- Email: [email protected]
An optimal antigen retrieval method suitable for different antibodies on human brain tissue stored for several years in formaldehyde fixative
Journal of Neuroscience Methods 72 (1997) 197 – 207
An optimal antigen retrieval method suitable for different antibodies on human brain tissue stored for several years in formaldehyde fixative P. Evers *, H.B.M. Uylings Graduate School Neurosciences Amsterdam, Netherlands Institute for Brain Research, Meibergdreef 33, 1105 AZ Amsterdam, The Netherlands Received 11 July 1996; accepted 9 October 1996
Abstract In the last 5 years the microwave oven has become a widely accepted apparatus for retrieving all kinds of antigens that are masked by prolonged formaldehyde fixation. However, it has so far not been possible to obtain good results for all the antigens we were interested in with just one method. This study offers a solution for this problem. We tested Tris buffered saline (TBS) pH 8–10 in order to see whether a microwave (MW) pretreatment in a solution with a high pH works on the antigens that have our particular interest: MAP-2, non-phosphorylated part of the neurofilament (SMI-32, SMI-311), phosphorylated part of the neurofilament (SMI-312) and the calcium binding proteins Calbindin D28-K, Parvalbumin and Calretinin. Furthermore we checked whether or not a lower temperature of 90°C (instead of boiling the tissue) could be used to achieve the same good staining results. These tests showed that an MW pretreatment in TBS pH 9.0 at full power (boiling) gave the best results. Pretreatment at a lower temperature gave also good results for most (but not all) antibodies when the irradiation time was prolonged to 30 min. The optimal antigen retrieval method also improved the quality of staining with an antibody against neuropeptide Y (NPY), for which, until now, no antigen retrieval appeared to be necessary, even after a long duration of fixation. © 1997 Elsevier Science B.V. Keywords: Calbindin; Calretinin; MAP-2; Microwave oven; NPY; Parvalbumin; SMI-32; SMI-312
1. Introduction The masking of antigens, induced by fixation of tissues in formaldehyde-containing fixatives, is a wellknown problem in both pathology and research laboratories. Different techniques as enzyme predigestion, formic acid pretreatment, microwave-heating, autoclave-heating and ultra-sound pretreatment were developed to overcome this problem (Battiflora and Kopinski, 1986; Kitamoto et al., 1987; Shi et al., 1991; Shin et al., 1991; Podkletnova and Alho, 1993). Comparisons between these different antigen retrieval meth-
* Corresponding author. Tel.: +31 20 5665500; fax + 31 20 6961006. 0165-0270/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 0 2 7 0 ( 9 6 ) 0 2 2 0 4 - 2
ods led to the conclusion that microwave-stimulated antigen retrieval is a good tool in the laboratory for many antigens (Cattoretti et al., 1993; Igarashi et al., 1994; Sherriff et al., 1994). The first procedures were introduced in 1991 by Shi et al., who used toxic heavy metal solutions in a microwave oven to retrieve a number of antigens in routinely formaldehyde-fixed, paraffin-embedded sections (Shi et al., 1991). Since then several authors have described different microwave methods, using less toxic solutions such as buffers, to retrieve all kinds of antigens (Cattoretti et al., 1993; Lucassen et al., 1993; Momose et al., 1993; Munakata and Hendricks, 1993; Suurmeijer and Boon, 1993; Ainley and Ironside, 1994; Evers and Uylings, 1994a,b; Shi et al., 1994; Yachnis and Trojanowski, 1994; Stra¨ter et al., 1995).
198
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
Our group described a method to retrieve antigens in whole tissue blocks, so before vibratome sections were cut, which then could be proceeded, free-floating, according to the standard staining protocols (Evers and Uylings, 1994a). We found that especially the pH of the antigen retrieval solution, the temperature and the irradiation time played an important role in unmasking the epitopes (Evers and Uylings, 1994b). For different antigens however the optimal microwave (MW) pretreatment was different. e.g. for the non-phosphorylated part of the neurofilament (antibody SMI-32) a long irradiation time (60 min) at 90°C in citrate buffer pH 2.5 was optimal, while for Microtubule Associated Protein-2 (MAP-2) a short irradiation time (10 min) at full power (boiling) in citrate buffer pH 4.5 was optimal. However, when it is necessary to apply different immunohistochemical stainings to the same tissue block, it is convenient to have a similar pretreatment procedure. Therefore, it was important to search for one method that would give satisfactory results for all antigens of interest. In 1995 Shi et al. also reported the important role of the pH of the antigen retrieval solution. They concluded that Tris – HCl or sodium acetate buffer at pH 8 – 9 could be suitable for most of the antigens they examined for paraffin sections (Shi et al., 1995). This report induced us to undertake the present study, in which we tried to find one optimal procedure for antigen retrieval for other antibodies frequently used in neuroscience research, such as MAP-2, SMI-32, SMI-311, SMI-312 and the calcium binding proteins calbindin D28-K, parvalbumin and calretinin. We explored different pH values and some different temperatures in combination with different durations of microwaving.
taken out of the brain. These slices were washed for several hours in running tap water and left overnight in one of the following solutions: Tris-buffered saline (TBS) with a pH of 8.0, 8.5, 9.0, 9.5 or 10.0 or 0.05 M citrate buffer with a pH of 2.5. The next morning the slices remained in their solution but were transferred to a plastic jar and put in a microwave oven (Miele Electronic M696). The time setting was 15 min at full power (700 W, boiling, the fluid level was checked every 5 min) or 15, 30 or 60 min. at 90°C (not boiling). The temperature was controlled using the temperature probe of the microwave oven. It took about 3 min to reach a temperature of 90°C. After the MW treatment the jars were allowed to cool for 15 min, after which the tissue slices were put in TBS pH 7.6. Sections, 95 mm thick, were cut on a vibratome and collected in plastic vials containing TBS pH 7.6. These sections were relatively thick because we are interested in staining whole neurons including neuronal processes to distinguish different morphological types. The sections were washed several times in TBS for 1 h and further processed according to our standard immunohistochemical staining procedures.
