A technique which has been widely used in quantitative studies of peripheral nerve involves fixation in Flemming’s fluid, embedding in paraffin wax, and staining with haematoxylin. An appraisal of the size changes induced by a standardized form of this technique has been made. Observations have been confined to changes in the thickness and external diameter of the internodal compact myelin sheath of the non-fasciculated nervus gastrocnemius medialis of the adult rabbit.
For a comparison of external diameters, sections from adjacent parts of the chosen nerve were prepared (a) by the standardized Flemming-Wolter technique; (b) by the rapid freezing technique, previously shown to reproduce accurately the dimensions present in fresh teased unfixed nerve-fibres. There was no significant difference in distribution of external diameters between sections prepared in the two ways.
For a comparison of sheath thicknesses, a regression line of 2 × myelin sheath thickness on external diameter was derived from each type of preparation. As external diameter was not appreciably affected by processing, a direct comparison of these regression lines could be made. It was found that regardless of external diameter or original sheath thickness, processing caused a reduction of sheath thickness of some 20% as a result of increase in internal diameter. This latter increase ranged from about 50% to under 20% with increasing fibre size.
Single frozen sections were passed through the series of reagents used in the Flemming-Wolter technique and photographed after each of seven stages. The principal changes in sheath thickness occurred in the first and last stages of processing. External diameter was not appreciably affected whereas the internal margin of the sheath was extremely labile. Observations suggest that slight inconsistencies in technique may be accompanied by relatively gross errors when measurements are undertaken. All stages need careful standardization, the most important in this respect being fixation, differentiation, final dehydration, and photographic exposure.
Most quantitative studies of peripheral nerve have been based on sections which have been fixed and stained. Such methods of preparation
are generally assumed to be accompanied by some degree of shrinkage and distortion (Baker, 1958), which will vary with the particular tissue component and reagents used.
It has been shown that the proportion of compact myelin to axoplasm in peripheral nerve-fibres varies appreciably not only with external fibre diameter (Taylor, 1942; Sanders, 1948; Evans and Vizoso, 1951), but also with the type of nerve studied (Wendell-Smith and Williams, 1958). It is possible that the size changes of axoplasm and myelin induced by a variety of reagents are comparable and proportionate so that the relation between them is unaffected. However, in view of the marked difference in the ultrastructure of these two components, this possibility seems remote, and on a priori grounds differential effects may be postulated.
A technique which has been widely used involves fixation in Flemming’s fluid, embedding in paraffin wax, and staining with haematoxylin. Much important comparative work has been carried out by a standardized form of this technique. The error involved has proved unimportant in this context. Moreover, most workers have limited their studies to external fibre diameter, and as will be shown this dimension is not appreciably affected by the technique. However, modern theories of conduction (Rushton, 1951) concern themselves in part with the ratio between the diameter of the axon and the thickness of the ensheathing structures, so that for their verification a knowledge of the effect of the standardized technique on these dimensions is necessary. Thus attention is focused not only on possible changes in the external diameter of the compact myelin sheath but more particularly on changes in internal diameter. Further, the validity of assuming that estimates of internal diameter performed on fixed and stained preparations provide a reliable index of axon diameter (e.g. Gasser and Grundfest, 1939) must be considered. In addition the importance of internal diameter measurements is stressed in relation to the observed differences between nerve-fibres of muscle and skin of the same external diameter, and to possible variations in the ratio of thickness of the myelin sheath to diameter of the axon along the length of the fibre (Evans and Vizoso, 1951; Sunderland and Roche, 1958).
The literature dealing with the size changes during processing has been reviewed by Sunderland and Roche (1958). A number of different techniques have been assessed by as many methods, but the majority of workers have confined their studies to changes in external diameter. Sanders (1948) studied the changes in both internal and external diameter as a result of a modified Weigert technique (Gutmann and Sanders, 1943). The principal stages of this technique are similar to those considered here. However, the nerve studied was the peroneal nerve of the rabbit, in which there are nerve-fibres of both muscle and skin. His basis of reference was the freshly teased nerve-fibre, and he made no allowance for the magnifying effect of the myelin lens system (Williams and Wendell-Smith, 1958).
