- Email: [email protected]
An improved silver staining technique as an alternative nuclear or combined nuclear nerve-fiber impregnation for comparative light-, secondary and backscattered electron scanning microscopy
Journal of Neuroscience Methods, 35 (1990) 3-8
3
Elsevier NSM 01128
An improved silver staining technique as an alternative nuclear or combined nuclear nerve-fiber impregnation for comparative light-, secondary and backscattered electron scanning microscopy D. von Langsdorff, S. Syed Ali and F. Niirnberger Department of Anatomy and Cytobiology, University of Giessen, Aulweg 123, D-6300 Giessen (F.R.G.) (Received 16 December 1989) (Revised received 6 March 1990) (Accepted 9 May 1990)
Key words: Silver staining technique; Electron microscopy Slices of rat brain were stained by a new silver impregnation technique. This method takes into consideration the pH-dependent differences of silver stain affinity of nerve tissues and can be used alternatively as a stain for nuclei or as a method for combined demonstration of nuclei nerves fibers. The slices were studied at the light microscopical (LM) level and subsequently with a scanning electron microscope, using secondary (SSEM = classical SEM), and backscattered electron detectors (BSEM). This new silver staining technique offers the opportunity of comparative studies with regard to different information acquired with LM, SSEM and BSEM. The described method allows to distinguish between nervous and glial tissue without necessarily damaging the glial tissue surrounding the nerve fibers. Specifically, scanning electron microscopy with backscattered electron detector of in situ preparations provides a higher contrast of stained and unstained tissue and increased depth of focus as compared to secondary electron detectors.
Introduction The application of the scanning electron microscope (SEM) for studying nerve tissue normally provides only information on the surface architecture, even after dissection, enzymatic digestion or tissue fracture of the surrounding tissue (Zeevi and Lewis, 1970; Lewis, 1971; Castej6n and Caraballo, 1980; Castej6n, 1981, 1984; Ulshafer and Allen, 1984). Thus, information on the environment of the nerve fibers is often lost because of the resulting tissue damage due to dissection or other factors related to the specific techniques. Brain development as well as the structure of the retina have been thoroughly examined with the
Correspondence: Dr. S. Syed Ali, Department of Anatomy and Cytobiology, University of Giessen, Aulweg 123, D-6300 Giessen (F.R.G.).
use of the SEM (Mestres and Rascher, 1977; Holly, 1982; Tayler and Roberts, 1983; Hanson, 1970; Meller and Tetzlaff, 1976; Davidorf and Sharpnack, 1977; Ulshafer and Allen, 1984). Only a few authors have used the possibility to stain nerve tissue for observation with SEM by use of silver contrasting. Based on its high atomic weight (DeNee et al., 1974, 1977), silver-impregnated nerve tissue can be distinguished from the surrounding tissue (Lewis, 1971). Backscattered electrons were used to enhance stained and unstained tissue components without damaging the environment of the nerve fibers (Tayler et al., 1984). The aim of the present investigation was the development of a method that allows the simultaneous observation of neuronal networks at the light and electron microscopical levels. Especially the connectivity of neuronal populations which can only be assumed at the light microscopical level should be detectable at higher resolution. In particular (1)
0165-0270/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
the efficiency of various tissue treatments on the degree of abundance of silver-stained neurons was tested. (2) The silver staining obtained in rather thick tissue slices was tested for its applicability and usefulness for scanning electron mciroscopy. (3) By m e a n s of SEM ~ t h ~ $ E detector, a high contrast and a large depth 9 f focus could be obtained which allows study of f i b e r contacts among different n e ~ e s with the help of three-dimensional observations.
the influence of the age of the different solutions were studied. Good results were achieved with the following staining protocol:
Silver impregnation solution (Agl), pH 3.5, t00 ml.
75 ml double-distilled water, 25 mt 100% ethanol, 20 g silver nitrate, 5 g potassium nitrate.
Ammoniac silver nitrate solutiuon (NH~Ag), pH 12, 100 ml. 75 ml double-distilled water, 25 ml
Materials and Methods
100% ethanol, 20 g silver nitrate, add concentrated ammonia until the initial precipitate disappears. Reducer solution (R), pH 4, 100 ml. 35 ml double-distilled water, 40 ml 100% ethanol, 25 ml 37% formalin.
