A method is described whereby quantum dot (QD) nanoparticles can be used for correlative immunocytochemical studies of epoxy embedded human pathology tissue. We employ commercial antibody fragment conjugated QDs that are visualized by widefield fluorescence light microscopy and transmission electron microscopy.
A method is described whereby quantum dot (QD) nanoparticles can be used for correlative immunocytochemical studies of human pathology tissue using widefield fluorescence light microscopy and transmission electron microscopy (TEM). To demonstrate the protocol we have immunolabeled ultrathin epoxy sections of human somatostatinoma tumor using a primary antibody to somatostatin, followed by a biotinylated secondary antibody and visualization with streptavidin conjugated 585 nm cadmium-selenium (CdSe) quantum dots (QDs). The sections are mounted on a TEM specimen grid then placed on a glass slide for observation by widefield fluorescence light microscopy. Light microscopy reveals 585 nm QD labeling as bright orange fluorescence forming a granular pattern within the tumor cell cytoplasm. At low to mid-range magnification by light microscopy the labeling pattern can be easily recognized and the level of non-specific or background labeling assessed. This is a critical step for subsequent interpretation of the immunolabeling pattern by TEM and evaluation of the morphological context. The same section is then blotted dry and viewed by TEM. QD probes are seen to be attached to amorphous material contained in individual secretory granules. Images are acquired from the same region of interest (ROI) seen by light microscopy for correlative analysis. Corresponding images from each modality may then be blended to overlay fluorescence data on TEM ultrastructure of the corresponding region.
Correlative light- and electron microscopy (CLEM) is a powerful approach for the analysis of transient dynamic events1, rare events2, 3 and complex systems4. There are many different technical permutations available5 depending on the question being asked however a common requirement is that the same structure in a single sample6 is imaged by multiple microscopy modalities. Our particular approach to CLEM was developed for the study of archival human pathology tissue and the case used here has been well characterized and published previously7. The aim was firstly, to maximize the analytical data from a single biopsy or surgical sample and secondly, to use fluorescence light microscopy to help clarify the context of the immunocytochemical labeling pattern seen at the ultrastructural level.
Quantum dot nanocrystals (QDs) offer the potential of a universal marker system able to be viewed by both, fluorescence light microscopy and electron microscopy8, 9, 10. Their crystalline core structure allows QDs of different sizes to generate a wide range of fluorescence emission peaks when excited by light at wavelengths far from their emission spectra11. Their atomic weight is sufficient to yield electron density that is detectable by transmission electron microscopy, scanning transmission electron microscopy (STEM) or field emission scanning electron microscopy. They are particularly suited to immunocytochemical studies as even single QDs may be observed giving an ultimate sensitivity of one QD per target molecule12. Furthermore, depending on the QD used they can possess an individual elemental signature suitable for mapping.
Human pathology samples offer significant benefits for translational biomedical research. Surgical tissue and biopsy samples are routinely submitted for biobanking and with appropriate ethics clearances can be accessed for research studies. Human tissue does not have issues of relevance or interpretation that can occur in animal or in vitro models of disease. However, specimen preparation of pathology samples often is not optimal. There can be delay in tissue being placed in fixative, inappropriate fixative used such as formalin rather than glutaraldehyde for TEM and inappropriate sampling. CLEM methods have the potential to optimize the diagnostic and prognostic information available from a single human sample. However, some newly developed correlative approaches such as those employing mini Singlet Oxygen Generator (miniSOG) are not available for use in pathology due to the need for the tag to be genetically encoded into the cell of interest13. For this reason we have explored the utility of QD labeling of routinely prepared TEM tissue for correlative immunocytochemical studies. QDs applied to etched epoxy or acrylic resin sections from lightly aldehyde fixed biopsy and tissue samples offer the possibility of obtaining correlative fluorescence light microscopy and TEM data from a single sample.
1. Tissue Dissection and Fixation
2. Tissue Processing and Embedding
3. Ultramicrotomy
4. Immunolabeling
5. Fluorescence Light Microscopy
6. Transmission Electron Microscopy
7. CLEM Imaging
The somatostatinoma tumor specimen used for this study comprised tumor cells forming ductal structures mixed with collagenous stomal tissue. By fluorescence light microscopy, individual tumor cells that contained abundant secretory granules showed positive labeling for the somatostatin hormone. Nuclei appeared as dark holes with minimal non-specific labeling detectable (Figure 1). At low magnifications, variably intense granular orange fluorescence was seen in the cytoplasm of these well characterized tumor cells. The orange color of fluorescence was determined by the size of QD used. For this study we used 585 nm QDs which emit an orange signal when excited with 365 nm light.