Material and Method
The nervus gastrocnemius medialis (N.G.M.) of the rabbit was used throughout the present investigation. Adult rabbits were used; no account was taken of breed, sex, or weight. The use of lipid-soluble inhalation anaesthetics was avoided and intravenous nembutal was used. The appropriate part of the sciatic nerve and its branches was removed in one piece. Sections from a short segment including the N.G.M. at about 3 cm proximal to its muscle were prepared in two ways.
(i) 5-or 6-μ sections of the fresh nerve were prepared by a rapid freezing technique (Williams, 1959). The sections were irrigated with a standard volume of normal saline and photographed immediately without a cover-slip, by polarized light. The volume of saline, 2 drops, has been found by experiment to be the most satisfactory. A larger volume often introduces an undesirable degree of meniscus curvature and is more sensitive to extraneous air currents, which produce streaming and interference with the polarization optics.
In such sections the dimensions of the compact myelin sheath do not differ significantly from those present in fresh teased fibres (Wendell-Smith and Williams, 1959).
(ii) The second method of preparation is given in full because it is believed that slight variations in technique may prove important when measurements are contemplated (Causey, 1948; Evans and Vizoso, 1951).
1. Remove the nerve carefully with the epineurium intact, and attach under slight tension to a frame prepared from a library card.
2. Fix for 24 h in 19 ml of freshly prepared Flemming’s fluid, with reduced acetic acid, in a 20 ml tightly stoppered bottle.
Fixative solution used
1% chromic acid in distilled water—15 ml.
2% osmium tetroxide in distilled water—4 ml. glacial acetic acid—1 drop (25 mgm).
3. Dehydrate without washing for 1 h each in 70%, 90%, and 95 % alcohol, and absolute alcohol (2 changes).
4. Clear in cedarwood oil, 15 b.
5. Vacuum-embed in paraffin wax (m.p. 56° C), 2 h.
6. Section at 5 or 6 μ and attach to slide.
7-3% potassium dichromate at 37° C, 6 h.
8. Rinse, stain in 100 ml Wolter’s haematoxylin for 15 h at 37° C.
haematoxylin 1 g—dissolve in 10 to 20 ml of absolute alcohol, distilled water to 100 ml.
glacial acetic acid 2 ml.
Filter before use.
9-3% potassium dichromate, 5 min at room temperature.
10. Differentiate in 0·25% potassium permanganate, 5 sec.
11. Flood with Pal’s bleach, 5 min.
12. Wash, dehydrate (95% and absolute alcohol, 1 min in each), clear in xylene (2 changes, 1 min in each), mount in Canada balsam.
All photographs were taken with a Leitz Panphot apparatus, preliminary focusing being on the ground glass screen and final critical focusing by means of the telescope.
As in previous experiments the fresh sections were photographed by light from a carbon arc source, and the exposure time was standardized at 3 sec. The fixed sections were photographed with light from a tungsten filament lamp source, with an Ilford micro 3 filter No. 404 and an exposure time of 10 sec. Ilford R. 20 quarter-plates were developed in I.D. 11 fine grain developer for 6 min at room temperature, and subsequently enlargements × 1,000 and × 2,000 diameters were made on Kodak W.S.G. 2. S. paper and developed in D. 163 for 2 min. After processing, the prints were allowed to dry between sheets of Fotonic paper. A squeegee was not used. A micrometer scale was photographed and enlarged at each session.
The optimum exposure times were determined by a comparison between the dimensions of the direct visual image and the dimensions of the photographic images produced after different exposure times. Ten fibres between
11 and 21 μ in external diameter were selected, and measurements of internal and external diameter made directly on the slide with a screw ocular micrometer to the nearest 0·125μ, and indirectly with a ruler to the nearest 0·25μ on photographic enlargements after exposures of , 5, 7, 10, 15, and 20 sec. The ratio of internal to external diameter was used as a basis for comparisons by the χ2 test.