2. Material and methods
Antibodies
Host
Dilution
MAP-2 (M4403, Sigma, St. Louis, MO) Non-phosph. neurofilament (SMI-32, Sternberger Monoclonals, Baltimore, MD) Non-phosph. neurofilament (SMI-311, Sternberger Monoclonals, Baltimore, MD) Phosph. neurofilament (SMI-312, Sternberger Monoclonals, Baltimore, MD) Calbindin D-28K (C8666, Sigma, St. Louis, MO) Parvalbumin (P3171, Sigma, St. Louis, MO) Calretinin (AB149, Chemicon, Temecula, CA) Neuropeptide Y (Neth. Inst. Brain Res, Amsterdam, Buijs et al., 1989)
M
1:2000
M
1:1000
M
1:2000
M
1:2000
M
1:1000
M
1:2000
R
1:4000
R
1:1000
We started this study on human brain tissue of a 22-year-old female, kindly provided by Prof. I. Kostovic´ of the Croatian Institute for Brain Research in Zagreb. The tissue was collected with the approval of and according to the regulations of the ethical committee of the school of medicine in Zagreb, Croatia. This brain, with a short postmortem interval of about 6 h, had been fixed for 4 years in 4% buffered formaldehyde. On this brain the optimal staining procedure has been developed for the different antibodies studied. Thereafter, this optimal staining procedure has been applied to about 20 human brains ranging in age from 30 weeks of gestation to 90 years old, with postmortem intervals ranging between 4 and 36 h.
2.1. Microwa6e pretreatment Cortical blocks with a thickness of about 5 mm were
2.2. Standard staining procedure To prevent endogenous peroxidase activity the sections were put in a TBS solution containing 3% hydrogen peroxide and 0.2% Triton X-100 for 30 min. After
Table 1 Dilutions of primary antibodies in a TBS solution containing 5% non-fat dried milk and 0.2% Triton X-100
M, mouse, R, rabbit.
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
199
Fig. 1. (a) MAP-2 staining after 4 years of fixation without MW pretreatment. Note that there is absolutely no immunostaining at all. (b) MAP-2 staining of a consecutive tissue slice after an MW pretreatment in TBS pH 9.0, at full power. There are a lot of positively stained neurons in all cortical layers. Bar=300 mm.
a thorough wash in TBS, the sections were put in a TBS solution containing 5% non-fat dry milk and 0.2% Triton X-100 for 1 h to prevent non-specific antibody binding to the tissue. Incubation of the primary antibody, directly after the milk step, took place overnight in a cold room at 4°C. The primary antibodies were diluted in TBS with 5% milk and 0.2%
Triton X-100 (Table 1). After incubation of the primary antibody the sections were thoroughly washed with TBS and incubated for 1 h in the secondary antibody. For the monoclonal antibodies raised in mice (MAP-2, SMI-32, SMI-311, SMI-312, calbindin and parvalbumin) we used peroxidase-conjugated rabbit anti-mouse (1:100 in 5% milk solution with 0.2%
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
200
Table 2 Comparison of different microwave pretreatments Ab
MAP-2 SMI-32 SMI-311 SMI-312 Calb Parv Calret
−MW
− − − ++ ++ ++ −
pH 2.5
8.0
8.5
9.0
9.5
10.0
+ +++ +++ + − − −
++ ++ NT NT ++ ++ ++
+++ +++ NT NT +++ +++ +++
++++ ++++ ++++ ++++ ++++ ++++ ++++
++++ ++++ NT NT ++++ ++++ ++++
+++ ++++ NT NT ++++ +++ ++++
Intensity of immunostaining was scored as −, no staining at all; +, very weak; ++, good staining of somata but weakly stained arborizations; +++, good staining but not in all cortical layers; ++++, optimal staining. Ab, antibody; −MW, no microwave pretreatment; NT, not tested.
Triton) as secondary antibody, after which visualization could take place. For the polyclonal antibodies against calretinin and neuropeptide Y, which were raised in rabbits, we used goat anti-rabbit (1:100 in milk-solution) as secondary antibody, followed (after a TBS-wash) by a tertiary antibody, peroxidase anti-peroxidase (PAP, 1:1000 in milk-solution). Visualization took place in 0.05% diaminobenzidine (DAB) enhanced with 0.2% nickel-ammonium sulphate.
3. Results The antigen retrieval technique was performed on 0.5 cm thick tissue slices. As we know from a former study it is not possible to apply the microwave-stimulated antigen retrieval technique on free-floating vibratome sections (Evers and Uylings, 1994a). Under none of the conditions mentioned below did we find a difference in immunostaining intensity between superficial and deep parts of the slices. Penetration of all the antibodies studied was through the whole section thickness of 95 mm.
3.2. Influence of the pH MW-pretreated tissue in an antigen retrieval solution pH 8.0–10.0 for 15 min at full power (boiling) showed a remarkable increase in immunostaining as compared to the staining results achieved with the standard protocol without pretreatment (Table 2). The best pHs appeared to be 9.0 and 9.5. In these cases immunostainings of all examined antibodies showed both well-stained somata and well-stained dendrites in all cortical layers as well as a very good axonal staining in the case of SMI-312 (Figs. 1–7). TBS, pH 10.0, in general also gave good staining results but it was difficult to cut good sections on the vibratome because the tissue slices had become softer during this pretreatment. Lower pHs (8.0 and 8.5) resulted in a weaker staining, especially in the superficial cortical layers and layer VI. The results obtained with MW pretreatment in citrate buffer pH 2.5 were comparable with those of our previous study (Evers and Uylings, 1994b). SMI-32 and SMI-311 showed a rather good staining of somata as well as dendrites. All the other stainings, however, had not improved at all at pH 2.5. SMI-312, calbindin and parvalbumin were even more weakly stained than without MW pretreatment.
3.1. Immunostaining without MW pretreatment Tissue, fixed for 4 years and immunostained according to our standard procedures without MW pretreatment, gave negative (Fig. 1a) or very poor (Fig. 3a and Fig. 4a) results (Table 2). Only SMI-312, an axonal marker, showed a moderate staining of axons. Calbindin and parvalbumin showed moderately stained somata and very few, faintly stained dendrites.
