The Outwandering of Cells in Tissue Cultures of Nerves Undergoing Wallerian Degeneration

When a nerve is sectioned the whole peripheral stump undergoes Wallerian degeneration, and a small part of the central stump near the cut also shows degenerative changes. While axons and myelin degenerate during this process the Schwann (or sheath) cells proliferate and in the peripheral stump become arranged in the well-known ‘bands of Büngner’. The endoneurial fibroblasts also proliferate. A study of the outwandering from explants of degenerating nerves in tissue culture should throw light on the changes in Schwann cells and fibroblasts, the physiology of which is little understood. It may also be of practical importance in nerve surgery, for it seems that outward migration of the Schwann cells from the cut surfaces of the nerve, which will be reflected by their outwandering in vitro, plays an important part in ensuring adequate healing (Young, 1942).

Several workers have grown Schwann cells in vitro, notably Ingebrigtsen (1916), Chlopin (1939) and Murray, Stout & Bradley (1940). Of these only Ingebrigtsen was concerned with the changes in the cells during Wallerian degeneration. He studied the peripheral stump during the first 19 days of degeneration, using as a criterion of activity the percentage of explants which showed any outwandering. Our work may be considered as an expansion of Ingebrigtsen’s suggestive beginning. We have investigated regions of both stumps, degenerated for periods of up to a year, and we have used for estimating activity a method which is more easily interpreted than that of Ingebrigtsen. The present paper is a general survey of quantitative differences of cell outwandering in vitro between various regions of the severed nerve and between nerves at different times after they have been severed. We are much indebted to Mr J. Z. Young for suggesting that we should undertake this investigation, and for criticizing the manuscript.

The explants were taken from the sciatic nerve of adult rabbits. In most of the experiments the peroneal branch was used; in a few, the tibial branch. There was no significant difference of behaviour of the two branches. At an initial operation a piece of varying length was cut from the nerve with sharp scissors in the thigh region at about the level of the third trochanter of the femur. After a varying interval of time we made a second operation and removed lengths of the central and peripheral stumps of the nerve for culture. In each experiment we cultured a number of explants from each of several different parts of the stumps of the severed nerve, as listed below.

In most experiments our only provision to prevent reinnervation of the peripheral stump was to make the gap between stumps large ; but when nerves were left to degenerate a long time the possibility that reinnervation had occurred was investigated by treating histological sections by Bodian’s method for staining of axons. In none of the peripheral stumps used for this paper were more than a very few isolated axons found.

In some experiments the central stump was killed at the initial operation to ensure that it did not reinnervate the peripheral stump. This was done by injecting the central stump with a 1 % aqueous solution of crystal violet (a method of axon destruction used by Guttmann & Medawar, 1942). The method was not entirely suitable for our purpose, because we found that the proximal part of the peripheral stump was sometimes a little damaged.

The ordinary hanging-drop method of tissue culture was used. Two explants were placed on each cover-slip. The medium consisted of fowl plasma and chick embryo extract. Through most of the series of experiments we made the extract from a dried embryo brei (see Peacock & Shukoff, 1940) kindly prepared for.us by Dr E. Chain of the Pathology Department, Oxford University. We used this at a concentration representing 20 % of pure embryo juice, the diluent being Pannet and Compton’s saline. The dried preparation was reasonably constant in action during the time it was in use. It was already several months old when we made most of our experiments, and had lost some of its original power. In a confirmatory set of experiments we used fresh embryo extract, which was considerably more stimulating. The epineurium was removed from the nerve before culture. The nerve was then divided into explants about 1 mm. long, consisting of the whole or half of the cross-section. Cultures were kept at 38° C., and in most experiments were cultured 4 days, in some cases 3–7 days, before fixing. Conditions of culture were kept as constant as possible throughout the series of experiments, and the different times of degeneration we used were well randomized with respect to dates at which we made the experiments. In many experiments normal and degenerated nerve, or nerves of different periods of degeneration, were cultured simultaneously.

