in a humid environment at 4°C overnight. After 3 rinses with phosphate-buffered saline (PBS), sections were mounted onto coverslips with Dako Cytomation fluorescent mounting medium.
Double immunolabeling was carried out by incubating the sections for 1 hour with a mixture of 2 primary antibodies, rat anti-F4/80 and goat anti-CD166 antibodies (R&D), followed by incubation with a mixture of anti-rat antibody labeled with FITC and anti-goat antibody labeled with Texas Red (Santa Cruz Biotechnology, Santa Cruz, California) as secondary antibodies. The sections were washed 3 times with PBS, incubated for 1 hour with secondary antibody, and then washed and mounted as above.
Features of AF in Kidney
Autofluorescence in kidney was characterized by visualizing its spatial distribution and by measuring its emission spectra using CLSM. The spectra of AF were then recorded at excitations of 405, 473, and 559 nm but not at 635 nm for the Far red channel (Figure 1, A). Autofluorescence spectra from mouse kidney were broad, occurring in the 3 tested channels (Figure 1, B). To examine the effect of AF on visualization of IF staining with CLSM, renal AF spectra were compared to spectra of 3 representative fluorophores: 4′,6-diamidino-2-phenylindole (DAPI), FITC, and Texas Red. Autofluorescence spectra excited at 405, 473, and 559 nm covered the entire emission spectra of DAPI, FITC, and Texas Red respectively, indicating broader AF emissions than from these 3 fluorophores. Autofluorescence in all the tested channels also exhibited similar emission peaks to those of the 3 representative fluorescent dyes, suggesting that AF interferes unavoidably with measurement of specific fluorescence labeling in the commonly observed channels of DAPI, FITC, and Texas Red.
Confocal images were acquired in DAPI, FITC, Texas Red, and far-red channels at the laser excitation wavelengths of 405, 473, 559, and 635 nm. Autofluorescence was observed in DAPI, FITC, and Texas Red channels in both normal and diseased sections, but stronger AF was seen in diseased (Adriamycin-induced nephropathy) kidney than in normal kidney (Figure 1, B). No detectable fluorescence was visualized in the far-red channel in sections from normal mouse or mouse with Adriamycin-induced nephropathy. In normal mouse kidney tissue, AF was widespread and found in glomeruli and tubules in the above 3 channels. The AF pattern of diseased kidney appeared generally similar to that of normal kidney tissue, with AF distributed throughout. However, bright and irregularly shaped patchy fluorescence was found sprinkled on background AF in diseased renal sections (Figure 1, B).
The Sources of AF
Autofluorescence arose from all renal cells in both normal and diseased kidney under FITC and Texas Red channels as shown in Figure 2. Autofluorescence was brighter in tubule cells than in glomeruli. Autofluorescence was distributed more evenly in normal kidney tissue than in diseased kidney, in which AF showed irregularly shaped granules with high-density fluorescence sprinkled in kidney cortex (arrows in Figure 2, A). These fluorescent granules were located in inflamed tubulointerstitium and were assumed to be injured tubule and interstitial cells. Some disk-shaped fluorescence in glomeruli and interstitium was identified as arising from red blood cells (arrows in Figure 2, B). Large patches with strong fluorescence were seen in dilated tubules from frozen kidney tissue (arrows in Figure 2, C). These bright patches were identified as protein casts using hematoxylin-eosin staining.
Blockade of AF by SBB
The blocking effect of SBB on AF was examined at different concentrations (0.001%–1%) of SBB diluted with 70% ethanol in both frozen and paraffin-embedded kidney tissue (Figure 3). Autofluorescence was reduced in renal tissues excited at 405, 473, and 559 nm in a dose-dependent fashion by SBB. However, AF intensity increased with SBB in high concentration at an excitation wavelength of 633 nm (data not shown).
In both frozen and paraffin kidney sections, 0.1% to 0.5% but not 0.01% SBB reduced AF effectively in all tested CLSM channels. Sudan black B precipitates were seen as black grains on sections treated with SBB at 1% on the differential interference contrast image (Figure 3, A and B). Therefore, 0.1% to 0.5% SBB should be chosen for reduction of background renal AF. In addition, the solvent 70% ethanol was also found to reduce AF moderately on frozen tissue, but not on paraffin sections (data not shown here).
Immunofluorescent staining after SBB blocking with CLSM
As AF was very weak in mouse renal tissue using the far-red channel under confocal microscopy (Figure 1), antibody against CD4 tagged with far-red fluorophore Alexa Fluor 633 was used to detect CD4 antigen in kidney sections. CD4 IF was visualized clearly in the far-red channel without using SBB blocking (Figure 4, A), suggesting a preference for far-red fluorophore labeling for single color IF staining of renal tissue. It was impossible to distinguish specific IF from background AF using anti-F4/80 antibody labeled with FITC without 1% SBB (Figure 4, B). When sections were double stained with F4/80 antibody labeled with FITC and anti-CD166 labeled with phycoerythrin, colocalization could be identified easily as yellow fluorescent signals (Figure 4, C) with 1% of SBB. However, without 1% SBB blockade, clear fluorescence signals indicating colocalization could not be identified with CLSM.
The findings and results of the normal and diseased kidneys obtained in this study are summarized in Table 2.