3.3. Effect of 15, 30 and 60 min MW irradiation at 90 °C In general an MW pretreatment at a lower temperature gave a lower background staining as compared to a pretreatment at full power. However, the specific staining also seemed to be less intensive, especially for the calcium binding proteins parvalbumin and calretinin. Even when the irradiation time was prolonged
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
201
Fig. 2. (a) SMI-32 immunostaining and (b) SMI-311 immunostaining. Both stainings are executed after an MW pretreatment in TBS pH 9.0, at full power after 4 years of fixation. Without MW pretreatment, there is no immunostaining at all. Bar= 300 mm.
to 1 h no improvement in the intensity was noted. Somata were well stained but small dendrites were often not visible when a lower temperature was used. On the other hand the Microtubule Associated Protein-2, the phosphorylated and non-phosphorylated parts of the neurofilament (SMIs) and calbindin stained well at 90°C when the irradiation time was prolonged to 30 or 60 min.
In general we conclude that an MW pretreatment of 15 min at full power (boiling) gives optimal staining results for all tested antibodies. This optimal procedure has been applied sofar on brain tissue of 20 subjects. They all showed an improved quality of staining as compared with the previous procedures for microwave stimulated antigen retrieval, especially the calcium binding proteins (Table 2). These improve-
202
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
Fig. 3. (a) Calbindin immunostaining without MW pretreatment after 4 years of fixation. Only the somata of the strongly positive cells are stained. (b) Calbindin after MW pretreatment of a consecutive tissue slice in TBS pH 9.0 at full power. Besides the somata also many cells with nicely stained arborizations are visible. Bar= 300 mm.
ments were similar to those mentioned above. Postmortem intervals does not seem to influence the results, for the antibodies studied, as long as the interval was kept below the 18 h. This optimal MW antigen retrieval procedure appears to be also optimal for an antigen, resistant to periods of a few years of fixation, such as Neuropeptide Y. Immunostaining after MW pretreatment showed even more intensely
stained somata and dendrites than without pretreatment (Fig. 7d).
4. Discussion The effects of fixation in formaldehyde solutions are due to the formation of methylene bridges (crosslinking) between proteins. These changes in secondary
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
203
Fig. 4. (a) Parvalbumin immunostaining without MW pretreatment after 4 years of fixation. The somata of many parvalbumin positive cells are stained, but hardly any dendrites are visible. (b) Parvalbumin staining of a consecutive tissue slice after an MW pretreatment in TBS pH 9.0 at full power. Many fibers are visible in all cortical layers and more somata in layer III. Bar= 300 mm.
and tertiary structure of the proteins are the reason why many antibodies do not recognize an antigen or why the antigen is not accessible for the antibody (Fox et al., 1985; Mason and O’Leary, 1991; Dapson, 1993). Although the exact mechanism of antigen retrieval is not yet understood, it is believed that the cross links between proteins are broken, which makes the antigen accessible or recognizible for the antibody.
Although this process takes place when the tissue or sections are washed in distilled water for a very long time (several weeks, Helander, 1994) a combination of immersion in a solution with a high pH and high temperatures will accelerate the process. We have shown in a former study that for non-phosphorylated neurofilament an MW-pretreatment in a solution with a low pH (pH 2.5) gives very satisfactory results but
204
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
not at pH 4.5 (Evers and Uylings, 1994b). Many other researchers described that a more neutral pH (pH 6) gave an optimal staining result in their studies of other antigens. In this study we showed that an MW pretreatment of a tissue block in a simple buffer with a high pH of 9.0 – 9.5 (but not pH 8.5 or 8.0) gives good results in immunostainings on vibratome
sections with seven antibodies frequently used in neuroscience research. This extends the results described recently by Shi et al. (1995), who applied the technique for the detection of other antigens on mounted paraffin sections and came to the conclusion that a buffer with a high pH (pH 8–9) gave good results in all their tests. Another important factor in antigen retrieval techniques is high temperature. According to the literature there are different possibilities to achieve this high temperature, i.e. besides microwave-heating, a combination of microwave and pressure cooker, conventional heating or autoclave (Shi et al., 1991; Shin et al., 1991; Igarashi et al., 1994). Many good results were achieved with all these techniques. In our situation the microwave oven is most convenient to achieve high temperatures. It appears possible to obtain good results by microwaving at full power (boiling) and at a lower temperature of 90°C. However, when using this lower temperature the time of pretreatment should be lengthened to obtain in general the same results as with higher temperatures (boiling). However, for a few calcium binding protein antibodies, parvalbumin and calretinin, no optimal staining was obtained at 90°C. Therefore, we prefer the high temperature of boiling at full power. We have applied the optimal technique of MW pretreatment in pH 9.0 solution at full power for 15 min on about 20 patients with excellent results. Also paraffin sections instead of whole tissue blocks were successfully stained for SMI-32 and SMI-311 with this technique. However, when fixation times were extremely long (8 years and longer) it was impossible to achieve optimally stained cells. On the other hand paraffin sections of human brain tissue fixed for 1 year in Bodian fixative (a formaldehyde-containing solution) stained rather well with SMI-32 antibody after the optimal pretreatment procedure, whereas no immunostaining was obtained without pretreatment. In conclusion, we obtained one antigen retrieval procedure which is optimal for all the antibodies used in this study. This enlarges greatly the staining of important and irreplaceable human tissues and sections that were believed to have been fixed for too long.
Acknowledgements
Fig. 5. Calretinin immunostaining after 4 years of fixation with MW pretreatment in TBS pH 9.0 at full power. Without MW pretreatment there is absolutely no immunostaining. Bar= 300 mm.
The authors would like to thank Mr. G. van der Meulen for his photographical assistance and W.T.P. Verweij for her secretarial assistance and for correcting the English.
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
205
Fig. 6. Detail of (a) MAP-2 in layer V, (b) SMI-32 in layer V, (c) SMI-311 in layer III, and (d) SMI-312 in layers I – III. All immunostainings performed after 4 years of fixation with MW pretreatment in TBS pH 9.0 at full power. Bar= 50 mm.
