Simultaneous double labelling of routinely processed ...

Research in Veterinary Science 88 (2010) 122–126

Contents lists available at ScienceDirect

Research in Veterinary Science journal homepage: www.elsevier.com/locate/rvsc

Simultaneous double labelling of routinely processed paraffin tissue sections using combined immunoperoxidase, immunofluorescence, and digital image editing L. Ressel, A. Poli * Department of Animal Pathology, School of Veterinary Medicine, University of Pisa, Viale delle Piagge 2, 56124, Pisa IT, Italy

a r t i c l e

i n f o

Article history: Accepted 7 July 2009

Keywords: Double immunohistological labelling Image editing Immunofluorescence and immunoperoxidase

a b s t r a c t An innovative image editing system based on a sequential immunoperoxidase–immunofluorescence technique on routine histological sections is described. With this technique it is possible to identify different antigens in different cells, as well as co-localised antigens in the same cell. The method uses digital image editing to mix two independently captured images into one merged image. The technique was performed with indirect immunoperoxidase, followed by sequential indirect immunofluorescence, digital image acquisition and image editing. Multiple staining examples using anti-cytokeratin, anti-vimentin and anti-calbindin antibodies on canine skin and cerebellum, and feline pleural mesothelioma sections were performed in order to investigate the capabilities of the proposed technique. Our data demonstrated that this method can be easily used to assess multiple protein staining studies with minimum laboratory equipment, and that it allows a better structural visualisation of the tissue morphology compared to double immunofluorescence. Moreover, in contrast to double-immunoperoxidase, with this method it is possible to easily co-localise two different antigens in the same cell compartment. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction A demonstration of multiple antigens in the same formalinfixed paraffin-embedded tissue section (FFPES) is often desirable for protein localisation or co-localisation studies. Double immunoperoxidase (IP) is widely used (Mason et al., 1983; Vandesande, 1988; Jeurissen et al., 2000; Ramalho et al., 2006) and enables the simultaneous visualisation of different antigens in the same tissue section, thus achieving a good structural view of the specimen. This technique is time consuming and it is often difficult to discriminate between colours when two antigens are spatially near to each other in the section. Thus this means that it is difficult to detect two co-localised antigens in the same cell (Valnes and Brandtzaeg, 1984; Mason et al., 2000; Borzacchiello and Roperto, 2006). Double or multiple immunofluorescence (IF) have been proposed by some authors (Mason et al., 2000; Bombardi et al., 2006; Borzacchiello and Roperto, 2006; Paciello et al., 2007) as quick techniques that can detect different antigens in the same tissue or co-localised proteins in the same cell. However, they do not give a good architectural view of the whole tissue. A few authors have tried to mix the benefits of these two techniques by performing sequential IP and IF staining on the same section and thus obtaining two separate pictures (Canese and Bussolati, 1977). Some have tried to get one merged microphotograph in which both signals are * Corresponding author. Tel.: +39 050 2216982; fax: +39 050 2216914. E-mail addresses: [email protected] (L. Ressel), [email protected] (A. Poli). 0034-5288/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.rvsc.2009.07.003

detectable using different filters and fluorescent alkaline phosphatase substrates (Lechago et al., 1979; Ramshaw and Parums, 1992; Tao et al., 1994). Recently Masuda et al. developed a simple double IP on routinely processed paraffin sections using AdobeÒ Photoshop software (Masuda et al., 2008). Here we describe an innovative double indirect immunoperoxidase (IIP) – indirect immunofluorescence (IIF) image processing technique on routinely processed paraffin sections in order to identify different antigens in different cells as well as co-localised antigens in the same cell. The technique uses digital image capture and editing to merge the two independently acquired images from the two different methods (IIP, IIF) of the same microscopic field into one composite image. Since feline mesothelioma cells express both cytokeratin and vimentin (Bacci et al., 2006), we used this neoplastic tissue to investigate the possibility of the simultaneous visualization of these two antigens in the same cell.