Tissue preparation
100 ml 50% ethanol.
The brains of rats (Wistar, both sexes) perfused with Bouin's fluid were excised, stored for 2 days in the same fixative, postfixed for 7 days in 10% formalin (phosphate-buffered solution, p H 7.2, 0.135 M) and then embedded in paraffin. Tissue was cut (15,-30 t~m) and mounted on glass slides (3 slices each), Prior to silver siaining, the slices were deparaffinized in a x y l o l / e t h a n o i sequence (100% xylol for 10 min; 2 × 100%, 95%, 80%, 60%, 50% ethanol, for 5 min each).
Staining schedule for combined demonstration of nucleus-nerve fibers. (1) AgI 30 min. (2) NH3Ag
Washing solutions (WS-1, WS-2, WS-3), 100 ml.
60 s, WS-1 5 s, R 25 s, WS-2 30 s. (3)-(4) NH3-Ag 30 s, WS-1 5 s, R 25 s, WS-2 30 s. (5) WS-3 10 min. For a more intense staining, step (4) can be repeated.
Schedule for nuclear staining Omit step (1), other .steps as above.
Microscopy Silver staining The slices were impregnated with the silver s61ution (AgI), then transferred with a oarrier to ammoniac silver nitrate solution (NI-I3.Ag, the developer), followed by reducer soluti0n (R), and 3 washing steps (specific data, see below). The temperature of the impregnation solution was 35 ° C and the developer 20 ° C. All the steps were carr/ed out in the dark to avoid the influence of light i:ausiag unspecific silver reduction. The different factors in the staining method used were varied to examine their influence on the staining quality. Some of these parameters were: composition, concentration, p H sequence, time, and temperature of the different solutions. Also
The slices were examined with respect to staining for nucleus, nerve fibers, perikaryon and the background. They were photographed with a Zeiss light microscope (Photomikroskop, Oberkochen, F.R.G.). From selected slides, the sections plus underlying glass slide were cut out by use of a dental drill and adhered to SEM stubs with carbon-silver adhesive. To avoid overcharging in the SEM, the edge of specimen was covered with the adhesive and small copper wires connected specimen and stub. The specimens were coated with gold in a sputtering device (Balzers, Liechtenstein) for 40 s, and studied with a Philips PSEM 500 (Eindhoven, The Netherlands) scanning electron microscope. The sectors photographed with
l,
F:i& 1. a-e: medulla oblongata, reticular formation. Combined nucleus-nerve fiber stain technique (a-c). Same section with el@rgements in (b) ( = d) and (c) ( -- e) in LM (a), SSEM (b,d) and BSEM (c,e). Note clear appearance of nuclei (n) and nerve fibers [~!n LM and BSEM compared to SSEM. Signal charging of tissue (arrows) disappears in BSEM. SSEM gives better imaging of StlrfaP.ebut the depth effect in imaging stained structures is low. Also very thin nerve fibers can be visualised in SEM (d,e). a-c: • ~ ~'. xT00; d,e: x1350.
Ib
L
the LM were reidentified in the SEM and photographed with the secondary electron detector as well as with the backscattered electron detector. In interesting cases, stereophotographs ( + 6 o C) were taken with both detectors to obtain three-dimensional information.
Results
LM, Silver-staining Nerve fiber staining (Fig. 1a-e).
The AgI with low pH, which itself causes no colouring, is necessary for nerve fiber staining and is more successful at a temperature of 35 o C. The fiber staining is more intensive when the developing time is repeated rather than increasing the time of one passage. Nuclear staining (Fig. 2a-c). Only nuclei and perikarya of some other larger nerve cells are stained without AgI. Accordingly, the staining of nuclei of nerve and glial cells and the neuronal perikarya seems to be based on N H 3 A g at high pH. These results were achieved by varying composition and schedule of the solutions (see Materials and methods for details).
lncrease in specificity and reduction of background staining. The increase in the staining of nerve fibers and the decrease in background staining was dependent on the low concentrations of the silver solutions, and the fact that several passages of developer are more successful for specific silver staining than one passage with increased time. Also the washing solutions, particularly the use of separate washing solutions after each step and the addition of nitrate (potassium nitrate) to the AgI solution, served the same purpose.
