Histol Histopathol (1998) 13: 275-281 Histology and 001: 10.14670/HH-13.275 Histopathology http://www.hh.um.es From Cell Biology to Tissue Engineering

Invited Review

Antigen detection on resin sections and methods for improving the immunogold labeling by manipulating the resin

S.-H. Brorson Department of Pathology, Ulleval Hospital, Oslo, Norway

Summary. Considering the importance of immuno­ qualities, but sections of methacrylate blocks were localization of cellular substances combined with good unstable in the electron beam and damage due to ultrastructure and ease of use, this review is focused on polymerisation were often present. In the sixties, epoxy the use of resin and the possibilities of manipulating the resins became popular because of good preservation of resin before and after embedding in order to improve the the ultrastructure and because of excellent stability in the immunolabeling of resin sections for electron electron beam. On the other hand, it was difficult to . The qualities of acrylic res ins and perform immunoelectron microscopy on epoxy sections. conventional epoxy resin for immunoelectron micro­ During the last 15-20 years methacrylate resins with scopy are discussed. Acrylic sections are usually more crosslinking qualities have been available, and these suited for immunoelectron microscopy than conven­ resins are well suited for immunoelectron microscopy tional epoxy sections. Different etching procedures and the stability of the sections in the electron beam is (sodium ethoxide or sodium metaperiodate) may be somewhat better than for the original methacrylates applied to conventional epoxy sections to enhance the (Newman and Hobot, 1987). Examples of such acrylic yield of immunolabeling. Lately, a method which does resins are Lowicryls and LR-White. Even if they have not involve any kind of etching has been developed for better qualities than older methacrylates for ultra­ enhancing the immunogold labeling of epoxy sections structural purposes, they do not match the epoxy resins up to about 8 times. This method involves increased in this respect. Immunolabeling of cryosections has also concentration of accelerator in the epoxy resin mixture been popular during the last decades, and by using that when processing the tissue. The ultrastructural technique no resin surrounds the tissue during immuno­ preservation of the tissue is important in immuno­ labeling (Roos and Morgan, 1990). electron microscopical procedures. and not only the intensity of the immunolabeling: in this respect no resin lB. Qualities of different resins may compete with the widely used epoxy resins. Epoxy resins are in several ways superior to other Key words: Acrylic resin, LR-White, Epoxy resin, resins as embedding medium. The polymerization occurs immunogold, Electron microscopy without significant shrinkage. Epoxy resin introduces intermolecular crosslinks, and these may act as a "fixative" on and nucleic acids. Epoxy sections are stable when exposed to the electron beam, which I. Introduction allows long observation times in the (Hayat, 1989). IA. History The true epoxy monomers in the resin mixture have a chemical structure where the reactive epoxy groups The most important progresses concerning may be concisely described as cyclic ethers with two embedding for electron microscopy occurred in the carbon atoms and one oxygen atom in the ring. Tertiary fifties and sixties. In the very late forties pure resin amines, often DMP-30 (Tris[dimethyl-aminomethyl] embedding was introduced by the use of butyl phenol), are used as accelerators to start the poly­ methacrylate (Newman et aI., 1949). The methacrylate merization process (Mark et aI., 1986). When enough resins became popular because of their good cutting epoxy monomers are present, secondary hydroxyl groups produced will add to the epoxies, and the Offprint requests to: Dr. Sverre·Henning Brorson, Department of polymerization process starts. To harden the plastic, two Pathology, Ulleval Hospitat. Kirkeveien 166, 0407 Oslo, Norway kinds of cyclic anhydrides of carboxyl acids are used: 276 detection on resin sections