2. Materials and methods 2.1. Specimens and antibodies Formalin fixed paraffin-embedded tissue samples from canine normal skin (n = 3) and cerebellum (n = 2) were selected from the archives of the department of Animal Pathology at the University of Pisa. A case of feline pleural mesothelioma was also investigated. For each block, four lm thick sections were cut and mounted on

123

L. Ressel, A. Poli / Research in Veterinary Science 88 (2010) 122–126

Superfrost Plus slides (Thermo Scientific, Menzel GmbH & Co., KG, Braunshweig, Germany) and dried overnight at 37 °C. Primary and secondary antibodies used in this study were all purchased from commercial producers (Vector Labs Inc., Burlingame, CA, USA; Dako Inc., Glostrup, Denmark; Novocastra Laboratories Ltd., Newcastle upon Tyne, UK; Millipore, Billerica, MA, USA) and their characteristics as well as the working dilutions used in each experiment are listed in Table 1. 2.2. Experimental design A series of five experiments were carried out on tissue sections to demonstrate the reliability of the proposed technique. All the experiments were done in sequential steps: IIP to detect the first antigen; IIF to detect the second antigen; haematoxylin counterstaining and mounting; digital image acquisition and image editing. Detailed informations on the experiments are presented in Table 2. 2.3. Indirect immunoperoxidase Sections were dewaxed in xylene, passed through a graded series of alcohols, and rehydrated in deionised water. Antigen retrieval (AR) was done with a citrate buffer pH 6.0 in a microwave oven for 15 min at 650W and cooled at room temperature for 20 min. To test the efficacy of the AR procedure for both antigens studied in the double labelling technique, individual IPs were performed and the protocol that gave the best result for both antigens was selected. Endogenous peroxidase was exhausted with 0.5% hydrogen peroxide for 30 min and after that, three washes were performed in 0.05% Tween Tris Buffered Saline solution at pH 7.6 (TBST). Normal serum from the host species of the secondary antibody diluted 1/10 in TBST was added to the sections and incubated for 30 min at room temperature. After three washes, a primary antibody diluted in TBST was applied and incubated for 1 h at room temperature. After three washes, a secondary antibody conjugated with horseradish peroxidase (HRP) was added and incubated for 30 min at room temperature. Tables 1 and 2 report the type, man-

ufacturer and working dilutions of the primary and secondary antibodies used in each experiment. The peroxidase reaction was developed for 10 min using diaminobenzidine following the manufacturer’s instructions (ImpactDAB, Vector Labs Inc., Burlingame, CA, USA) and blocked with deionised water. Negative controls for the target antigens were performed by replacing the primary antibodies with irrelevant antibodies from the host species in which the immunoglobulins were developed (Rabbit immunoglobulins fraction normal X093 and Mouse IgG1 X0931, Dako Inc., Glostrup, Denmark). Negative controls replacing the primary antibody with a buffer solution only (TBST) were also performed. 2.4. Indirect Immunofluorescence Slides with developed peroxidase reactions were washed three times with TBST. The primary antibody for the second antigen was then added and incubated for 1 h at room temperature. After three washes, a secondary antibody conjugated with fluorescein isothiocyanate (FITC) was added and incubated for 30 min at room temperature. Negative controls for the target antigen were performed by replacing the primary antibody with an irrelevant antibody (Rabbit immunoglobulins fraction normal X093 and Mouse IgG1 X0931, Dako Inc., Glostrup, Denmark). Negative controls replacing the primary antibody with a buffer solution only (TBST) were also performed. In some cases (data not shown) sections from the same tissue were stained with IIP for the same antigen to confirm the IIF signal localisation pattern. After three washes in TBST, slides were mounted with a coverslip using a Crystal Mount (Biomeda Corp., Foster City, CA., USA) and air dried for 10 min (Experiments Nos. 1 and 2) or counterstained with Mayer’s haematoxylin for 45 s, washed in deionised water, mounted with a coverslip, and air dried for 10 min (Experiments 3–5). 2.5. Image acquisition Slides were examined using a microscope (Leica DMR, Leica Microsystems, GmbH, Wetzlar, Germany) equipped with a cooled CCD Colour Digital Camera (DS-2M, Nikon, Japan) and epifluorescence