Cultures were fixed in 4 % formaldehyde and stained in Ehrlich’s haematoxylin. Wandering cells, fibroblasts and Schwann cells occurred in the zone of outwandering. The wandering cells were sporadic in occurrence, and we do not report on them here. Fibroblasts were sometimes very numerous, and the amount of their outwandering was estimated by eye, using an arbitrary scale with 4 divisions, the index 4 representing the highest number. The amount of Schwann cell outwandering was obtained by.counting the Schwann nuclei in the zone of outwandering. This could be fairly accurately done because the number of nuclei in most cultures was small and it seldom exceeded 1000. Since mitosis of Schwann cells in the zone of outwandering is very rare the nuclear count corresponds closely to the number of nuclei which have wandered out. The length in profile of the transverse cut surface of the explant was measured, and the Schwann cell activity of a given region of nerve was then expressed as the mean number of Schwann nuclei which have wandered out per mm. of this measurement (usually during 4 days in vitro). Such a standardized measure of activity, even though imperfect, is an advantag since there was considerable variation of the size of the nerves, but it was not worth applying to the rough estimates of fibroblast activity. It may be added that the nuclear count reflects approximately the distance the Schwann cells have migrated into the clot.

The Schwann cell in vitro has been described by Ingebrigtsen (1916), Chlopin (1939) and Murray et al. (1940). It is unnecessary to add anything to these descriptions. At a later date we intend to publish a detailed account of the cytology of the Schwann cell in vitro. It is not difficult to distinguish the majority of Schwann cells from fibroblasts; but difficulties sometimes arise when separating very elongated fibroblasts, such as may occur in the substance of the clot, from the more richly cytoplasmic Schwann cells. A series of intermediate forms can be traced. So the decision is sometimes arbitrary; but these doubtful cells are very few in number compared with the clearly identifiable cells. Since it may have introduced a systematic error into our results, a more serious difficulty is that a new type of cell appears in the cultures from long-degenerated stumps. Although typical Schwann cells and typical fibroblasts still occur, many of the cells are of extraordinarily ramifying and straggling form. They have characteristics of, and grade into, both typical Schwann cells and typical fibroblasts. We believe them to be Schwann cells and have counted them as such; but if we are mistaken, our counts of Schwann cells from 139 days onward are over-estimated by about 50 %.

In the conditions of culture used in these experiments, normal (i.e. not predegenerated) nerve is very inactive. Nineteen experiments with normal nerve were made, and of these thirteen showed no outwandering whatever. Of the 213 explants cultured in these experiments, 191 (90 %) were blank after 4–6 days in vitro ; the remaining twenty-two explants, which were confined to six of the experiments, realized between them a total of thirty Schwann nuclei and a few fibroblasts. Half of this total came from explants taken at different times from the right and left nerves, of one particular rabbit. The mean number of Schwann cells per mm. of explant was about 0·1 and of fibroblasts about 1, for all normal nerve cultures.

One of the active nerves was sectioned and examined histologically. A group of degenerated fibres was found in it. The activity shown by some normal nerves can perhaps be accounted for by the occasional occurrence of a few degenerated fibres.

In view of the sporadic occurrence of activity in normal nerve, we made, when necessary, cultures of normal nerve from the same rabbit simultaneously with those from previously cut nerves. These controls were always blank. For purposes of comparison with previously cut nerves, normal nerve can therefore be considered inactive.

Several different regions of the peripheral stump were cultured, the nerve having always been cut about half-way down the.thigh. The regions used were as follows :

Bulb, the terminal swelling derived from the hernia produced by the retraction of the perineurium and epineurium from the nerve fibres at the cut surface.

Traumatic, a region 5 mm. long immediately peripheral to the bulb.

Standard, the region stretching from the traumatic region to the biceps blood vessels.

Knee, the peroneal nerve at knee level.

Shank, anterior tibial branch of peroneal in. shank.

Schwannoma region, the connective tissue immediately central to the bulb, into which cells migrate from the bulb forming eventually a tumour-like mass.

Most of our data are derived from the traumatic and standard regions, and in the early stages of the work these two regions were not distinguished. So we shall first consider the traumatic and standard regions jointly.