NMA ( Nadic Me thy l Anh y dride) : a nd DDS A intro du c ti o n were no t o ri g in a ll y con s truc te d fo r (Dodecenyl-Succinic Anhydride). These anhydrides will immunoelectron mi croscopy. For in stance, LR-White compete with the secondary hydroxyl groups me ntioned was fa bricated to be a non-toxic altern ati ve to epoxy above and add to the epoxy groups (Mark et aI. , 1986). resins fo r use in conventional, ultrastructural EM . Other The poly val ent character of the epoxy monomer, makes acrylic resin s, like Lowicryl K4M , were constructed to po ly me ri zed e poxy resin hig hly crosslinked . The be a hydro phil ic resin . The pri mary reasoning is th at reacti ve epoxy groups have a great te nde ncy to react when the water is re moved from the ti ssue, hydrophilic with hydroxyl and amino groups, which are che mi cal components in the monomer mi xture will interact with side groups present in biological macromolecules such the m ac ro m o lec ul es, a nd the re b y the n e twork of as pro te in s a nd nuc le ic ac ids . The refore the bi o­ hy droge n bo unds is m a in ta ined. In thi s w a y the mo lecul es w ill be parts of the polymer network whe n biological structures will be kept as natural as possible e mbedded in epoxy resin (Causton, 1984: Kell enberger (Carl emalm et at., 1982). et at.. 1987; Hayat, 1989). The acry li c resins showed generall y better qualities Acrylic res in s (Lowic ry ls, LR-White) a re po ly ­ for immunoEM than epoxy resins, and it was suggested me rized by a no the r mecha ni sm : free rad ical c ha in tha t the reason was t hat aq ueous immuno reagents polymeri zation. Free radicals react with double bounds penetrated easily into the ' hydrophilic' secti ons. It was of the acryli c monomer, and a new radi cal, which is one claimed by Newman and Hobot ( 1987) that monomer larger, is produced. Monomers will continue to and secondary peroxidase reagents can penetrate about be added in this way and the polymer grows larger until 200 11m into LR-White sections with an incubation time its growth is terminated (Munk, 1989). The free radicals of 30 minutes for the primary . Similar results have no affinity for proteins and nucle ic acids, a nd were c la imed by E llinger a nd Pavelka ( 1985). The the refore bio mo lecules are no t incorporated into the theory of pe ne tra tio n o f a ntibodies into LR-White polymer network (Ke llenberger et a!. , 1987). secti ons was rejected by Brorson et a!. ( 1994), and the This a rti c le rev iews the use o f res in a nd the claimed ' penetrati on' was un veiled as an artefact. Other possibilities of ma nipula ting the res in in o rde r to sc ie n tis ts examine d the a bility of a ntibodies a nd improve the immunolabeling of resin secti ons. To do this ant ibody gold conj ugates to pe ne trate into fixed and properl y, it is necessary to cons ider the mechani sm fo r thawed cryosecti ons (S tierhof et a I. , 1986: Sti erhof and how are detected when being located in resin Sc h war z, 1989) b y ree m bed d ing immuno la be led sectio ns w ith s pec ia l inte rest b e ing s how n to the cryosecti ons in epoxy resin . No general penetrati on of interacti on between resin and antigens. How knowledge antibodies or the ir conjugates was observed . Immuno­ of these mechanisms is used to improve the immuno­ markers were onl y seen in side the reembedded c ryo­ labeling of epoxy sections will be illustrated . secti ons in areas where th e secti ons were inte rrupted . Examinati on of the capability of antibodies to penetrate II. Immunoelectron microscopy on thin sections de plasticized epoxy sections did not show any general antibody pe netrati on into these secti ons (Brorson and There a re two fundame nta ll y diffe re nt ways of S kj 0rte n , 1995). ' Pe ne tration ' of anti bodies was onl y perfo rming postembedding immunoelectron microscopy observed in the periphery of compact structures such as of sectio n s o n g rids. The firs t o ne is to pe rfo rm hormone vesicles. immunoEM on a secti on where the embedding media is Another theory to expl ain why acrylic secti ons are present. The altern ative is to use a immunoEM technique genera ll y better suited fo r immunoelectron microscopy where the embedding media is removed from the secti on than epoxy secti on was proposed by Ke ll enberger et al. prior to immunolabeling. Intuitively, it seems reasonabl e ( 1987): "Epoxies are abl e to fo rm covalent bonds with that immunoreagents w ill have be tter opportunities to bi o log ical ma te ri al , partic ul a rl y w ith prote in s. Co­ pe ne tra te in to the secti o ns if the re is no e mbedding polymeri zati on of epoxies with embedded ti ssues occurs, media present. An example of the first type is immuno­ w hile polyme ri zed acry li c resins permeate e mbedded labeling of acrylic sections and epoxy sections, and an ti ssues w ithout binding to the m . Accord ingly, durin g example of the second type is immunolabeling of cryo­ cleavage the behaviour of embedments with epox ies and sections and de plasti c ized e po xy sectio ns ( Ma r a nd acry li cs is di ffe re nt. Witho ut co-po ly me ri zati o n Wi ght , 1988; Roos and Morgan, 1990). We may also (acrylics), the surface of cleavage tends to fo ll ow the pe rfo rm ty pes o f postembedding immunoelectro n areas of least resistance, e.g., the inte rfaces be tween mi croscopy whi ch may be classifi ed between the two res in a nd pro teins. In e poxy-e mbedded m a te ri a l, groups mentioned above whe re th e resin is che mi call y however, the resistance in these interfaces is not ve ry ma nipulated prio r to immuno labe ling w itho ut be ing much less than in the proteins, and when cutting epoxy­ removed . e mbedded ti ssue, the surface of cleavage has greater tendency to di vide the proteins. Therefore antigens wi ll /lA. Immunolabeling of acrylic sections and epoxy protrude more easil y from acrylic secti ons than from sections epoxy sections, and acrylic secti ons will achieve larger amplitudes on their sllIface". This means that penetrati on Some of the ' new' acryli c resins mentioned in the of immunoreagents into the res in section is not necessary 277 Antigen detection on resin sections