206
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
Fig. 7. Detail of (a) calbindin in layer II, (b) parvalbumin in layer V, (c) calretinin in layer II, and (d) NPY in layer V. All immunostainings performed after 4 years of fixation with MW pretreatment in TBS pH 9.0 at full power. Bar= 50 mm.
References Ainley, C.D. and Ironside, J.W. (1994) Microwave technology in diagnostic neuropathology, J. Neurosci. Methods, 55: 183 – 190.
Battiflora, H. and Kopinski, M. (1986) The influence of protease digestion and duration of fixation on the immunostaining of keratins, J. Histochem. Cytochem., 34: 1095 – 1100. Buijs, R.M., Pool, C.W., Van Heerikhuize, J.J., Sluiter, A.A., Van der Sluis, P.J., Ramkema, M., Van der Woude, T.P. and Van
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207 der Beek, E. (1989) Antibodies to small transmitter molecules and peptides: production and application of antibodies to Dopamine, Serotonine, GABA, Vasopressin, Vasoactive Intestinal Peptide, Neuropeptide Y, Somatostatin and Substance P. Biomed. Res., 10: 213 – 221. Cattoretti, G., Pileri, S., Parravicini, C., Becker, M.H.G., Poggi, S., Bifulco, C., Key, G., D’Amato, L., Sabattini, E., Feudale, E., Reynolds, F., Gerdes, J. and Rilke, F. (1993) Antigen unmasking on formalin-fixed, paraffin-embedded tissue sections, J. Pathol., 171: 83 – 98. Dapson, R.W. (1993) Fixation for the 1990s: a review of needs and accomplishments, Biotechnol. Histochem., 68: 75–82. Evers, P. and Uylings, H.B.M. (1994a) Effects of microwave pretreatment on immunocytochemical staining of vibratome sections and tissue blocks of human cerebral cortex stored in formaldehyde fixative for long periods, J. Neurosci. Methods, 55: 163 – 172. Evers, P. and Uylings, H.B.M. (1994b) Microwave-stimulated antigen retrieval is pH and temperature dependent, J. Histochem. Cytochem., 42: 1555–1563. Fox, C.H., Johnson, F.B., Whiting, J. and Roller, P.P. (1985) Formaldehyde fixation, J. Histochem. Cytochem., 33: 845 – 853. Helander, K.G. (1994) Kinetic studies of formaldehyde binding in tissue, Biotechnol. Histochem., 69: 177–179. Igarashi, H., Sugimura, H., Maruyama, K., Kitayama, Y., Ohta, I., Suzuki M., Tanaka, M., Dobashi, Y. and Kino, I. (1994) Alteration of immunoreactivity by hydrated autoclaving, microwave treatment, and simple heating of paraffin-embedded tissue sections. APMIS, 102: 295–307. Kitamoto, T., Ogomori, K., Tateishi, J. and Prusiner, S.B. Formic acid pretreatment enhances immunostaining of cerebral and systemic amyloids, Lab. Invest., 57: 230–236. Lucassen, P.J., Ravid, R., Gonatas, N.K. and Swaab, D.F. (1993) Activation of the human supraoptic and paraventricular nucleus neurons with aging and in Alzheimer’s disease as judged from increasing size of the Golgi apparatus, Brain Res., 632: 105 – 113. Mason, J.T. and O’Leary, T.J. (1991) Effects of formaldehyde fixation on protein secondary structure: a calorimetric and infrared spectroscopic investigation, J. Histochem. Cytochem., 39: 225 – 229. Momose, H., Metha, P. and Battiflora, H. (1993) Antigen retrieval by
.
207
microwave irradiation in lead thiocyanate, Appl. Immunochem., 1: 77 – 82. Munakata, S. and Hendricks, J.B. Effect of fixation time and microwave oven heating time on retrieval of the Ki-67 antigen from paraffin-embedded tissue, J. Histochem. Cytochem., 41: 1241– 1246. Podkletnova, I. and Alho, H. (1993) Ultrasound-amplified immunohistochemistry, J. Histochem. Cytochem., 41: 51 – 56. Sherriff, F.E., Bridges, L.R. and Jackson, P. (1994) Microwave antigen retrieval of b-amyloid precursor protein immunoreactivity. Neuroreport, 5: 1085 – 1088. Shi, S.-R., Key, M.E. and Kalra, K.L. (1991) Antigen retrieval in formalin-fixed, paraffin-embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections, J. Histochem. Cytochem., 39: 741 – 748. Shi, S.-R., Chaiwun, B., Young, L., Imam, A., Cote, R.J. and Taylor, C.R. (1994) Antigen retrieval using pH 3.5 glycine-HCl buffer or urea solution for immunohistochemical localization of Ki-67, Biotechnol. Histochem., 69: 213 – 215. Shi, S.-R., Imam, A., Young, L., Cote, R.J. and Taylor, C.R. (1995) Antigen retrieval immunohistochemistry under influence of pH using monoclonal antibodies, J. Histochem. Cytochem., 43: 193– 201. Shin, R.-W., Iwaki, T., Kitamoto, T. and Tateishi, J. (1991) Hydrated autoclave pretreatment enhances TAU immunoreactivity in formalin-fixed normal and Alzheimer’s disease brain tissues, Lab. Invest., 64: 693 – 702. Stra¨ter, J., Gu¨nthert, A.R., Bru¨derlein, S. and Mo¨ller, P. (1995) Microwave irradiation of paraffin-embedded tissue sensitizes the TUNEL method for in situ detection of apoptotic cells, Histochemistry, 103: 157 – 160. Suurmeijer, A.J.H. and Boon, M.E. (1993) Optimizing keratin and vimentin retrieval in formalin-fixed, paraffin-embedded tissue with the use of heat and metal salts, Appl. Immunohistochem., 1: 143 – 148. Yachnis, A.T. and Trojanowski, J.Q. (1994) Studies of childhood brain tumors using immunohistochemistry and microwave technology: methodological considerations, J. Neurosci. Methods, 55: 191 – 200.