Table 1 Antibodies sources for immunohistochemical tests. Antibody

Antigen

Producer

Id code

Type

Host

Dilution

Conjugation

Primary antibody Primary antibody Primary antibody Secondary antibody Secondary antibody Secondary antibody Secondary antibody

Vimentin Pan cytokeratin Calbindin Rabbit IgG Mouse IgG Rabbit IgG Mouse IgG

Novocastra Novocastra Millipore Vector Vector Dako Vector

V-9 CKp D-28K PI 1000 PI 2000 F0054 SAM

Monoclonal Polyclonal Polyclonal Polyclonal Polyclonal Polyclonal Polyclonal

Mouse Rabbit Rabbit Goat Horse Swine Sheep

1/100 1/100 1/100 1/200 1/200 1/20 1/20

None None None HRP HRP FITC FITC

Refer to Table 2 for the use of each antibody in the experiments; HRP, Horseradish peroxidase; FITC, Fluorescein isothiocyanate; Id code, producer’s identification code or clone.

Table 2 Sequential double immunoperoxidase–immunofluorescence. Xp

Species/tissue

IIP

IIF

I Ab 1 2 3 4 5

Dog/skin Dog/skin Dog/skin Dog/cerebellum Cat/mesothelioma

Mab anti-vim Pab anti-CK Mab anti-vim Pab anti-calbindin Mab anti-vim

II Ab + + + + +

Anti-mouse IgG HRP Anti-rabbit IgG HRP Anti-mouse IgG HRP Anti-rabbit IgG HRP Anti-mouse IgG HRP

H.C.

I Ab + + + + +

Pab anti-CK Mab anti-Vim Pab anti-CK Mab anti-Vim Pab anti-CK

II Ab + + + + +

Anti-rabbit igG FITC Anti-mouse igG FITC Anti-rabbit igG FITC Anti-mouse igG FITC Anti-rabbit igG FITC

None None Yes Yes Yes

Xp, Experiment number; IIP, indirect immunoperoxidase; IIF, indirect immunofluorescence; I Ab, primary antibody; II Ab, secondary antibody. For antibodies details refer to Table 1; HC, haematoxylin counter-stain; Mab, Monoclonal antibody; Pab, Polyclonal antibody; Vim, vimentin; CK, Cytokeratin; HRP, Horseradish peroxidase; FITC, Fluorescein isothiocyanate.

124

L. Ressel, A. Poli / Research in Veterinary Science 88 (2010) 122–126

unit with a FITC filter. A representative field for IIP staining was selected on brightfield light alternating with observation of the same field using a FITC filter, representative of the IIF stain. Indirect immunoperoxidase and IIF stains were considered positive for the presence of a cellular brown precipitate and bright green luminescence in cell structures, respectively. For each representative microscopic field chosen, two images were captured: the first was a colour Red–Green–Blue (RGB) format image captured on brightfield light representative of the IIP stain and, when present, for the haematoxylin counterstain. The second was a monochrome image acquired changing the settings of the microscope to epifluorescence to capture the IIF stain without changing the microscopic field. Images were saved in ‘‘.TIF” format, 2560  1920 pixels width and 8-bit depth. 2.6. Image editing Image files were processed using Adobe Photoshop 6.0 (Adobe, San Jose, CA, USA). Brightfield images were adjusted just in terms of contrast. Epifluorescence monochrome greyscale images captured with an FITC filter were adjusted for the black level to enhance the fluorescence signal over the black background. Black pixels were selected using the ‘‘colour interval” function and removed from the image. The remaining pixels representing the FITC signal were copied and pasted over the brightfield image creating a new level (level one) over the brightfield image (background level). Level one was converted from greyscale to the RGB format and coloured using the Adobe Photoshop’s ‘‘colour balance” function; we used green for experiments 1, 2 and 5, yellow for experiment 3, and red for experiment 4. Finally, the first level (fluorescence image) and the background level (brightfield image) were merged using the ‘‘transparency setting” in order to see a composite image of the two signals. The merging percentage of the two images was considered optimal when both the FITC signal (from level one) and the underlying morphological details and IIP stain of the brightfield image (background level) were both easily visible. When counter-