to expl ain the amount of immunogold labeling observed epoxy sec ti ons was not cru ciall y different from th e (see also Enes trom and Kniola, 1995). corresponding labeling of LR-White secti ons. By a It ha s also been proposed th at the hydrophobic mathemati cal approach, Brorson and Skj0rten (1996b) charac ter of the conventional epoxy sections has reduced concluded th at proteins embedded in acryli c res in gain their ability to be used for immunolabeling. But it was immunolabeling by hav in g less tendency to get th e demonstrated by Durrenberger et al. (1 99 1) that there is outermost calotte cut off during the cutting procedure now signi ficant difference in immunolabeling due to (Kell enberger et aI. , 1987), while they lose by hav in g whether hydrophobic or hydrophilic secti ons are used, some parts hidden by resin . Depl as ti cized s, ass uming that un spec ific labeling is prevented . This was ori ginall y embedded in epoxy res in , win by havin g no proved by show ing that the hydrophobic Low icryl resin part hidden by resin , while they lose by havin g a great HM 20 is equ a ll y we ll suited for immunoelec tron tendency to get the outermost calotte cut off during the mi croscopy as the hydrophilic Low icryl res in K4M. cutting procedure. Often, hi gh concentrati ons of sodium eth ox ide do not themselves destroy in such a liB. Etching and dep/asticizing of epoxy sections way th at th e intensity of the immunogold labeling suffe rs (Brorson, 1997). Etching age nts such as Nal04 and H20 2 have bee n used pri or to the immunoprocedure to enhance the yield III. Theoretical study of the intensity of immuno­ of immunolabeling (Herrera et aI. , 1993). If the ti ssue labeling of acrylic sections and epoxy sections has been fixed by osmium tetroxide before embedding, th ese ox id ati on agents will reoxidize the osmium As mentioned in chapte r II , Ke ll enberger et a l. tetrox id e and remove it fro m th e secti on. This may ( 1987) proposed that epoxy secti ons are less suited fo r contribute to facilitate immunolabelin g (Baskin et aI., immunoelectron mi c roscopy than ac ryli c secti ons 1979) . But the immunolabelin g is increased after such because of th e 'e pox ies' tendency to react chem icall y etchin g even if osmium tetroxide is not initiall y used as with th e anti ge ns resulting in low amplitudes on the fixa ti ve (Enestrom and Kni ola, 1995 ). But the immuno­ surface of the epoxy secti ons. In spired by this theory the labe ling is normally not enhanced so mu ch with these rati o of immunogold labeling of LR-White sections and etchin g agent s as with sodium e thoxide, even if epoxy sec ti on (Llrw/Lep) was deduced mathemati call y inte restin g res ults are ac hi eved by usin g bo iling by Brorson and Skj0rten (1996c), and is give n by the solutions of sodium metaperiodate (Stirling and Graffs, fo rmul a: 1995). The ox idati on of the secti on will introduce hydro­ ph ili c chemi cal groups on th e epoxy resin (Causton . 1984). It has bee n claimed that thi s increased hydro­ philicity fac ilitates immunodetection. From the results of Durrenberger et al. ( 1991 ) menti oned in chapter lIa, thi s explanati on seems dubious. A more reliable theory proposes that the ox idati on breaks epoxy res in bonds and thereby increases the accessibility of the epitopes on the sur face o f th e secti on. This th eory has also been suggested to ex pl ain the observa ti on that oxidized epox y secti ons are easier to stain than untreated secti ons (Pfe iffer. 1982) . where'd ' is the di ameter of the prote in carryin g the Heating the thin secti ons by mic rowaves or by other epitopes. The two parameters rep and rl)'w are measures meth ods has been used fo r antige n retrieval by releas ing (in nanometers) for how tightl y the antIgen molec ul e is some of the fixa tions bounds produced by formaldehyde bound to the polymer network of epoxy res in and LR ­ or glutaraldehyde (Wil son et aI. , 1996). It is possible that White, respecti ve ly (the lower th e r- value, the more such treatment also releases some of the bounds between ti ghtl y th e antige n is bound to the polymer). The ' r­ epoxy resin and the fixed ti ss ue. values' are mani fes ted in different height amplitudes on The epoxy resin may be totall y removed from the the surface of the LR -White and epoxy sections (Fig. I), sec ti ons by treating them with strong sodium ethox ide and are specifi c for each protein. Typi cal amplitudes of so lution, which will enhance the immunolabeling (Mar an epoxy secti on are J -3 nm , and corresponding va lues and Wi ght , 1988; Baigent and Muller, 1990; Brorson of an acryli c secti on are 3-6 nm (Kell enberger et a!. , and Skj0rt en, 1995, 1996b), but the ultrastructural 1987). A consequence of the fo rmula for Llrw/Lep is that preserva tion may suffer. It was confirmed by Brorson th e advantage of usin g ac ryli c resin to epoxy resin is and Skj 0rt en ( 1995) th at ge neral penetrati on into the largest when immunolabeling large proteins (Fi g. 2) . depl as ti cized sec ti ons is not the mechani sm for th e improved yie ld of immunogold labeling . Antibody IV. Improvement of the labeling of epoxy sections ' penetrati on' was onl y detected in th e periphery of without using any etching agents co mpac t stru ctures such as hormone vesicles. It was also demonstrated th at the labeling intensity for depl asti cized As menti oned above, the yi eld of immunolabel in g 278 Antigen detection on resin sections