Fig. 2. a-c: medulla oblongata, reticular formation. Nuclear stain technique. × 350. Same section in LM (a), SSEM (b) and BSEM (c). Nuclei (n) and nerve cell perikarya (p) are stained and show good contrast in LM and BSEM. Note nucleoli (asterisks) staining in all figures, better stained-unstained contrast and less signal charging (arrows) in BSEM than in SSEM.
A time period of 10 rain in the WS-3 step causes fading of the background staining and omits the necessity of final fixation. Duration of storage of solutions. The NH3Ag solution should be stored in the dark in a refrigerator and for no longer than 1 month, and renewed after every 10-15 series of staining. The AgI is still useable after 30 stainings and for several months, when stored in the dark in a refrigerator. The most stable solution is R, which can be used for more than 50 stainings and stored for more than 1 year. The use of the washing solutions, especially the application of separate WS, decreases the deterioration of the NH3Ag and R solutions. The WS should be renewed after every 5-10 stainings and not be stored.
Findings in comparative microscopy The stained structures viewed in the LM are also illustrated after BSEM (Figs. la,c,e and 2a,c). In BSEM, higher contrast and better signal ratio were obtained from stained structures, which were also located deeper in the tissue (depth range), than in SSEM (Figs. l b - e and 2b,c). Also the signals from overcharging do not produce such a disturbing effect in BSEM as in SSEM; SSEM often does not allow to distinguish stained structures from overcharging effects (arrows in Figs. l b - e and 2b,c). The application of stereophotography in SEM aids in solving three-dimensional considerations with a high resolution in the depth of the thick tissue slice.
to produce a latent image with a slightly but highly specific impregnation to nerve fibers and then to develop these invisible primary silver grains with a highly effective developer. The Golgi method stains randomly the cells whereas with our method all (argyrophilic) cells are stained. We found in comparison to other authors (Lillie and Fulmer, 1976; Cox, 1977) that addition of potassium nitrate and the reduction in staining intensity by extended times and lower concentrations increase the specificity of the staining and decrease background staining. Moreover, the different pH affinities in nerve tissues are taken into account by the use of two silver solutions differing in pH. The following aspects led us to choose a nonammoniac silver solution for the impregnation and an ammoniac silver solution for the development: The most critical aspect of silver staining is the colouration of nerve fibers which takes place during the highly specific impregnation step. Nerve fibers seem to have low-pH affinity and were not stained when the impregnation step at the low-pH AgI was omitted. Ammoniac silver solution is too potent for this step and has a high pH. Thus, we used a low-pH, non-ammoniac silver nitrate solution for this step with longer staining time and a higher temperature for more effectivity, and an ammoniac silver nitrate solution for the developing step where the nuclei and some perikarya with high-pH affinity were also stained. Evidently, lipophillic structures are present since a more lipophillic solution produced by adding ethanol improves the nerve fiber staining.
Discussion
Comparative microscopy Silver staining One of the most interesting results was to show that the low-pH AgI, which itself gives no stained image, was necessary to stain nerve fibers. Without the AgI only nuclei and the perikarya of larger nerve cells were stained with the high-pH NH3Ag solution. Thus, this staining technique gives the possibility for combined nucleus-nerve fiber staining or separate nuclear staining. The strategy of this silver staining technique was to follow procedures similar to photography:
For nerve tissue research the different silver staining methods, often modifications of the Golgi technique, represent some of the most important methods in LM. The use of SEM is not as frequent as EM and in most cases it is a pure surface contemplation rarely distinguishing nerve structures or the surrounding glial structures. This paper demonstrates a useful method to combine the different information from LM, SSEM and BSEM based on a variable silver staining technique without damage to glial tissue. The
L M gives the best s t a i n e d - u n s t a i n e d contrast but the resolution is very low. T h e S E M can b e regarded as a link b e t w e e n L M a n d t r a n s m i s s i o n electron microscopy, the disadvantage being the thickness of sections a n d the small size of the tissue specimens. W i t h BSEM the high resolution a n d the increased d e p t h of focus of the SEM can be comb i n e d with a higher s t a i n e d - u n s t a i n e d contrast a n d a higher range in depth t h a n within SSEM. F u r t h e r m o r e , the u n d e s i r a b l e signal charging in S E M is reduced in BSEM. O n the other hand, SSEM is more a d v a n t a g e o u s for i n t e r p r e t a t i o n of surface architecture. T h e s t e r e o p h o t o g r a p h y in SEM gives more detailed i n f o r m a t i o n a b o u t plasticity, particularly of fine structures where L M resolution is n o t sufficient.