staining and IIP co-localised signals were present below the FITC level, merging was adjusted to obtain the best simultaneous visualization of the signals. 3. Results An example of the sequential steps of image acquisition and editing is given in Fig. 1. The results of the double IIP-IIF technique are shown in Fig. 2. 4. Discussion and conclusions Indirect immunoperoxidase and sequential IIF successfully stained two different antigens on routine tissue sections without masking or interfering with each other. This finding is also supported by previous similar studies (Canese and Bussolati, 1977; Lechago et al., 1979; Van Vlierberghe et al., 2005) in which double immunolabelling was successfully performed with different markers from cytokeratin and vimentin, but no digital merging of the captured images was performed. In our method, as in double IHC in general, only one AR step is performed for two antigens. In our experience the pH 6.0 cytrate buffer heat induced AR give good results for the majority of antigens, but in some rare cases we suggest to test a different good AR for both antigens in single IHCs before of submitting the samples to the double IHC. In previous studies, counterstaining was not used extensively, but the use of hematoxylin gives better structural details of the specimen and might help in the interpretation of results. Restrictions caused by the use of a filter that is able to capture contemporarily the brightfield light and a fluorescent source, for the simultaneous visualisation of double staining, caused a reduction in one of the two signals in the micrograph in one study (Lechago et al., 1979), but we did not experience this problem with our digital image capture and editing system. Compared to double IP, observing and capturing the two different signals with different optical systems produced no masking

Fig. 1. Summary of the double immunoperoxidase–immunofluorescence and image editing technique. Experiment 5: sequential steps in image acquisition and editing are represented by a flow of images (A–H) and arrows (scale bars = 5 lm). Images in the Figure are from a cat pleural mesothelioma specimen double labelled with vimentin (indirect immunoperoxidase) and cytokeratin (indirect immunofluorescence); (A) image of the specimen with brightfield light (vimentin); (B) image of the same microscopic field as in ‘‘A” observed with a fluorescence FITC filter (cytokeratin); (C) colour RGB image captured with brightfield light (D) monochrome image captured with the fluorescence FITC filter. (E) Editing of the FITC image: black background is removed; (F) digital merging of images C and E. In this composite image there is a background level (from image C) and an additional ‘‘level one” (from image E); (G) conversion to RGB and green colouring of ‘‘level one”; (H) merging of the two levels: the final image is obtained by merging the two levels thus giving an optimal visualization of both structures and double labelling.

L. Ressel, A. Poli / Research in Veterinary Science 88 (2010) 122–126

125

Fig. 2. Digital merging of double immunoperoxidase–immunofluorescence images. A, B, C: Experiment 1: IIP vimentin and IIF cytokerain on dog skin without haematoxylin counterstaining (scale bars = 100 lm). (A) brown vimentin labelling of mesenchymal dermal cells observed with brightfield light; (B) green fluorescent cytokeratin labelling of epidermal cells observed with FITC filter; (C) merged image of ‘‘A” and ‘‘B”. Vimentin labelling is evident as a brown stain in the cytoplasm of mesenchymal dermal cells, while cytokeratin is expressed in the epidermal cell cytoplasm as a green stain. D, E, F: Experiment 2: IIP cytokeratin and IIF vimentin on dog skin without haematoxylin counterstain (scale bars = 100 lm). (D) brown cytokeratin stain in epidermis observed with brightfield light; (E) Green fluorescent vimentin stain in the mesenchymal dermal cells observed with FITC filter; (F) merged image from images ‘‘D” and ‘‘E”. Cytokeratin stain is evident as a brown stain in the cytoplasm of epidermal cells, while vimentin is expressed by mesenchymal dermal cells and melanocytes in the basal layer of epidermis as a green stain. (G) Experiment 3: IIP vimentin and IIF cytokerain on dog skin (scale bar = 20 lm); merged image of follicular adnexa with haematoxylin counterstaining: brown vimentin stain in the mesenchimal dermal cells (arrowhead) and green fluorescent cytokeratin stain in cytoplasm of apocrine (arrow) and sebaceous glands. (H) Experiment 4: IIP vimentin and IIF calbindin on dog cerebellum with haematoxylin counterstaining (scale bar = 20 lm). In the merged image a brown vimentin stain in a capillary endothelial cell (arrowhead) and a red cytoplasmic calbindin stain in a Purkinje cell (arrow) are seen. I, L, M: Experiment 5: IIP vimentin and IIF cytokeratin on feline pleural mesothelioma cells with haematoxylin counterstaining (scale bar = 20 lm). (I) brown vimentin stain in the cytoplasm of mesothelial cells observed under brightfield light. (L) Green fluorescent cytokeratin stain in cytoplasm of mesothelial cells observed with an FITC fluorescence filter. (M) Merged image of ‘‘I” and ‘‘L”. Vimentin stain is evident as a brown stain in the cytoplasm of mesothelial cells, and cytokeratin is present in the cytoplasm of the same cells as a green stain. Most of the cells express both vimentin and cytokeratin and in some cases co-localization of the two markers is evident as a mixture of brown and green stains (arrow).