on epoxy sections may be increased by using strong epoxy polymer network. The practical solution was to etching agents to remove the resin partly or totally. But increase the concentration of accelerator, DMP-30, in the the etching agents may damage the ultrastructural infiltration and embedding steps when processing the preservation (Mar and Wight, 1988; Baigent and MUller, tissue. This embedding resin is called high-accelerator 1990; Brorson and Skj¢rten, 1995, 1996b). epoxy resin , and the sections cut from the resulting Another way of improving the immunolabeling was blocks are named high-accelerator epoxy sections. When worked out by Brorson and Skj¢rten (1996a,c). The method has been produced from the mathematically based theory mentioned in chapter III , which shows that the immunolabeling of epoxy sections would increase if 6 we could make chemical changes in the epoxy resin mixture that made the antigens less tightly linked to the

2

surface of secl)oo

• = epltopes exposed 2 4 6 10 12 DJameler (nm) Fig. 1. Illustrates the surface of an LR-White section. The figure shows the maximum height difference rlrw (= 2h 1rw). for the surface when cutting this particular protein (a figure for an epoxy section will be Fig. 2. This graph shows Ll rv/Lep as a function of the diameter of the similar). protein when the rap -value is 1.9 nm and the rlrw-value is 4.5 nm.

Fig. 3. a. Illustrates a high­ accelerator epoxy section of fibrin (8% acce lerator in the first infiltration step, 4% accelerator in the second one, and 6% in the embedding step) immuno­ labeled with anti-fibrinogen and 5 nm gold particles. There is high density of immunogold labeling, much more intense than for conventional embedding (Fig.3b). b. Illustrates a conventional epoxy section of fibrin (normal amount of accelerator) immunolabeled with anti-fibrinogen and 5 nm gold particles. There is a low density of immunogold labeling. x 22,500. Bar: 200 nm. 279 Antigen detection on resin sections

using high-accelerator sections based on 8% accelerator microscopy (IF). Immunogold labeling of the high­ in the first infiltration step, 4% accelerator in the second accelerator epoxy resin showed improved sensitivity for one , and 6% in the embedding step, the immunogold detecting small immune complex deposits, especially for labeling increased 7-8 times from the labeling seen in IgA (Fig. 4), compared to the IF-method ; otherwise there conventional, unetched epoxy sections (Fig. 3a-b). This was good correlation between immunoelectron increase was only seen for large proteins like fibrinogen, microscopy and IF. The ultrastructural presentation of thyroglobulin and, later on (Brorson et aI., 1997), the glomerular tissue was good on high-accelerator immunog lobulins (lgA, IgO , and Ig M), and is an epoxy sections when tannic acid was used to enhance the observati on which agrees with theoretical considerations contrast. Both the stabi Iity in the electron beam and (Brorson and Skj0rten, 1996c). The immunolabeling on ultrastructural preservation was significantly better than the hi gh-accelerator epoxy sections was only 30-40% for LR-White sections. This shows that immunoelectron less intense than for corresponding LR-White sections. microscopy on high-accelerator epoxy sections is well The following theory was proposed to explain th e suited for routine use. increase in the intensity of immunogold labeling of the hi gh-accelerator epoxy sections: " Increase of the V. Fixation and processing of the tissue before concentration of DMP-30 tends to block the polymer embedding chain synthesis at a low molecular weight (Hayat, 1970; Mark et aI., 1986). The epoxy components react with the There are more factors affecting the yield of side g ro ups of proteins (co-polymerization) when immunolabeling for electron microscopy than the ones normal amount of accelerator is used (Kellenberger et mentioned in chapters I-IV. But since this review is a I. , 19 87). The blocked pol y mer chains