An optimal antigen retrieval method suitable for different antibodies on human brain tissue stored for several years in formaldehyde fixative P. Evers *, H.B.M. Uylings Graduate School Neurosciences Amsterdam, Netherlands Institute for Brain Research, Meibergdreef 33, 1105 AZ Amsterdam, The Netherlands Received 11 July 1996; accepted 9 October 1996
Abstract In the last 5 years the microwave oven has become a widely accepted apparatus for retrieving all kinds of antigens that are masked by prolonged formaldehyde fixation. However, it has so far not been possible to obtain good results for all the antigens we were interested in with just one method. This study offers a solution for this problem. We tested Tris buffered saline (TBS) pH 8–10 in order to see whether a microwave (MW) pretreatment in a solution with a high pH works on the antigens that have our particular interest: MAP-2, non-phosphorylated part of the neurofilament (SMI-32, SMI-311), phosphorylated part of the neurofilament (SMI-312) and the calcium binding proteins Calbindin D28-K, Parvalbumin and Calretinin. Furthermore we checked whether or not a lower temperature of 90°C (instead of boiling the tissue) could be used to achieve the same good staining results. These tests showed that an MW pretreatment in TBS pH 9.0 at full power (boiling) gave the best results. Pretreatment at a lower temperature gave also good results for most (but not all) antibodies when the irradiation time was prolonged to 30 min. The optimal antigen retrieval method also improved the quality of staining with an antibody against neuropeptide Y (NPY), for which, until now, no antigen retrieval appeared to be necessary, even after a long duration of fixation. © 1997 Elsevier Science B.V. Keywords: Calbindin; Calretinin; MAP-2; Microwave oven; NPY; Parvalbumin; SMI-32; SMI-312
1. Introduction The masking of antigens, induced by fixation of tissues in formaldehyde-containing fixatives, is a wellknown problem in both pathology and research laboratories. Different techniques as enzyme predigestion, formic acid pretreatment, microwave-heating, autoclave-heating and ultra-sound pretreatment were developed to overcome this problem (Battiflora and Kopinski, 1986; Kitamoto et al., 1987; Shi et al., 1991; Shin et al., 1991; Podkletnova and Alho, 1993). Comparisons between these different antigen retrieval meth-
* Corresponding author. Tel.: +31 20 5665500; fax + 31 20 6961006. 0165-0270/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 0 2 7 0 ( 9 6 ) 0 2 2 0 4 - 2
ods led to the conclusion that microwave-stimulated antigen retrieval is a good tool in the laboratory for many antigens (Cattoretti et al., 1993; Igarashi et al., 1994; Sherriff et al., 1994). The first procedures were introduced in 1991 by Shi et al., who used toxic heavy metal solutions in a microwave oven to retrieve a number of antigens in routinely formaldehyde-fixed, paraffin-embedded sections (Shi et al., 1991). Since then several authors have described different microwave methods, using less toxic solutions such as buffers, to retrieve all kinds of antigens (Cattoretti et al., 1993; Lucassen et al., 1993; Momose et al., 1993; Munakata and Hendricks, 1993; Suurmeijer and Boon, 1993; Ainley and Ironside, 1994; Evers and Uylings, 1994a,b; Shi et al., 1994; Yachnis and Trojanowski, 1994; Stra¨ter et al., 1995).
198
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
Our group described a method to retrieve antigens in whole tissue blocks, so before vibratome sections were cut, which then could be proceeded, free-floating, according to the standard staining protocols (Evers and Uylings, 1994a). We found that especially the pH of the antigen retrieval solution, the temperature and the irradiation time played an important role in unmasking the epitopes (Evers and Uylings, 1994b). For different antigens however the optimal microwave (MW) pretreatment was different. e.g. for the non-phosphorylated part of the neurofilament (antibody SMI-32) a long irradiation time (60 min) at 90°C in citrate buffer pH 2.5 was optimal, while for Microtubule Associated Protein-2 (MAP-2) a short irradiation time (10 min) at full power (boiling) in citrate buffer pH 4.5 was optimal. However, when it is necessary to apply different immunohistochemical stainings to the same tissue block, it is convenient to have a similar pretreatment procedure. Therefore, it was important to search for one method that would give satisfactory results for all antigens of interest. In 1995 Shi et al. also reported the important role of the pH of the antigen retrieval solution. They concluded that Tris – HCl or sodium acetate buffer at pH 8 – 9 could be suitable for most of the antigens they examined for paraffin sections (Shi et al., 1995). This report induced us to undertake the present study, in which we tried to find one optimal procedure for antigen retrieval for other antibodies frequently used in neuroscience research, such as MAP-2, SMI-32, SMI-311, SMI-312 and the calcium binding proteins calbindin D28-K, parvalbumin and calretinin. We explored different pH values and some different temperatures in combination with different durations of microwaving.
taken out of the brain. These slices were washed for several hours in running tap water and left overnight in one of the following solutions: Tris-buffered saline (TBS) with a pH of 8.0, 8.5, 9.0, 9.5 or 10.0 or 0.05 M citrate buffer with a pH of 2.5. The next morning the slices remained in their solution but were transferred to a plastic jar and put in a microwave oven (Miele Electronic M696). The time setting was 15 min at full power (700 W, boiling, the fluid level was checked every 5 min) or 15, 30 or 60 min. at 90°C (not boiling). The temperature was controlled using the temperature probe of the microwave oven. It took about 3 min to reach a temperature of 90°C. After the MW treatment the jars were allowed to cool for 15 min, after which the tissue slices were put in TBS pH 7.6. Sections, 95 mm thick, were cut on a vibratome and collected in plastic vials containing TBS pH 7.6. These sections were relatively thick because we are interested in staining whole neurons including neuronal processes to distinguish different morphological types. The sections were washed several times in TBS for 1 h and further processed according to our standard immunohistochemical staining procedures.