Changes of Schwann cell activity with time of degeneration shown by these regions of the nerve are given in Table 1 ; they are typical for the rest of the peripheral stump. Only explants cultured 4 days in vitro are included. The results of eighteen other experiments, in which explants were cultured for a longer or shorter time, substantiate these changes. Column 4 gives the mean number of Schwann nuclei outwandered per mm. of explant and these values are plotted against time of degeneration in Fig. 1. The experiments indicated by * in Table 1 were made more recently and all within a short time of each other (four were made simultaneously on rabbit O). They differ from previou experiments in that freshly prepared embryo extract was used. The extract proved to be more stimulating than the dried form, so the results cannot be combined with the previous ones. They are shown in a separate curve in Fig. 1.

Fig. 1 is of course constructed with data from many different nerves, but it may be considered to show approximately changes of activity with time in a single nerve after section. After an initial period of inactivity or of slight activity (the time of onset of activity is analysed below), a rapid rise starts at about 4 days of degeneration and carries the activity to a peak at about 19–25 days. At the peak, activity is about 40 times as great as at 4 days. After the peak there is a decline,, probably at first nearly as steep as the rise, but soon becoming less steep ; after about 60 days the fall is very slow. After 1 year of degeneration activity is still considerable, being 5–7 times as high as at 4 days.

Since it is clear that there is considerable variation of experimental conditions and from animal to animal, useful confirmation of the general trend of the curve is provided by those experiments in which two nerves from the same animal but of different degeneration times were cultured simultaneously, and therefore in conditions as nearly identical as possible. The results of the ‘paired’ experiments (indicated in Table 1, column 2, by the rabbit’s protocol designation) were as follows (the figures refer to days after cutting ; the sign indicates which time had the greater outwandering): 1 < 3; 2 < 4; 10 < 20; 16 < 25; 39 > 60. In one case (rabbit O) peroneals and tibials were simultaneously cultured; of the peroneals, 25 > 97, of the tibials, 20 > 35; while comparing a tibial with a peroneal, 20 < 25 and 35 > 97. The difference in activity between the members of a pair was in three instances statistically significant; between 10 and 20 days of degeneration (mean activity and standard error 71 ± 20 and 462 ± 34 respectively); between 20 and 35 days of degeneration (392 ± 67 and 188 ± 26 respectively); and between 25 and 97 days of degeneration (470 ± 57 and 130 ± 12 respectively).

It is known that the region of the peripheral stump imrpediately adjacent to the cut undergoes precocious degeneration (Cajal, 1928). The effect of this on Schwann cell activity was investigated in twenty-two experiments comprising 130 traumatic and 166 standard explants.

In each experiment explants from the two regions of a single nerve were cultured simultaneously, and their activity compared. We found that the mean activity of the traumatic region explants was greater than that of the standard region explants in seventeen of the experiments (owing to the small numbers of explants counted, however, the difference was only rarely statistically significant in an individual experiment). In four other experiments the mean activities were the same. In the remaining experiment, at 1 day of degeneration, we found no activity of either region. In the early stages of degeneration (2–4 days), the mean activity of the traumatic region explants is very much higher (4–20 times) than that of the corresponding standard region. With further degeneration it is a small but rather constant amount higher, averaging 154 ± 16 % of that of the standard region, a significant difference. It is clear that the raised activity of the traumatic region persists for 2–3 months after the initial cut ; and possibly, though our data are insufficient, for a year.

We found indications that a traumatic effect was also produced by recutting a nerve which had already degenerated. Two experiments of this kind were made. A nerve of 15 days’ degeneration was recut and cultured after a further 18 days, when the new traumatic region was 510 % of the standard region; similarly a nerve recut after 30 days’ regeneration and cultured after a further 18 days had a new traumatic region 145 % of its standard region.

The fact that the activity of the traumatic region is so markedly in advance of that of the standard region during the early stages of degeneration calls for more detailed investigation of the time of onset of activity in the two regions after the initial cut.

After 4 days in vitro both standard and traumatic regions of one day of degeneration were entirely inactive. At 2 days of degeneration the standard region was inactive, but the traumatic region had produced an average of 0·5 Schwann nucleus per explant. At 3 days of degeneration the standard region produced an average of 0·2 Schwann nucleus per explant, the traumatic 3·5. Thereafter the activity of both increased, the traumatic region maintaining its lead, but proportionately a much reduced one.