Acknowledgements The authors are grateful to Dr. R.L. Snipes for r e a d i n g a n d correcting the m a n u s c r i p t . M a n y thanks are also due to Mrs. B. W i l d n e r for skilfully p r e p a r i n g the m a n u s c r i p t . Part of this work was done in partial fulfilment for a n M D degree of D.v.L.
References Castej6n, O.J. (1981) Light microscope, SEM and TEM study of fish cerebellar granule cells. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, t981, IV, SEM Inc., AMF O'Hare, Chicago, IL, pp. 105-113. Castej6n, O.J. (1984) Low resolution scanning electron microscopy of cerebellar neurons and neuroglial cells of the granular layer. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1989, III, SEM Inc., AMF O'Hare, Chicago, IL, pp. 1391-1400. Castej6n, O.J. and Caraballo, A.J. (1980) Light and scanning electron microscopic study of cerebeUar cortex of teleost fishes, Cell Tissue Res., 207: 211-226. Cox, C. (1977) Neuropathological techniques. In: Bancroft, J.D. and Stevens, A. (Eds.), Theory and practice of histological techniques. Churchill Livingstone, Edinburgh, London, New York, pp. 249-273.
Davidorf. F.H. and Sharpnack, D.D. (1977) Surface ultrastructure of the human retina. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1977, II, SEM Inc., AMF O'Hare, Chicago, IL, pp. 459-466. DeNee, P.B., Abraham, J.L. and Willard, P.A. (1974) Histological stains for the scanning electron microscope - qualitative and semiquantitative aspects of specific silver stains. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1974, I, SEM Inc., AMF O'Hare, Chicago, IL, pp. 259-266. DeNee, P.B., Frederickson, R.G. and Pope, R.S. (1977) Heavy metal staining of paraffin, epoxy and glycol methacrylate embedded biological tissue for scanning electron microscope histology. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1977, II, SEM Inc., AMF O'Hare, Chicago, IL, pp. 83-90. Hansen, H.A. (1970) Scanning electron microscopy of the rat retina. Z. Zellforsch., 170: 23-44. Holly, J.A. (1982) Early development of the circumferential axonal pathway in mouse and chick spinal cord. J. Comp. Neurol., 205: 371-382. Lewis, E.R. (1971) Studying neuronal architecture and organization with SEM. In: Becket, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1971, I. SEM Inc., AMF O'Hare, Chicago, IL, pp. 281-288. Lillie, R.D. and Fullmer, H.M. (1976) Histopathologic Technique and Practical Histochemistry, 4th edition, McGrawHill Book Company, New York, pp. 765-786. Meller, K. and Tetzlaff, W. (1976) Scanning electron microscopic studies on the development of the chick retina. Cell Tissue Res., 170; 145-159. Mestres, P. and Rascher, K. (1977) Some aspects of the early development of the rat hypothalamus: a scanning electron microscopic (SEM) study. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1977, IL SEM Inc., AMF O'Hare, Chicago, IL, pp. 381-386. Tayler, J.S.H. and Roberts, A. (1983) The early development of the peripheral nervous system in an amphibian embryo; a scanning electron microscope study. J. Embryol. Exp. Morphol., 75: 31-47. Tayler, J.S.H., Fawcett, J.W. and Hirst, L. (1984) The use of backscattered electrons to examine selectivelystained nerve fibers in the scanning microscope. Stain Technol., 59: 335341. Ulshafer, R.J., Allen, C.B. (1984) Scanning electron microscopy of the retina in an animal model of hereditary blindness. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1984, II, SEM Inc., AMF O'Hare, Chicago, IL, pp. 841-848. Zeevi, Y.Y. and Lewis, E.R. (1970) A new technique for exposing neuronal surfaces for viewing with the scanning electron microscope. In: Microscopie Electronique 1970, Societ6 Francaise de Microscopie Electronique, Paris, vol. 1, pp. 481-482.