which are focused on the use of resin and the possibilities of obtained when using hi gh-accelerator epoxy resin , are manipulating the resin in order to improve the immuno­ not able to react with proteins. The co-polymerization is labeling, fixation and processing of the tissue will only therefore significantly reduced. On the basis of the be discussed briefly here. Optional fixation for ultra­ theory of Kellenberger et al. (1987), it is reasonable to structural electron microscopy involves perfusion believe that it will be more favourable for the surface of fixation with 2% glutaraldehyde, and postfixation with c leavage to follow the interface between resin and I % osmium tetroxide. But such a strong fixation will proteins during the cutting process than if the prote in s often be too damaging for sensitive antigens, and a were embedded in a conventional, more co-polymerized compromise between ultrastructural preservation and epoxy network." immunoreactivity has to be achieved. This is often Another way to enhance the immunolabeling of obtained by using a mixture of 4% paraformaldehyde epoxy sections is to use propylene oxide as an additional and 0.1-0.5% glutaraldehyde, osmium tetroxide being dehydration/infiltration agent in addition to ethanol or to avoided. Probably, the fixation network also cooperates use 5-10% propylene oxide in the embedding mixture with the polymer network to bind the antigens more or (Brorson, 1996). The mechanism for this kind of less tightly in the block (Kellenberger et aI., 19 87), enhanci ng is probably that propylene oxide blocks the resulting in lower or higher amplitudes on the surface of polymerization in a similar way as mentioned for the the sections after cutting (Chapter III and Fig. I). high accelerator procedure. This way of improving the The dehydration process may introduce disturbance immunolabeling is less potent than the one based on at the cellular and molecular level. To avoid this, increased concentration of accelerator. The increase in cryosectioning may be used (Roos and Morgan, 1990). immunolabeling using additional amounts of propylene Another alternative is free ze substitution which oxide is largest when the concentration of accelerator is combines freezing technique with resin e mbedding smallest, but the increase cannot compete with the effect (Hippe-Sanwald, 1993). Freeze substitution dissolves the of usi ng high-accelerator epoxy resin. Another important cellular ice in the frozen specimen by an organic solvent consequence of using additions of propylene oxide is which us ually contains chemical fixatives. This th at th e e mbedding mixture becomes less viscous . procedure is carried out at a very low temperature. The Similar enhancing results have been achieved by mixing ti ssue is finally embedded in resin , often the acrylic resin water into the partly water miscible epoxy resin Quetol Lowicryl. In this way the molecular disarrangements are 651 (Abad, 1992), and the mechanism proposed for this held at a very low level facilitating the immunolabeling enhancing was also reduction of the polymer cross­ of sensitive antigens on resin sections . linking by water blocking the polymerization . The practical use of high-accelerator epoxy resin has VI. Concluding remarks been demonstrated in the diagnosis of renal biopsies (Brorson et aI., 1997). Renal biopsies were embedded in Immunoelectron microscopy of resin sections has hi g h-accelerator epoxy res in, and postembedding been widely used. For research purposes cryosectioning immunoelectron microscopy was performed with and freeze substitution have become popular. Twenty antibodies directed against IgA, IgO, IgM and C3c to years ago electron microscopy was expected to have an detect immune complex deposits. Parallel samples of exploding development in routine pa thological renal ti ssue were subjected to immunofluorescence diagnostics, but this scenario has failed . Immunocyto- 280 Antigen detection on resin sections