Comparison of external diameters
The non-fasciculated part of the N.G.M. was removed as above, and carefully cut into two segments. Sections from the two adjoining cut surfaces were prepared by the freezing technique and the Flemming-Wolter technique respectively. Estimates of external diameter were made by two methods. A Perspex disk inscribed with concentric circles 2 mm apart was used to allocate the magnified images of fibres to sizecategories differing in diameter by 2 μ steps. This method was chosen initially because it has frequently been adopted by other workers in this field. Since minor differences might not be apparent when a 2 μ grouping was used, a more accurate estimation was made for one N.G.M. by taking the mean of two measurements at right angles made with a 0·5 mm scale. Histograms were constructed and a comparison of the distributions made by the χ 2 test.
Comparison of sheath thicknesses
Sections were prepared by the two techniques and estimates of external and internal diameter were made on photographs at final magnifications of 1,000 diameters (10 to 20 μ fibres) and 2,000 diameters (<10 μ fibres). For each fibre the estimator was the mean of two measurements made at right angles. The maximum error of the measuring technique was ±4%.
For each type of preparation a scatter diagram was constructed by plotting external diameter against 2 × sheath thickness. Regression lines and standard errors of estimate were calculated.
The results from frozen sections and fixed stained sections were compared and the changes in sheath thickness and internal diameter demonstrated graphically.
Illustration of lability of sheath dimensions during processing
Frozen sections were prepared and photographed after each of the following successive stages:
Flemming’s fluid (15 min).
70%, 90%, 95%, and absolute alcohol (15 min each).
4. Cedarwood oil (15 min).
5. Molten paraffin wax (15 min) and return to xylene (15 min).
6. Through the alcohols to water.
7. Wolter’s haematoxylin staining; dehydration, clearing, and mounting in Canada balsam.
Twenty fibres, all approximately 15 μ in external diameter, were followed through the successive stages by measurements of internal and external diameter. The mean changes in dimensions were used to construct a diagram.
It was found that the thickness of the myelin sheath decreased with increasing exposure time; that this decrease was regular and due to changes in both internal and external diameter (fig. 1). The ratios of internal to external diameter for the 10 fibres as determined with the screw ocular micrometer were taken as the best estimates. The ratios as determined on photographic images after 10 sec exposure did not differ significantly from those estimates. This and other results are given in table 1.
Comparison of external diameters
It was found that there was no significant difference in distribution between sections prepared by the two techniques (table 2).
Histograms for the N.G.M. which was examined in more detail are shown in fig. 2.
Comparison of sheath thicknesses
It having been demonstrated that external diameter was not appreciably different after processing, a direct comparison of the regression lines of sheath thickness on external diameter gave a measure of reduction in sheath thickness.
Fig. 3, A is a scatter diagram of 2 × sheath thickness against external diameter for 204 fibres from fresh frozen sections. A calculated regression line and relevant statistics are given in fig. 3, B. A similar scatter diagram and regression line for 172 fibres from the fixed and stained preparation are given in fig. 4, A, B.
The regression lines for fresh frozen sections and for fixed stained sections are compared in fig. 5. Comparison of these lines shows that processing resulted in a reduction of sheath thickness of some 20% regardless of external diameter (fig. 6, A) or original sheath thickness (fig. 6, B). This reduction in sheath thickness was the result of an increase in internal diameter. It will be noted that within the range of fibres studied, as fibre diameter increased, the increase in internal diameter changed from some 50% to under 20% (fig. 6, c).
Lability of sheath dimensions
The results of this investigation are illustrated in fig. 7. The principal changes in sheath thickness occurred in the first and last stages of processing. External diameter was not appreciably affected, whereas the internal margin of the sheath was extremely labile. A typical 15-μ. fibre as seen in a section prepared by the standard Flemming-Wolter technique is included for comparison.
It is axiomatic that for quantitative histology a standardized technique is necessary. Different aspects of the technique requiring standardization will be considered.
The selection of a particular nerve and site is important in an investigation of this type. As has been pointed out, a mixed population of fibres to muscle and skin should not be used. Fasciculation has been shown to be accompanied by branching of myelinated fibres (Quilliam, 1956), which is clearly undesirable. The evidence concerning peripheral tapering of myelinated nerve-fibres has been reviewed by Sunderland (1958). Causey (1948) and Evans and Vizoso (1951) found no evidence of tapering along the greater part of the N.G.M. The latter authors, however, present data suggesting a gradual increase in the proportion of myelin peripherally. For these reasons, our sections were taken from a short segment (<5 mm) of the non-fasciculated portion of the N.G.M. of adult rabbits, 3 cm proximal to its point of entry into the muscle.