2. Material and methods
Antibodies
Host
Dilution
MAP-2 (M4403, Sigma, St. Louis, MO) Non-phosph. neurofilament (SMI-32, Sternberger Monoclonals, Baltimore, MD) Non-phosph. neurofilament (SMI-311, Sternberger Monoclonals, Baltimore, MD) Phosph. neurofilament (SMI-312, Sternberger Monoclonals, Baltimore, MD) Calbindin D-28K (C8666, Sigma, St. Louis, MO) Parvalbumin (P3171, Sigma, St. Louis, MO) Calretinin (AB149, Chemicon, Temecula, CA) Neuropeptide Y (Neth. Inst. Brain Res, Amsterdam, Buijs et al., 1989)
M
1:2000
M
1:1000
M
1:2000
M
1:2000
M
1:1000
M
1:2000
R
1:4000
R
1:1000
We started this study on human brain tissue of a 22-year-old female, kindly provided by Prof. I. Kostovic´ of the Croatian Institute for Brain Research in Zagreb. The tissue was collected with the approval of and according to the regulations of the ethical committee of the school of medicine in Zagreb, Croatia. This brain, with a short postmortem interval of about 6 h, had been fixed for 4 years in 4% buffered formaldehyde. On this brain the optimal staining procedure has been developed for the different antibodies studied. Thereafter, this optimal staining procedure has been applied to about 20 human brains ranging in age from 30 weeks of gestation to 90 years old, with postmortem intervals ranging between 4 and 36 h.
2.1. Microwa6e pretreatment Cortical blocks with a thickness of about 5 mm were
2.2. Standard staining procedure To prevent endogenous peroxidase activity the sections were put in a TBS solution containing 3% hydrogen peroxide and 0.2% Triton X-100 for 30 min. After
Table 1 Dilutions of primary antibodies in a TBS solution containing 5% non-fat dried milk and 0.2% Triton X-100
M, mouse, R, rabbit.
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
199
Fig. 1. (a) MAP-2 staining after 4 years of fixation without MW pretreatment. Note that there is absolutely no immunostaining at all. (b) MAP-2 staining of a consecutive tissue slice after an MW pretreatment in TBS pH 9.0, at full power. There are a lot of positively stained neurons in all cortical layers. Bar=300 mm.
a thorough wash in TBS, the sections were put in a TBS solution containing 5% non-fat dry milk and 0.2% Triton X-100 for 1 h to prevent non-specific antibody binding to the tissue. Incubation of the primary antibody, directly after the milk step, took place overnight in a cold room at 4°C. The primary antibodies were diluted in TBS with 5% milk and 0.2%
Triton X-100 (Table 1). After incubation of the primary antibody the sections were thoroughly washed with TBS and incubated for 1 h in the secondary antibody. For the monoclonal antibodies raised in mice (MAP-2, SMI-32, SMI-311, SMI-312, calbindin and parvalbumin) we used peroxidase-conjugated rabbit anti-mouse (1:100 in 5% milk solution with 0.2%
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
200
Table 2 Comparison of different microwave pretreatments Ab
MAP-2 SMI-32 SMI-311 SMI-312 Calb Parv Calret
−MW
− − − ++ ++ ++ −
pH 2.5
8.0
8.5
9.0
9.5
10.0
+ +++ +++ + − − −
++ ++ NT NT ++ ++ ++
+++ +++ NT NT +++ +++ +++
++++ ++++ ++++ ++++ ++++ ++++ ++++
++++ ++++ NT NT ++++ ++++ ++++
+++ ++++ NT NT ++++ +++ ++++
Intensity of immunostaining was scored as −, no staining at all; +, very weak; ++, good staining of somata but weakly stained arborizations; +++, good staining but not in all cortical layers; ++++, optimal staining. Ab, antibody; −MW, no microwave pretreatment; NT, not tested.
Triton) as secondary antibody, after which visualization could take place. For the polyclonal antibodies against calretinin and neuropeptide Y, which were raised in rabbits, we used goat anti-rabbit (1:100 in milk-solution) as secondary antibody, followed (after a TBS-wash) by a tertiary antibody, peroxidase anti-peroxidase (PAP, 1:1000 in milk-solution). Visualization took place in 0.05% diaminobenzidine (DAB) enhanced with 0.2% nickel-ammonium sulphate.
3. Results The antigen retrieval technique was performed on 0.5 cm thick tissue slices. As we know from a former study it is not possible to apply the microwave-stimulated antigen retrieval technique on free-floating vibratome sections (Evers and Uylings, 1994a). Under none of the conditions mentioned below did we find a difference in immunostaining intensity between superficial and deep parts of the slices. Penetration of all the antibodies studied was through the whole section thickness of 95 mm.
3.2. Influence of the pH MW-pretreated tissue in an antigen retrieval solution pH 8.0–10.0 for 15 min at full power (boiling) showed a remarkable increase in immunostaining as compared to the staining results achieved with the standard protocol without pretreatment (Table 2). The best pHs appeared to be 9.0 and 9.5. In these cases immunostainings of all examined antibodies showed both well-stained somata and well-stained dendrites in all cortical layers as well as a very good axonal staining in the case of SMI-312 (Figs. 1–7). TBS, pH 10.0, in general also gave good staining results but it was difficult to cut good sections on the vibratome because the tissue slices had become softer during this pretreatment. Lower pHs (8.0 and 8.5) resulted in a weaker staining, especially in the superficial cortical layers and layer VI. The results obtained with MW pretreatment in citrate buffer pH 2.5 were comparable with those of our previous study (Evers and Uylings, 1994b). SMI-32 and SMI-311 showed a rather good staining of somata as well as dendrites. All the other stainings, however, had not improved at all at pH 2.5. SMI-312, calbindin and parvalbumin were even more weakly stained than without MW pretreatment.
3.1. Immunostaining without MW pretreatment Tissue, fixed for 4 years and immunostained according to our standard procedures without MW pretreatment, gave negative (Fig. 1a) or very poor (Fig. 3a and Fig. 4a) results (Table 2). Only SMI-312, an axonal marker, showed a moderate staining of axons. Calbindin and parvalbumin showed moderately stained somata and very few, faintly stained dendrites.