3
Elsevier NSM 01128
An improved silver staining technique as an alternative nuclear or combined nuclear nerve-fiber impregnation for comparative light-, secondary and backscattered electron scanning microscopy D. von Langsdorff, S. Syed Ali and F. Niirnberger Department of Anatomy and Cytobiology, University of Giessen, Aulweg 123, D-6300 Giessen (F.R.G.) (Received 16 December 1989) (Revised received 6 March 1990) (Accepted 9 May 1990)
Key words: Silver staining technique; Electron microscopy Slices of rat brain were stained by a new silver impregnation technique. This method takes into consideration the pH-dependent differences of silver stain affinity of nerve tissues and can be used alternatively as a stain for nuclei or as a method for combined demonstration of nuclei nerves fibers. The slices were studied at the light microscopical (LM) level and subsequently with a scanning electron microscope, using secondary (SSEM = classical SEM), and backscattered electron detectors (BSEM). This new silver staining technique offers the opportunity of comparative studies with regard to different information acquired with LM, SSEM and BSEM. The described method allows to distinguish between nervous and glial tissue without necessarily damaging the glial tissue surrounding the nerve fibers. Specifically, scanning electron microscopy with backscattered electron detector of in situ preparations provides a higher contrast of stained and unstained tissue and increased depth of focus as compared to secondary electron detectors.
Introduction The application of the scanning electron microscope (SEM) for studying nerve tissue normally provides only information on the surface architecture, even after dissection, enzymatic digestion or tissue fracture of the surrounding tissue (Zeevi and Lewis, 1970; Lewis, 1971; Castej6n and Caraballo, 1980; Castej6n, 1981, 1984; Ulshafer and Allen, 1984). Thus, information on the environment of the nerve fibers is often lost because of the resulting tissue damage due to dissection or other factors related to the specific techniques. Brain development as well as the structure of the retina have been thoroughly examined with the
Correspondence: Dr. S. Syed Ali, Department of Anatomy and Cytobiology, University of Giessen, Aulweg 123, D-6300 Giessen (F.R.G.).
use of the SEM (Mestres and Rascher, 1977; Holly, 1982; Tayler and Roberts, 1983; Hanson, 1970; Meller and Tetzlaff, 1976; Davidorf and Sharpnack, 1977; Ulshafer and Allen, 1984). Only a few authors have used the possibility to stain nerve tissue for observation with SEM by use of silver contrasting. Based on its high atomic weight (DeNee et al., 1974, 1977), silver-impregnated nerve tissue can be distinguished from the surrounding tissue (Lewis, 1971). Backscattered electrons were used to enhance stained and unstained tissue components without damaging the environment of the nerve fibers (Tayler et al., 1984). The aim of the present investigation was the development of a method that allows the simultaneous observation of neuronal networks at the light and electron microscopical levels. Especially the connectivity of neuronal populations which can only be assumed at the light microscopical level should be detectable at higher resolution. In particular (1)
0165-0270/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
the efficiency of various tissue treatments on the degree of abundance of silver-stained neurons was tested. (2) The silver staining obtained in rather thick tissue slices was tested for its applicability and usefulness for scanning electron mciroscopy. (3) By m e a n s of SEM ~ t h ~ $ E detector, a high contrast and a large depth 9 f focus could be obtained which allows study of f i b e r contacts among different n e ~ e s with the help of three-dimensional observations.
the influence of the age of the different solutions were studied. Good results were achieved with the following staining protocol:
Silver impregnation solution (Agl), pH 3.5, t00 ml.
75 ml double-distilled water, 25 mt 100% ethanol, 20 g silver nitrate, 5 g potassium nitrate.
Ammoniac silver nitrate solutiuon (NH~Ag), pH 12, 100 ml. 75 ml double-distilled water, 25 ml
Materials and Methods
100% ethanol, 20 g silver nitrate, add concentrated ammonia until the initial precipitate disappears. Reducer solution (R), pH 4, 100 ml. 35 ml double-distilled water, 40 ml 100% ethanol, 25 ml 37% formalin.
Tissue preparation
100 ml 50% ethanol.