When removing the nerve for processing we have found it important not to dissect fine nerves from the parent trunk nor to remove the epineurium, as these manœuvres lead to stretching and distortion. Removal of the appropriate part of the sciatic nerve and its branches in one piece avoids this.
When attaching the nerves to card frames, attempts were made to restore their original length (Fernand and Young, 1951).
The effects on penetration, autolysis, and fibre contour of variations in the composition of Flemming’s fixative have been studied by Causey (1948), who found that 1 drop of glacial acetic acid delivered from a standard pipette gave optimal results. We have used a 25-mgm drop in 19 ml of chrome-osmium mixture for each specimen.
Evans and Vizoso (1951) found that the most critical factor in the staining process is the time allowed for differentiation in the potassium permanganate bath. This in turn is dependent upon density of staining. Staining in 100 ml of Wolter’s haematoxylin at 37° C for 15 h allows the differentiation time to be fixed at 5 sec. Wolter’s haematoxylin is made immediately before use and the variability of natural ripening processes is eliminated.
The experiment demonstrating the lability of sheath dimensions during processing showed that the major changes occurred during initial fixation and final dehydration and clearing before mounting. Frequently the latter stages are the least well controlled with respect to time. During this experiment reagents were poured directly on to the freshly cut surface of a microscopic section, which was attached to a slide. This situation is not directly comparable with the standard Flemming-Wolter technique and thus has little quantitative value. Nevertheless, it is believed to provide information of a qualitative type concerning the site of maximum lability (the internal border of the compact myelin) and the stages of processing principally concerned.
General measures have been taken to standardize the photographic technique according to the work of Rexed (1944). However, Evans and Vizoso (1951) point out that a serious error may be introduced by variations in the intensity of the light and in the exposure time. Using direct projection on to bromide paper they found marked differences in the ratio of myelin area to axon area when exposures were made for 15 sec and 30 sec, and therefore standardized their exposure time, presumably at an arbitrary level. We have determined and used the exposure time which leads to an image of dimensions comparable to those determined with a sensitive screw ocular micrometer. It is interesting to note that this exposure time corresponded to that chosen empirically and quite independently by an experienced photographer to give a ‘normal’ negative after development over the recommended interval for the emulsion used. No experiments with different emulsions were made.
Some importance attaches to the finding that external diameter was little affected by the reagents used in the Flemming-Wolter technique in view of the wide use of similar techniques for quantitative studies, involving external! diameter only.
Studies involving internal diameter have been more limited. They include aspects of growth (Evans and Vizoso, 1951) and regeneration (Sanders, 1948), as well as theoretical (Rushton, 1951) and practical (Gasser and Grundfest, 1939) functional studies. In this connexion it is necessary to consider the reduction in sheath thickness of some 20%, associated with increases in internal diameter varying with fibre size between 20 and 50%. Clearly a measurement of the internal diameter of the myelin sheath made on a fixedstained preparation bears no simple relation to the axon diameter.
Measurements made on such preparations do, however, give useful information on the behaviour of compact myelin under the conditions of the experiment. It has been shown that the Flemming-Wolter technique causes a constant size-change per unit of myelin thickness (fig. 6, B). This may be regarded as an expression of systematic change affecting the lamellar subunits of myelin. Such a view receives strong support from the work of Fincan (1958), who used X-ray diffraction techniques to study quantitatively the effects of reagents on the ultrastructural distances of nerve myelin. His basis of reference was the diffraction pattern of the freshly dissected nerve. The percentage change per radial repeating unit (after dehydration) was of a similar order to that described above. Furthermore, a number of substantial changes of proportion were shown to be reversible. This accords well with the observations on sheath lability during processing illustrated here.
We wish to thank Professor R. Warwick for his interest in this work, Mr. A. N. Finch for technical assistance and photography, and the Department of Medical Illustration, Guy’s Hospital Medical School, for preparing figs. 2-7. This investigation has been supported by grants from the Central Research Fund, University of London.