Immunofluorescence staining with one (direct) or several layers (indirect) is a gold standard technique to evaluate renal biopsy specimens.17 Autofluorescence arising from endogenous fluorophores is an intrinsic property of cells as well as a major obstacle to IF analysis.18 Our results showed that mouse kidneys have strong AF in the commonly observed channels with CLSM, and thus support the importance of methodology to reduce AF.
In this study, murine renal AF of mouse kidney was examined with CLSM. Kidney tissue emitted a broad AF with the emission spectra covering DAPI, FITC, and Texas Red channels, therefore interfering with analysis of fluorescence emission of most routinely used fluorescent probes, such as DAPI/Alexa Fluor 405 (425–525 nm), FITC/Alexa Fluor 488 (485–585 nm), and Texas Red (570–670 nm). Thus, AF makes it difficult to localize specific antigens in cells or tissues by single-fluorescent labeling, and to colocalize by double- or multiple-fluorescent staining, especially when using CLSM. Interestingly, kidney AF was emitted much less in the far-red spectrum (650–750 nm), indicating that it is unnecessary to block AF if using an antibody conjugated with far-red fluorescence. Staining with a single fluorescent antibody such as anti-CD4 antibody tagged with Alexa Fluor 633 showed clear specific IF signals without AF, providing a simple and practical protocol for single color IF staining in renal tissues. However, far-red fluorescence is used less than red, green, or blue fluorescence, because far-red fluorescence is relatively weak as light-scattering intensity drops off with increasing wavelength. Therefore, strategies are required to reduce AF when using red or green fluorescence. Based on the renal AF spectra obtained in this study, which were broader than those of commonly used fluorophores, kidney AF can be separated from specific IF by switching fluorescence microscope filters or adjusting emission bands of a confocal microscope equipped with spectral capabilities to select different fluorescence collection ranges. For example, when the renal fluorescence signal using a green fluorescent antibody is suspected to be nonspecific AF, a channel for red fluorescence can be selected for confirmation. The signal can be confirmed as nonspecific AF if it persists in the red channel.
The pattern of kidney AF differed between normal and diseased kidney: a relatively weak but evenly distributed AF was seen in normal kidney, whereas a stronger AF signal with irregular dots/patches was seen in diseased kidney. As a consequence, AF in diseased kidney can be easier to confuse with specific IF signals that often present as bright fluorescent dots/patches, especially when the signals of interest are dim, such as in direct IF or IF tracking studies. Use of strongly fluorescing antibodies/probes or indirect IF to amplify intensity of specific IF could limit the interference by AF.
Autofluorescence can arise intrinsically or be caused by fixation media used in tissue processing. It has been reported that murine kidney contains a high level of lipofuscin, which is believed to be the main resource of AF generated in normal kidney.16 We found that strong AF came from red blood cells probably related to intracellular rich lipofuscin products, and the formation of conjugated Schiff base compounds from aldehydes derived from lipid peroxidation and amino groups of phospholipids or cell proteins.19 Protein casts are a common histologic feature of diseased kidney. We observed that protein casts in diseased kidney could also contribute to AF. Moreover, reactive or injured renal tubular cells with intracytoplasmic pigmented granules (hemosiderin or lipofuscin) could result in AF, consistent with the results of Ohsaki et al.20
Various methods have been developed to subtract background fluorescence, including ammonia-ethanol, sodium borohydride, SBB, CuSO4,13 or irradiation with light.21 Treatment with chemicals such as CuSO4 can also reduce the intensity of specific immunofluorescent labeling. Sodium borohydride can reduce brilliant AF in erythrocytes that otherwise remains inconspicuous in formaldehyde-fixed tissue.11 Because SBB has the ability to dissolve many types of lipids, including lipofuscin, it serves as an AF suppressor. It has been reported that SBB can mask AF without interfering with the surface fluorescent label in sections of blood vessels, myocardium, and nerve tissue at various concentrations ranging from 0.1 to 1%.11,13,22 In our study, 0.5% SBB was proven effective at minimizing AF in both frozen and paraffin sections of kidney tissue. 0.5% SBB was able to block nearly all AF from commonly used red, green, and blue fluorescent dyes, and was efficient at reducing nonspecific fluorescence signals during single or double staining. Interestingly, 70% ethanol was also effective at reducing AF on frozen but not on paraffin sections of kidney tissue. This effect of ethanol is understandable, as ethanol can degrease and extract lipid and thus remove the fluorescence background induced by lipids. Similarly, improved extraction and dissolution of AF entities by ethanol with ammonia has been reported,11 because ammonia-ethanol may dissolve negatively charged lipid derivatives, phenols, or polyphenols, and degrade weak esters by hydrolysis. However, there was virtually no effect of 70% ethanol on AF on paraffin slides. This may relate to the paraffin-embedding process, which fixes lipid and lipofuscin and renders them difficult to dissolve by ethanol.
In conclusion, we examined the characteristics of AF from murine kidney tissue and blocked renal AF by SBB. This study provides a practical protocol for identifying and eliminating AF during IF staining of murine kidney tissue.
From the Department of Pediatrics, Provincial Hospital Affiliated to Shandong University, Jinan, China (Dr Sun); and the High Tech Centre (Dr Yu) and the Centre for Transplantation and Renal Research (Messrs Zheng and Cao and Drs Ya Wang, Harris, and Yiping Wang), The University of Sydney at Westmead Millennium Institute, Westmead, Australia.
The authors have no relevant financial interest in the products or companies described in this article.