3.3. Effect of 15, 30 and 60 min MW irradiation at 90 °C In general an MW pretreatment at a lower temperature gave a lower background staining as compared to a pretreatment at full power. However, the specific staining also seemed to be less intensive, especially for the calcium binding proteins parvalbumin and calretinin. Even when the irradiation time was prolonged
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
201
Fig. 2. (a) SMI-32 immunostaining and (b) SMI-311 immunostaining. Both stainings are executed after an MW pretreatment in TBS pH 9.0, at full power after 4 years of fixation. Without MW pretreatment, there is no immunostaining at all. Bar= 300 mm.
to 1 h no improvement in the intensity was noted. Somata were well stained but small dendrites were often not visible when a lower temperature was used. On the other hand the Microtubule Associated Protein-2, the phosphorylated and non-phosphorylated parts of the neurofilament (SMIs) and calbindin stained well at 90°C when the irradiation time was prolonged to 30 or 60 min.
In general we conclude that an MW pretreatment of 15 min at full power (boiling) gives optimal staining results for all tested antibodies. This optimal procedure has been applied sofar on brain tissue of 20 subjects. They all showed an improved quality of staining as compared with the previous procedures for microwave stimulated antigen retrieval, especially the calcium binding proteins (Table 2). These improve-
202
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
Fig. 3. (a) Calbindin immunostaining without MW pretreatment after 4 years of fixation. Only the somata of the strongly positive cells are stained. (b) Calbindin after MW pretreatment of a consecutive tissue slice in TBS pH 9.0 at full power. Besides the somata also many cells with nicely stained arborizations are visible. Bar= 300 mm.
ments were similar to those mentioned above. Postmortem intervals does not seem to influence the results, for the antibodies studied, as long as the interval was kept below the 18 h. This optimal MW antigen retrieval procedure appears to be also optimal for an antigen, resistant to periods of a few years of fixation, such as Neuropeptide Y. Immunostaining after MW pretreatment showed even more intensely
stained somata and dendrites than without pretreatment (Fig. 7d).
4. Discussion The effects of fixation in formaldehyde solutions are due to the formation of methylene bridges (crosslinking) between proteins. These changes in secondary
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
203
Fig. 4. (a) Parvalbumin immunostaining without MW pretreatment after 4 years of fixation. The somata of many parvalbumin positive cells are stained, but hardly any dendrites are visible. (b) Parvalbumin staining of a consecutive tissue slice after an MW pretreatment in TBS pH 9.0 at full power. Many fibers are visible in all cortical layers and more somata in layer III. Bar= 300 mm.
and tertiary structure of the proteins are the reason why many antibodies do not recognize an antigen or why the antigen is not accessible for the antibody (Fox et al., 1985; Mason and O’Leary, 1991; Dapson, 1993). Although the exact mechanism of antigen retrieval is not yet understood, it is believed that the cross links between proteins are broken, which makes the antigen accessible or recognizible for the antibody.
Although this process takes place when the tissue or sections are washed in distilled water for a very long time (several weeks, Helander, 1994) a combination of immersion in a solution with a high pH and high temperatures will accelerate the process. We have shown in a former study that for non-phosphorylated neurofilament an MW-pretreatment in a solution with a low pH (pH 2.5) gives very satisfactory results but
204
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
not at pH 4.5 (Evers and Uylings, 1994b). Many other researchers described that a more neutral pH (pH 6) gave an optimal staining result in their studies of other antigens. In this study we showed that an MW pretreatment of a tissue block in a simple buffer with a high pH of 9.0 – 9.5 (but not pH 8.5 or 8.0) gives good results in immunostainings on vibratome
sections with seven antibodies frequently used in neuroscience research. This extends the results described recently by Shi et al. (1995), who applied the technique for the detection of other antigens on mounted paraffin sections and came to the conclusion that a buffer with a high pH (pH 8–9) gave good results in all their tests. Another important factor in antigen retrieval techniques is high temperature. According to the literature there are different possibilities to achieve this high temperature, i.e. besides microwave-heating, a combination of microwave and pressure cooker, conventional heating or autoclave (Shi et al., 1991; Shin et al., 1991; Igarashi et al., 1994). Many good results were achieved with all these techniques. In our situation the microwave oven is most convenient to achieve high temperatures. It appears possible to obtain good results by microwaving at full power (boiling) and at a lower temperature of 90°C. However, when using this lower temperature the time of pretreatment should be lengthened to obtain in general the same results as with higher temperatures (boiling). However, for a few calcium binding protein antibodies, parvalbumin and calretinin, no optimal staining was obtained at 90°C. Therefore, we prefer the high temperature of boiling at full power. We have applied the optimal technique of MW pretreatment in pH 9.0 solution at full power for 15 min on about 20 patients with excellent results. Also paraffin sections instead of whole tissue blocks were successfully stained for SMI-32 and SMI-311 with this technique. However, when fixation times were extremely long (8 years and longer) it was impossible to achieve optimally stained cells. On the other hand paraffin sections of human brain tissue fixed for 1 year in Bodian fixative (a formaldehyde-containing solution) stained rather well with SMI-32 antibody after the optimal pretreatment procedure, whereas no immunostaining was obtained without pretreatment. In conclusion, we obtained one antigen retrieval procedure which is optimal for all the antibodies used in this study. This enlarges greatly the staining of important and irreplaceable human tissues and sections that were believed to have been fixed for too long.
Acknowledgements
Fig. 5. Calretinin immunostaining after 4 years of fixation with MW pretreatment in TBS pH 9.0 at full power. Without MW pretreatment there is absolutely no immunostaining. Bar= 300 mm.
The authors would like to thank Mr. G. van der Meulen for his photographical assistance and W.T.P. Verweij for her secretarial assistance and for correcting the English.
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
205
Fig. 6. Detail of (a) MAP-2 in layer V, (b) SMI-32 in layer V, (c) SMI-311 in layer III, and (d) SMI-312 in layers I – III. All immunostainings performed after 4 years of fixation with MW pretreatment in TBS pH 9.0 at full power. Bar= 50 mm.