,

"\;'" . "'. 4

Fig, 4. Th e ultrastructure and immunolabeling with anti-lgA and 15 nm gold particles of an immune complex deposit in a human glomerulus (Arrowheads show deposits) . x 7.500. Bar: 500 nm . - .. ._------

281 Antigen detection on resin sections

chemistry on frozen sections or paraffin sections for (1991). Polar or apolar lowicryl resin for immunolabeling. In : light microscopy has taken over the role that EM was Colloidal gold: Principles. methods and applications. Vol. 3. Hayat expected to achieve. Immunolabeli ng of conventional MA (ed) . Academic Press. pp 73-85. epoxy sections often requires st rong etching or Ellinger A. and Pavelka M. (1985). Post-embedding locali zation of deplasticizing procedures to be used. Acrylic sections glycoproteins by means of lectins on thin secti ons of ti ssue may be us ed, but have th ei r shortcomin gs, like in ­ embedded in LR-White. Histochem. J. 17. 1321-1336. stabi lity in the electron beam, reduced ultrastructural En estrbm S. and Kniola B. (1995). Resin embedding for quantitative preservation , sensiti vit y to oxygen during th e poly­ immunoelectron microscopy. A comparative compurized image merization process and reduced CUllin g qu alities. The analysis. Biotechn. Histochem. 25, 135-146. use of hi gh-accelerator epoxy resin for immunoelectron Hayat M.A. (1970). Polymerization . In: prinCiples and techniques of microscopy is we ll suited for routine use, and may electron microscopy. Vol. 1. Van Nostrand Reinh old Company. New contribute to an in crease in the interest for electron York. p 140. microscopy in pathology laboratori es. Hayat MA (1989) . Rinsing, dehydration and embedding. In: Principles and techniques of electron microscopy. Th e Macmillan Press Ltd. References London. pp 92-95. Herrera G.A. , Turbat-Herrera E.A. and Lockard G. V. (1993) . Abad A.R. (1992) . Medium temperature epoxy re sin for immunocyto­ Ultrastructural immunolabeling in the evaluation, diagnosis and chemistry: Quetol 651 with water. Microsc. Res. Techn. 20, 274-280. characterization of neuroendocrine neoplasms. Ultrastruct. Pathol. Baigent C.L. and Muller G. (1990). A new approach to the immuno­ 17, 93- 11 3. staining of epoxy-res in-embedded material. J. Microsc. 158. 73- Hippe-Sanwald S. (1993). Impact of freeze substitution on biological 80. electron microscopy. Microsc. Res. Techn. 24, 400-422. Baskin D.G. , Erlandsen S.I. and Pa rsons A. (1979). Influence of Kellenberger E., DLirrenberger M .. Villiger W., Carlemalm E. and Wurtz hydrogen peroxide or alcoholic sodium hydroxide on the immuno­ M. (1987). The efficiency of immu nolabel on Lowicryl secti ons cytochemical detection of growth hormone and prolactin aft er compared to theoretical predictions. J. Histochem. Cytochem. 35. osmium fixation. J. Histochem. Cytochem. 27, 1290-1292. 959-969. Brorson S.H. (1996). Improved immunogold labeling of epoxy sections Mar H. and Wight T.N . (1988). Colloidal gold immunostaining on by the use of propylene oxide as additional agent in dehydration, de plasticized ultra-thin sections. J. Histochem. Cytochem. 36, 1387- in filtration and embedding. Micron 27, 345-353. 