effects and enabled us to collect all the information from the two different techniques. The use of subsequent image editing may affect the quantitative score of the stain, but in any case this technique, as with IP in general, does not produce strong quantitative results compared with other techniques – in fact the aim of our approach was merely qualitative. Commonly double labelling of tissue sections is performed after individual labelling has identified the localisation of each marker. In the case of image processing as in double IF (Mason et al., 2000) or in our technique the previous demonstration of single IHC localisation of each marker can avoid the risk of producing manipulated results. The image editing software used in this study was Adobe Photoshop as this software is commonly used in a wide range of applications and was used in previous studies (Mason et al., 2000; Masuda et al., 2008; Nga et al., 2008). Other image processing software could be used to obtain the same results such as open source software. Recently Masuda et al. (2008) proposed a IHC technique similar to ours based on a double IP in which pictures were taken after the first and after the second IP (both performed with a DAB stain) and then digitally merged with Photoshop. Our method has some advantages over theirs: we took pictures from the same microscopic field at the same time, thus avoiding the need to find the

same microscopic field as the first IP image to obtain the second IP. Furthermore, our method stains co-localised antigens in the same cells unlike Masuda’s in which the use of DAB in both sequential IP reactions makes co-localisation unfeasible. The possibility of choosing the colour of the FITC level could be useful to enhance the contrast against the brown DAB colour, especially in co-localization studies. In our study we found the green colouring of the IIF signal coupled with the IP image to be an optimal solution for co-localisation studies (Fig. 2M arrow). This hybrid technique seems to have some advantages over double IP and double IF. Our technique is simpler to perform than double IP, and seems to overcome the problem of sequential immunoenzymatic staining not being very feasible for the simultaneous detection of different antigens in the same cell when the two antigens are expressed in the same cell compartment (Ramos-Vara, 2005). Compared to double IF, our technique gives more information about the structure of the examined tissue because of the brightfield component of the image. Four-colour staining combining three fluorescent and one immuno-gold stains on a confocal microscope has been developed (van Vlierberghe et al., 2005) and gave excellent results but required expensive equipment. The equipment for computer-based immunofluorescence and immunoenzymatic labelling is relatively inexpensive compared

126

L. Ressel, A. Poli / Research in Veterinary Science 88 (2010) 122–126

to more sophisticated systems such as a confocal microscope and we agree with some authors (Mason et al., 2000; Borzacchiello and Roperto, 2006) that there are no strong arguments for the use of more expensive equipment other than an epifluorescence microscope and a computer-based image acquisition and editing system to assess double immunolocalisation. In summary our technique make possible to gain an easily qualitative co-localization of two antigens and tissue architecture information from the FFPES specimen analysed and might be useful within a research context when rapid and inexpensive co-localisation of markers coupled with tissue structure and topography are needed. Conflict of interest statement Neither of the authors of this paper has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper. Acknowledgement We would like to thank Dott. Carlo Cantile for his help during the review process and Dott. Emanuele Ricci for providing the anti-calbindin antibody. References Bacci, B., Morandi, F., De Meo, M., Marcato, P.S., 2006. Ten cases of feline mesothelioma: an immunohistochemical and ultrastructural study. Journal of Comparative Pathology 134, 347–354. Bombardi, C., Grandis, A., Chiocchetti, R., Bortolami, R., Johansson, H., Lucchi, M.L., 2006. Immunohistochemical localization of alpha(1a)-adrenoreceptors in muscle spindles of rabbit masseter muscle. Tissue and Cell 38, 121–125. Borzacchiello, G., Roperto, F., 2006. Double-labeling immunofluorescence: an alternative to detect multiple antigens in tissue sections. Veterinary Pathology 43, 83. Canese, M.G., Bussolati, G., 1977. Immuno-electron-cytochemical localization of the somatostatin cells in the human antral mucosa. Journal of Histochemistry and Cytochemistry 25, 1111–1118. Jeurissen, S.H., Claassen, E., Boonstra-Blom, A.G., Vervelde, L., Marga Janse, E., 2000. Immunocytochemical techniques to investigate the pathogenesis of infectious