206
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207
Fig. 7. Detail of (a) calbindin in layer II, (b) parvalbumin in layer V, (c) calretinin in layer II, and (d) NPY in layer V. All immunostainings performed after 4 years of fixation with MW pretreatment in TBS pH 9.0 at full power. Bar= 50 mm.
References Ainley, C.D. and Ironside, J.W. (1994) Microwave technology in diagnostic neuropathology, J. Neurosci. Methods, 55: 183 – 190.
Battiflora, H. and Kopinski, M. (1986) The influence of protease digestion and duration of fixation on the immunostaining of keratins, J. Histochem. Cytochem., 34: 1095 – 1100. Buijs, R.M., Pool, C.W., Van Heerikhuize, J.J., Sluiter, A.A., Van der Sluis, P.J., Ramkema, M., Van der Woude, T.P. and Van
P. E6ers, H.B.M. Uylings / Journal of Neuroscience Methods 72 (1997) 197–207 der Beek, E. (1989) Antibodies to small transmitter molecules and peptides: production and application of antibodies to Dopamine, Serotonine, GABA, Vasopressin, Vasoactive Intestinal Peptide, Neuropeptide Y, Somatostatin and Substance P. Biomed. Res., 10: 213 – 221. Cattoretti, G., Pileri, S., Parravicini, C., Becker, M.H.G., Poggi, S., Bifulco, C., Key, G., D’Amato, L., Sabattini, E., Feudale, E., Reynolds, F., Gerdes, J. and Rilke, F. (1993) Antigen unmasking on formalin-fixed, paraffin-embedded tissue sections, J. Pathol., 171: 83 – 98. Dapson, R.W. (1993) Fixation for the 1990s: a review of needs and accomplishments, Biotechnol. Histochem., 68: 75–82. Evers, P. and Uylings, H.B.M. (1994a) Effects of microwave pretreatment on immunocytochemical staining of vibratome sections and tissue blocks of human cerebral cortex stored in formaldehyde fixative for long periods, J. Neurosci. Methods, 55: 163 – 172. Evers, P. and Uylings, H.B.M. (1994b) Microwave-stimulated antigen retrieval is pH and temperature dependent, J. Histochem. Cytochem., 42: 1555–1563. Fox, C.H., Johnson, F.B., Whiting, J. and Roller, P.P. (1985) Formaldehyde fixation, J. Histochem. Cytochem., 33: 845 – 853. Helander, K.G. (1994) Kinetic studies of formaldehyde binding in tissue, Biotechnol. Histochem., 69: 177–179. Igarashi, H., Sugimura, H., Maruyama, K., Kitayama, Y., Ohta, I., Suzuki M., Tanaka, M., Dobashi, Y. and Kino, I. (1994) Alteration of immunoreactivity by hydrated autoclaving, microwave treatment, and simple heating of paraffin-embedded tissue sections. APMIS, 102: 295–307. Kitamoto, T., Ogomori, K., Tateishi, J. and Prusiner, S.B. Formic acid pretreatment enhances immunostaining of cerebral and systemic amyloids, Lab. Invest., 57: 230–236. Lucassen, P.J., Ravid, R., Gonatas, N.K. and Swaab, D.F. (1993) Activation of the human supraoptic and paraventricular nucleus neurons with aging and in Alzheimer’s disease as judged from increasing size of the Golgi apparatus, Brain Res., 632: 105 – 113. Mason, J.T. and O’Leary, T.J. (1991) Effects of formaldehyde fixation on protein secondary structure: a calorimetric and infrared spectroscopic investigation, J. Histochem. Cytochem., 39: 225 – 229. Momose, H., Metha, P. and Battiflora, H. (1993) Antigen retrieval by
.
207
microwave irradiation in lead thiocyanate, Appl. Immunochem., 1: 77 – 82. Munakata, S. and Hendricks, J.B. Effect of fixation time and microwave oven heating time on retrieval of the Ki-67 antigen from paraffin-embedded tissue, J. Histochem. Cytochem., 41: 1241– 1246. Podkletnova, I. and Alho, H. (1993) Ultrasound-amplified immunohistochemistry, J. Histochem. Cytochem., 41: 51 – 56. Sherriff, F.E., Bridges, L.R. and Jackson, P. (1994) Microwave antigen retrieval of b-amyloid precursor protein immunoreactivity. Neuroreport, 5: 1085 – 1088. Shi, S.-R., Key, M.E. and Kalra, K.L. (1991) Antigen retrieval in formalin-fixed, paraffin-embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections, J. Histochem. Cytochem., 39: 741 – 748. Shi, S.-R., Chaiwun, B., Young, L., Imam, A., Cote, R.J. and Taylor, C.R. (1994) Antigen retrieval using pH 3.5 glycine-HCl buffer or urea solution for immunohistochemical localization of Ki-67, Biotechnol. Histochem., 69: 213 – 215. Shi, S.-R., Imam, A., Young, L., Cote, R.J. and Taylor, C.R. (1995) Antigen retrieval immunohistochemistry under influence of pH using monoclonal antibodies, J. Histochem. Cytochem., 43: 193– 201. Shin, R.-W., Iwaki, T., Kitamoto, T. and Tateishi, J. (1991) Hydrated autoclave pretreatment enhances TAU immunoreactivity in formalin-fixed normal and Alzheimer’s disease brain tissues, Lab. Invest., 64: 693 – 702. Stra¨ter, J., Gu¨nthert, A.R., Bru¨derlein, S. and Mo¨ller, P. (1995) Microwave irradiation of paraffin-embedded tissue sensitizes the TUNEL method for in situ detection of apoptotic cells, Histochemistry, 103: 157 – 160. Suurmeijer, A.J.H. and Boon, M.E. (1993) Optimizing keratin and vimentin retrieval in formalin-fixed, paraffin-embedded tissue with the use of heat and metal salts, Appl. Immunohistochem., 1: 143 – 148. Yachnis, A.T. and Trojanowski, J.Q. (1994) Studies of childhood brain tumors using immunohistochemistry and microwave technology: methodological considerations, J. Neurosci. Methods, 55: 191 – 200.