1395. Brorson S. H. (1997). How to examine the antigen-damaging effect of Mark H.F., Bikales N.M ., Overberger C.G. and Menges G. (1986) . sodium ethoxide on deplasticized epoxy sections. J. Histochem. Encyclopaedia of polymer sciences and engineerin g. Vol. 6. 2. ed. Cytochem. 45, 143-146. Kroschwitz J.1. (ed). Pu bl. John Wileys and Sons pp. 335-400 Brorson S.H. and Skjorten F. (1995). Mechanism for antigen detection Munk P. (1989). Introduction to macromolecular sciences. John Wiley on deplasticized epoxy sections. Micron 26, 301-310. and Sons publ. Brorson S.H. and Skjorten F. (1996a) . Improved technique for immuno­ Newman G.R. and Hobot J .A. (1987) . Modern acrylics for post­ electron microscopy. How to prepare epoxy resin to obtain embeddin g immunostain ing techniques. J. Histochem . Cytochem. approximate the same immunogold labeling for epoxy sections as 35,971-981 for acrylic sections without any etching . Micron 27, 211-217. Newman J.B., Borysko E. and Swerdlow M. (1949). New sectioning Brorson S. H and Skjorten F. (1996b). The th eoretical ra tio of immuno­ techniques fo r light and electron microscopy. Science 110, 66. gold labeling of de plasticized epoxy secti ons and acrylic sections. Pfeiffer SW. (1982). Use of hydrogen peroxide to accelerate staining of Micron 27, 203-209. ultrathin spurr sections. Stain Technol. 57. 137-142. Brorson S.H. and Skjerten F. (1996c) . The th eo retical re lationship of Roos N. and Morgan A.J . (1990). Cryopreparation of thin biologica l immunogold labeling on acrylic sections and epoxy sections. Micron specimens for electron microscopy: Methods and appli cations. 27, 193-201. Royal Microsc. Soc. Oxford Un iversity Press. pp 69-79. Brorson S.H .. Roos N. and Skjorten F. (1994). Antibody penetration into Stierhof Y.D. and Schwarz H. (1989). Labeli ng properties of sucrose­ LR-White sections. Micron 25, 453-460. infiltrated cryosections. Scanning Microsc. Suppl. 3, 35-46. Brorson S.H., Strom E.H. and Skjorten F. (1997) . Immunoelectron Stierhof Y.D .. Schwarz H. and Frank H. (1986). Transverse sectioning of microscopy on epoxy sections without deplasticizing to detect plastic-embedded immunolabeled cryosections: morphology and glomerular immunoglobulin and complement depOsits in renal permeability to protein A-colloidal gold complexes. J. Ultrastru ct. diseases. APMIS 105. 139-14 9. Mol. Struct. Res. 97, 187-196. Carlemalm E .. Garavito R.M. and Villiger W. (1982). Resin development Stirling JW. and Graffs P.S . (1995). Antigen unmasking for immuno­ for electron microscopy and an analysis of embedding at low electron microscopy: labeling is improved by treating with sodium temperature. J. Microsc. 126/2, 123-1 43. ethoxide or sodium meta period ate, th en heating on retrieval Causton B.E . (1984) . Th e choice of re sin s for electron immunocyto­ medium. J. Histochem. Cytochem. 43, 115-123. chemistry. In : Immunolabeling for electron microscopy. Polak J.M. Wilson D.F. , Jiang D.J., Pierce A.M., and Wi ebkin OW. (1996) . Antigen and Varndell I.M. (eds). Elsevier Science Publishers. Amsterdam, retrieval for electron microscopy using a microwave techn iq ue for New York. pp 29-36. epithelial and basal lamina antigens. Appl. Immunohistochem . 4, 66- Durrenberger M .. Villiger W., Arnold B., Humbel B.M. and Schwarz H. 71