micro-organisms and the concurrent immune response of the host. Developmental and Comparative Immunology 24, 142–151. Lechago, J., Sun, N.C., Weinstein, W.M., 1979. Simultaneous visualization of two antigens in the same tissue section by combining immunoperoxidase with immunofluorescence techniques. Journal of Histochemistry and Cytochemistry 27, 1221–1225. Mason, D.Y., Abdulaziz, Z., Falini, B., Stein, H., 1983. Single and double immunoenzymatic techniques for labeling tissue sections with monoclonal antibodies. Annals of the New York Academy of Sciences 420, 127– 133. Mason, D.Y., Micklem, K., Jones, M., 2000. Double immunofluorescence labelling of routinely processed paraffin sections. Journal of Pathology 191, 452–461. Masuda, A., Nishikawa, T., Yamamoto, T., Kobayashi, M., 2008a. Simple method for Photoshop-aided double immunohistochemistry – usage of ‘image stack’ function. Histopathology 53, 609–610. Masuda, A., Nishikawa, T., Yamamoto, T., Kobayashi, M., 2008b. Simple method for Photoshop-aided double immunohistochemistry – usage of ‘image stack’ function. Histopathology 53, 609–610. Nga, M.E., Lim, K.H., Tan E.Y., Chan, P., Tan, S.Y., Walford, N. 2008. Malignant adenomyoepithelial tumor of the breast: multi-immunolabeling technique and detailed immunophenotypic study. 16, 100–104. Paciello, O., Shelton, G.D., Papparella, 2007. Expression of major histocompatibility complex class I and class II antigens in canine masticatory muscle myositis. Neuromuscolar Disorders 17, 313–320. Ramalho, L.N., Ribeiro-Silva, A., Cassali, G.D., Zucoloto, S., 2006. The expression of p63 and cytokeratin 5 in mixed tumors of the canine mammary gland provides new insights into the histogenesis of these neoplasms. Veterinary Pathology 43, 424–429. Ramos-Vara, J.A., 2005. Technical aspects of immunohistochemistry. Veterinary Pathology 42, 405–426. Ramshaw, A.L., Parums, D.V., 1992. Combined immunohistochemical and immunofluorescence method to determine the phenotype of proliferating cell populations. Journal of Clinical Pathology 45, 1015–1017. Tao, Q., Srivastava, G., Loke, S.L., Chan, E.Y., Ho, F.C., 1994. Improved double immunohistochemical staining method for cryostat and paraffin wax sections, combining alkaline phosphatase anti-alkaline phosphatase and indirect immunofluorescence. Journal of Clinical Pathology 47, 597–600. Valnes, K., Brandtzaeg, P., 1984. Paired indirect immunoenzyme staining with primary antibodies from the same species. Application of horseradish peroxidase and alkaline phosphatase as sequential labels. The Histochemical Journal 16, 477–487. Van Vlierberghe, R.L., Sandel, M.H., Prins, F.A., Van Iersel, L.B., Van, de Velde C.J., Tollenaar, R.A., Kuppen, P.J., 2005. Four-color staining combining fluorescence and brightfield microscopy for simultaneous immune cell phenotyping and localization in tumor tissue sections. Microscopy Research and Technique 67, 15–21. Vandesande, F., 1988. Double and multiple immunoenzymatic labelling of tissue sections for light microscopy. Acta Histochemica Supplementband 35, 107– 115.