The brains of rats (Wistar, both sexes) perfused with Bouin's fluid were excised, stored for 2 days in the same fixative, postfixed for 7 days in 10% formalin (phosphate-buffered solution, p H 7.2, 0.135 M) and then embedded in paraffin. Tissue was cut (15,-30 t~m) and mounted on glass slides (3 slices each), Prior to silver siaining, the slices were deparaffinized in a x y l o l / e t h a n o i sequence (100% xylol for 10 min; 2 × 100%, 95%, 80%, 60%, 50% ethanol, for 5 min each).
Staining schedule for combined demonstration of nucleus-nerve fibers. (1) AgI 30 min. (2) NH3Ag
Washing solutions (WS-1, WS-2, WS-3), 100 ml.
60 s, WS-1 5 s, R 25 s, WS-2 30 s. (3)-(4) NH3-Ag 30 s, WS-1 5 s, R 25 s, WS-2 30 s. (5) WS-3 10 min. For a more intense staining, step (4) can be repeated.
Schedule for nuclear staining Omit step (1), other .steps as above.
Microscopy Silver staining The slices were impregnated with the silver s61ution (AgI), then transferred with a oarrier to ammoniac silver nitrate solution (NI-I3.Ag, the developer), followed by reducer soluti0n (R), and 3 washing steps (specific data, see below). The temperature of the impregnation solution was 35 ° C and the developer 20 ° C. All the steps were carr/ed out in the dark to avoid the influence of light i:ausiag unspecific silver reduction. The different factors in the staining method used were varied to examine their influence on the staining quality. Some of these parameters were: composition, concentration, p H sequence, time, and temperature of the different solutions. Also
The slices were examined with respect to staining for nucleus, nerve fibers, perikaryon and the background. They were photographed with a Zeiss light microscope (Photomikroskop, Oberkochen, F.R.G.). From selected slides, the sections plus underlying glass slide were cut out by use of a dental drill and adhered to SEM stubs with carbon-silver adhesive. To avoid overcharging in the SEM, the edge of specimen was covered with the adhesive and small copper wires connected specimen and stub. The specimens were coated with gold in a sputtering device (Balzers, Liechtenstein) for 40 s, and studied with a Philips PSEM 500 (Eindhoven, The Netherlands) scanning electron microscope. The sectors photographed with
l,
F:i& 1. a-e: medulla oblongata, reticular formation. Combined nucleus-nerve fiber stain technique (a-c). Same section with el@rgements in (b) ( = d) and (c) ( -- e) in LM (a), SSEM (b,d) and BSEM (c,e). Note clear appearance of nuclei (n) and nerve fibers [~!n LM and BSEM compared to SSEM. Signal charging of tissue (arrows) disappears in BSEM. SSEM gives better imaging of StlrfaP.ebut the depth effect in imaging stained structures is low. Also very thin nerve fibers can be visualised in SEM (d,e). a-c: • ~ ~'. xT00; d,e: x1350.
Ib
L
the LM were reidentified in the SEM and photographed with the secondary electron detector as well as with the backscattered electron detector. In interesting cases, stereophotographs ( + 6 o C) were taken with both detectors to obtain three-dimensional information.
Results
LM, Silver-staining Nerve fiber staining (Fig. 1a-e).
The AgI with low pH, which itself causes no colouring, is necessary for nerve fiber staining and is more successful at a temperature of 35 o C. The fiber staining is more intensive when the developing time is repeated rather than increasing the time of one passage. Nuclear staining (Fig. 2a-c). Only nuclei and perikarya of some other larger nerve cells are stained without AgI. Accordingly, the staining of nuclei of nerve and glial cells and the neuronal perikarya seems to be based on N H 3 A g at high pH. These results were achieved by varying composition and schedule of the solutions (see Materials and methods for details).
lncrease in specificity and reduction of background staining. The increase in the staining of nerve fibers and the decrease in background staining was dependent on the low concentrations of the silver solutions, and the fact that several passages of developer are more successful for specific silver staining than one passage with increased time. Also the washing solutions, particularly the use of separate washing solutions after each step and the addition of nitrate (potassium nitrate) to the AgI solution, served the same purpose.