CLEM analysis of the same cell was possible (Figure 2). Using the same grid as viewed by light microscopy for TEM examination allowed ROIs to be recognized and imaged in both modalities. The same tissue architecture in relation to grid bars as seen by light microscopy was observed by TEM and used for navigation. The ROI was found by referring to a 20X light microscopy image and noting tissue features in relation to grid bars.
TEM showed the cytoplasm of the somatostatin positive cells to contain abundant round granules of varying electron density. Granules of various diameters were present in individual cells and QD immunolabeling density over the granules was also variable. The overlaying of fluorescence data on TEM images suggested strong positive labeling corresponded to granules containing moderately electron dense material within the vesicle limiting membrane (Figure 3).
At higher magnification the moderately electron dense granules could be seen to be intensely immunolabeled with QD nanocrystals (Figure 4).
Figure 1: Light Microscopy View of an Ultra-thin Section with Somatostatinoma Cells Labeled for Somatostatin Hormone. The ultrathin section was mounted on a TEM grid, immunolabeled, then placed on a glass slide in water under a coverslip. Somatostatinoma cells (arrows) show round nuclei and abundant somatostatin hormone positive granules in the cytoplasm. Considerable variation in labeling density is seen in the somatostatinoma cell population (200X). Please click here to view a larger version of this figure.
Figure 2: CLEM Composite View from an Ultrathin Section Through Somatostatinoma Tissue Labeled for Somatostatin. A) Somatostatinoma cells by widefield fluorescence microscopy (120X); B) Same cells surrounding a lumen (L) seen by TEM at low magnification (570X); C) Higher power TEM view showing brightly fluorescent cell from (A) with nucleus (N) (2,000X); D) Detail of cytoplasmic granules showing intense labeling with QDs over the amorphous material contained within the granule (asterisk) (20,000X). Please click here to view a larger version of this figure.
Figure 3: CLEM Overlay Image. A transparent image from fluorescence microscopy blended with a TEM image of the corresponding area (2,000X). Please click here to view a larger version of this figure.
Figure 4: Transmission Electron Microscopy (TEM) View of Somatostatin Positive Granules from Somatostatinoma Cell Cytoplasm. Somatostatin hormone positive granules in the cytoplasm (arrows) show QD localization over the amorphous contents largely within the limiting membrane. Surrounding cytoplasmic background is clean with few QD probes evident. Some non-specific nuclear labeling is seen. The vesicle membrane surrounding the granules (arrows) is still visible (50,000X). Please click here to view a larger version of this figure.
This study has demonstrated the potential utility of QDs as universal probes for CLEM studies. The 585 nm QD nanoparticles used showed bright and stable fluorescence when viewed by widefield light microscopy and were readily observed by TEM. A previous study by one of the present authors has shown QDs also to be suitable for super-resolution light microscopy7. Their photostability was particularly useful for extended viewing periods and long imaging exposures. QDs can also be used for multiplex immunohistochemistry with different sized probes simultaneously emitting different colored spectra under the same excitation wavelength.
QDs possess sufficient atomic weight to be visualized by TEM but this was found to be challenging depending on the size of the nanoparticle used. For light microscopy, we have found that 525 nm QDs (green) produce less background labeling compared to the larger 585 nm (orange) and 655 nm (red) forms. However, we chose to use the larger 585 nm QD in the present study to enable more convenient visualization of the labeling by TEM. The smaller 525 nm QDs are sized approximately 3 – 5 nm whereas the larger 585 nm forms are approximately 6 – 8 nm and the 655 nm QDs approximately 8 – 10 nm. The 585 nm QDs possessed an irregular crystalline shape with moderate electron density when viewed by TEM. They were not as round or as electron dense as the more traditional colloidal gold immunocytochemical probes. However, QDs have been shown to yield up to 10X more efficient labeling than colloidal gold9 and we have confirmed here that a high labeling density was readily achievable, particularly with the biotin-streptavidin probe linking system. CdSe QDs also possess a characteristic elemental signature that may allow detection and mapping using energy dispersive spectroscopy (EDS), electron energy loss spectroscopy (EELS) or energy filtered transmission electron microscopy (EFTEM)8. Scanning transmission electron microscopy (STEM) systems can also be employed due their exploitation of Z-contrast or atomic number imaging which would enhance contrast between nanoparticles and biological material of lower atomic number14. Large area mapping and visualization of 3D distribution information15 with field emission scanning electron microscopy should also be possible.
A critical step in our method is the antigen unmasking or section etching procedure. We use a single treatment with sodium metaperiodate for 30 min at room temperature. Longer than this and aggregation of structures or loss of membrane definition can occur. Others have used a two-step method involving deplasticizing with sodium methoxide followed by de-osmication with sodium periodate and this may provide more control over the process if necessary16.
CLEM was possible by using light microscopy to carefully log the structural features of ROIs in relation to grid bars. The same grid is then placed in the TEM and orientated so that the grid bars and landmarks correspond to the light microscopy view. A low magnification of approximately 1,400X is most suitable for this. Tissue structures such as ductal formations and patterns of nuclei proved to be most useful for re-finding ROIs in the tissue. However, accurate correlation with same cell resolution can be challenging. Care had to be taken with orientation of the grid to reflect the same orientation seen by light microscopy. We have achieved reasonably accurate subcellular spatial correlation between modalities with overlay images produced using Photoshop. The fluorescence overlay image was blended with the TEM background image however some misalignment was detectable. This may be due to the different imaging systems used by each modality or to radiation damage to our unsupported ultrathin sections in the TEM. Automated systems for storing coordinates of ROIs seen by light microscopy and transferring these to the electron microscope are now being marketed and would greatly simplify this procedure.
The utility of the CLEM approach we have presented for immunocytochemical studies is by necessity limited by the specimen preparation method used, i.e., heavy metal staining and epoxy resin embedding inevitably masking available epitopes. However, the technique does allow the considerable benefit of obtaining fluorescence and TEM data from pathology samples. Obtaining structural and functional information from different microscopy modalities using a single sample is something that has not generally been available to studies of human pathology cells and tissue that were processed primarily for diagnostic purposes.
It would be expected that more sensitive immunocytochemical localization could be obtained from a freeze fixation approach rather than a protocol based on chemical fixation. Adoption of cryofixation principles17 for biobanking of human tissue followed by cryoultramicrotomy and room temperature QD-based immunolabeling as in the classic Tokuyasu technique18, 19 would facilitate sensitive and specific correlative immunocytochemical studies of human disease pathogenesis.
The authors have nothing to disclose.
The authors wish to acknowledge the support of Xiao Juan Wu (Immunohistochemistry Laboratory) and the Department of Anatomical Pathology, Sydney South West Pathology Service (SSWPS), NSW Health Pathology, Liverpool, New South Wales, Australia.
Sodium cacodylate | Proscitech | C0205 | Harmful chemical |
Osmium tetroxide | Proscitech | C010 | Use only in fume hood |
Uranyl acetate | Univar-Ajax | 569 | Hazardous chemical |
Ethanol 100% | Fronine | JJ008 | |
Acetone 100% | Fronine | JJ006 | |
ERL 4221 | Proscitech | C056 | |
DER 732 | Proscitech | C047 | |
NSA | Proscitech | C059 | |
DMAE | Proscitech | C050 | |
Sodium metaperiodate | Analar BDH | 10259 | |
anti-somatostatin antibody | Dako | A0566 | |
Antibody diluent | Dako | S3022 | |
Qdot 585 Streptavidin Conjugate | Invitrogen | Q10113MP | |
Biotinylated goat anti-rabbit IgG antibody | Sigma | B7389-1ML | |
Glutaraldehyde 50% | EMS | 16320 | |
Normal goat serum | Invitrogen | PCN5000 | |
PBS "Dulbecco A" | Oxoid | BR0014G | |
BSAc (10%) | Aurion | 900.022 | |
Parafilm | Pechiney PP | M | |
pH indicator strips (pH 2.0 – 9.0) | Merck | 1.09584.0001 | |
Micromoulds | Proscitech | RL063 | |
Diamond knife | Diatome | Ultra 45 | |
Transmission electron microscope | FEI | Morgagni 268D | |
Fluorescence light microscope | Carl Zeiss | Axioscope A1 | |
Grids 300 mesh nickel (thin bar) | Agar Scientific | G2740N | |
Ultramicrotome | RMC | Powertome | |
TEM camera control software | Soft Imaging System | AnalySIS | Version 3.0 |
Image processing software | Adobe Systems Incorporated | Photoshop CS2 |