Fig. 2. a-c: medulla oblongata, reticular formation. Nuclear stain technique. × 350. Same section in LM (a), SSEM (b) and BSEM (c). Nuclei (n) and nerve cell perikarya (p) are stained and show good contrast in LM and BSEM. Note nucleoli (asterisks) staining in all figures, better stained-unstained contrast and less signal charging (arrows) in BSEM than in SSEM.
A time period of 10 rain in the WS-3 step causes fading of the background staining and omits the necessity of final fixation. Duration of storage of solutions. The NH3Ag solution should be stored in the dark in a refrigerator and for no longer than 1 month, and renewed after every 10-15 series of staining. The AgI is still useable after 30 stainings and for several months, when stored in the dark in a refrigerator. The most stable solution is R, which can be used for more than 50 stainings and stored for more than 1 year. The use of the washing solutions, especially the application of separate WS, decreases the deterioration of the NH3Ag and R solutions. The WS should be renewed after every 5-10 stainings and not be stored.
Findings in comparative microscopy The stained structures viewed in the LM are also illustrated after BSEM (Figs. la,c,e and 2a,c). In BSEM, higher contrast and better signal ratio were obtained from stained structures, which were also located deeper in the tissue (depth range), than in SSEM (Figs. l b - e and 2b,c). Also the signals from overcharging do not produce such a disturbing effect in BSEM as in SSEM; SSEM often does not allow to distinguish stained structures from overcharging effects (arrows in Figs. l b - e and 2b,c). The application of stereophotography in SEM aids in solving three-dimensional considerations with a high resolution in the depth of the thick tissue slice.
to produce a latent image with a slightly but highly specific impregnation to nerve fibers and then to develop these invisible primary silver grains with a highly effective developer. The Golgi method stains randomly the cells whereas with our method all (argyrophilic) cells are stained. We found in comparison to other authors (Lillie and Fulmer, 1976; Cox, 1977) that addition of potassium nitrate and the reduction in staining intensity by extended times and lower concentrations increase the specificity of the staining and decrease background staining. Moreover, the different pH affinities in nerve tissues are taken into account by the use of two silver solutions differing in pH. The following aspects led us to choose a nonammoniac silver solution for the impregnation and an ammoniac silver solution for the development: The most critical aspect of silver staining is the colouration of nerve fibers which takes place during the highly specific impregnation step. Nerve fibers seem to have low-pH affinity and were not stained when the impregnation step at the low-pH AgI was omitted. Ammoniac silver solution is too potent for this step and has a high pH. Thus, we used a low-pH, non-ammoniac silver nitrate solution for this step with longer staining time and a higher temperature for more effectivity, and an ammoniac silver nitrate solution for the developing step where the nuclei and some perikarya with high-pH affinity were also stained. Evidently, lipophillic structures are present since a more lipophillic solution produced by adding ethanol improves the nerve fiber staining.
Discussion
Comparative microscopy Silver staining One of the most interesting results was to show that the low-pH AgI, which itself gives no stained image, was necessary to stain nerve fibers. Without the AgI only nuclei and the perikarya of larger nerve cells were stained with the high-pH NH3Ag solution. Thus, this staining technique gives the possibility for combined nucleus-nerve fiber staining or separate nuclear staining. The strategy of this silver staining technique was to follow procedures similar to photography:
For nerve tissue research the different silver staining methods, often modifications of the Golgi technique, represent some of the most important methods in LM. The use of SEM is not as frequent as EM and in most cases it is a pure surface contemplation rarely distinguishing nerve structures or the surrounding glial structures. This paper demonstrates a useful method to combine the different information from LM, SSEM and BSEM based on a variable silver staining technique without damage to glial tissue. The
L M gives the best s t a i n e d - u n s t a i n e d contrast but the resolution is very low. T h e S E M can b e regarded as a link b e t w e e n L M a n d t r a n s m i s s i o n electron microscopy, the disadvantage being the thickness of sections a n d the small size of the tissue specimens. W i t h BSEM the high resolution a n d the increased d e p t h of focus of the SEM can be comb i n e d with a higher s t a i n e d - u n s t a i n e d contrast a n d a higher range in depth t h a n within SSEM. F u r t h e r m o r e , the u n d e s i r a b l e signal charging in S E M is reduced in BSEM. O n the other hand, SSEM is more a d v a n t a g e o u s for i n t e r p r e t a t i o n of surface architecture. T h e s t e r e o p h o t o g r a p h y in SEM gives more detailed i n f o r m a t i o n a b o u t plasticity, particularly of fine structures where L M resolution is n o t sufficient.
Acknowledgements The authors are grateful to Dr. R.L. Snipes for r e a d i n g a n d correcting the m a n u s c r i p t . M a n y thanks are also due to Mrs. B. W i l d n e r for skilfully p r e p a r i n g the m a n u s c r i p t . Part of this work was done in partial fulfilment for a n M D degree of D.v.L.
References Castej6n, O.J. (1981) Light microscope, SEM and TEM study of fish cerebellar granule cells. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, t981, IV, SEM Inc., AMF O'Hare, Chicago, IL, pp. 105-113. Castej6n, O.J. (1984) Low resolution scanning electron microscopy of cerebellar neurons and neuroglial cells of the granular layer. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1989, III, SEM Inc., AMF O'Hare, Chicago, IL, pp. 1391-1400. Castej6n, O.J. and Caraballo, A.J. (1980) Light and scanning electron microscopic study of cerebeUar cortex of teleost fishes, Cell Tissue Res., 207: 211-226. Cox, C. (1977) Neuropathological techniques. In: Bancroft, J.D. and Stevens, A. (Eds.), Theory and practice of histological techniques. Churchill Livingstone, Edinburgh, London, New York, pp. 249-273.
Davidorf. F.H. and Sharpnack, D.D. (1977) Surface ultrastructure of the human retina. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1977, II, SEM Inc., AMF O'Hare, Chicago, IL, pp. 459-466. DeNee, P.B., Abraham, J.L. and Willard, P.A. (1974) Histological stains for the scanning electron microscope - qualitative and semiquantitative aspects of specific silver stains. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1974, I, SEM Inc., AMF O'Hare, Chicago, IL, pp. 259-266. DeNee, P.B., Frederickson, R.G. and Pope, R.S. (1977) Heavy metal staining of paraffin, epoxy and glycol methacrylate embedded biological tissue for scanning electron microscope histology. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1977, II, SEM Inc., AMF O'Hare, Chicago, IL, pp. 83-90. Hansen, H.A. (1970) Scanning electron microscopy of the rat retina. Z. Zellforsch., 170: 23-44. Holly, J.A. (1982) Early development of the circumferential axonal pathway in mouse and chick spinal cord. J. Comp. Neurol., 205: 371-382. Lewis, E.R. (1971) Studying neuronal architecture and organization with SEM. In: Becket, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1971, I. SEM Inc., AMF O'Hare, Chicago, IL, pp. 281-288. Lillie, R.D. and Fullmer, H.M. (1976) Histopathologic Technique and Practical Histochemistry, 4th edition, McGrawHill Book Company, New York, pp. 765-786. Meller, K. and Tetzlaff, W. (1976) Scanning electron microscopic studies on the development of the chick retina. Cell Tissue Res., 170; 145-159. Mestres, P. and Rascher, K. (1977) Some aspects of the early development of the rat hypothalamus: a scanning electron microscopic (SEM) study. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1977, IL SEM Inc., AMF O'Hare, Chicago, IL, pp. 381-386. Tayler, J.S.H. and Roberts, A. (1983) The early development of the peripheral nervous system in an amphibian embryo; a scanning electron microscope study. J. Embryol. Exp. Morphol., 75: 31-47. Tayler, J.S.H., Fawcett, J.W. and Hirst, L. (1984) The use of backscattered electrons to examine selectivelystained nerve fibers in the scanning microscope. Stain Technol., 59: 335341. Ulshafer, R.J., Allen, C.B. (1984) Scanning electron microscopy of the retina in an animal model of hereditary blindness. In: Becker, R.P. and Johari, O. (Eds.), Scanning Electron Microscopy, 1984, II, SEM Inc., AMF O'Hare, Chicago, IL, pp. 841-848. Zeevi, Y.Y. and Lewis, E.R. (1970) A new technique for exposing neuronal surfaces for viewing with the scanning electron microscope. In: Microscopie Electronique 1970, Societ6 Francaise de Microscopie Electronique, Paris, vol. 1, pp. 481-482.