A longer cultivation of the apparently inactive nerves was undertaken. Even after 8 days in vitro both regions of the stump at 1 day of degeneration are quite inactive. But while no activity was detectable in the 2-day degenerated standard region after 4 days in vitro, slight activity was shown after 7 days in vitro (mostly in the form of filaments of Schwann cell cytoplasm, without nuclei). Strictly parallel.experiments done with normal undegenerated nerve showed complete inactivity after 7 days in vitro. It can therefore be inferred that the traumatic region becomes active at a slightly earlier tima of degeneration than the standard region (probably late on the first day), and that the standard region first acquires activity on the second day of degeneration.

The terminal bulb of the peripheral stump was not used to derive the curve of Schwann cell activity shown in Fig. 1 owing to its variability and the frequency with which the Schwann cells are obscured by fibroblasts. In spite of considerable variability the changes of the Schwann cell activity of the bulb with time of degeneration repeat in general form the course of activity shown in Fig. 1 ; activity starts on the second day, and from the fourth day rises fast to a peak at 15–25 days after cutting, thereafter declining. During these changes, but excluding the very long degeneration times of 344 and 402 days, the bulb is approximately as active as the traumatic region (though it varies between 50 and 400 % of this region) ; it is therefore (in twelve out of fourteen comparable experiments) more active than the standard region. However, in the two experiments with nerves of very long degeneration (344 and 402 days) the activity of the bulb was about one-fifth or less of that of the rest of the peripheral stump. The bulb was studied in 25 experiments, comprising 90 explants.’

The standard, traumatic and bulb regions discussed above are all parts of the sciatic nerve in the thigh. It seemed possible that activity might differ in the more peripheral parts of the nerve. In several experiments in which the nerve was cut in the usual place, pieces were taken for culture from the peroneal nerve at knee level (16 experiments, 100 explants) and from the anterior tibial branch of the peroneal half-way down the shank (9 experiments, 52 explants).

In both regions the changes of activity with time of degeneration are approximately the same as those of the thigh region, though a good deal more irregular. There is the same rise to a peak, at 20–23 days’ degeneration, and the same fall after the peak, at first rapid, then slow. The general level of activity proved to be higher than that of the standard region of the same nerve in nine out of eleven experiments on the knee region and in all eight experiments on the shank where comparison is possible. When the activities of the knee and shank regions were expressed as percentages of those of the standard region, much the highest values were obtained in the experiments at the shortest degeneration times. These were for the knee region 640% at 5 days and 500% at 7 days: for the shank 500% at 5 days and 760% at 10 days. At longer degeneration times the percentages decreased. They varied for the knee region (13–402 days) between 30 and 200%, the mean and its standard error being 120 ± 19%; and for the shank region (21–97 days) between 112 and 200%, the mean and its standard error being 210 ± 26%. In corhparable experiments the activities of the shank and knee regions were not significantly different from each other.

It should be noted that the shank region, separated by about 8 cm. from the cut in the nerve, cannot be affected by the initial operation except in so far as Wallerian degeneration is concerned.

It is known that Schwann cells migrate out from the cut surface of the peripheral stump (i.e. the bulb) in vivo, and infiltrate the neighbouring connective tissue, forming a tumour-like mass (Nageotte, 1913) m the later, stages of degeneration. We cultured the tissue from immediately beyond the tip of the bulb after various times of degeneration to test the activity of these emigrant Schwann cells. Twenty-two such experiments comprising 105 explants were made.

Explants from the schwannomå region are of variable activity, because it is not homogeneous nervous tissue, and some of the explants turn out to be connective tissue only. For this reason detailed quantitative treatment is not useful. But the general trend of activity with time of degeneration is clear. No Schwann cells appeared from explants taken from nerves of 6 days’ degeneration or less (five experiments). At 7 and 11 days of degeneration only very few Schwann cells were observed (two experiments). But from 13 to 25 days of degeneration (seven experiments) there were many explants which produced abundant Schwann cells, fully as many as the bulb or traumatic region. With longer degeneration (60-139 days, six experiments) activity is less, far below that of the peripheral stump proper, e.g. at 139 days the schwannoma region had an activity 20% of that of the bulb. Finally, in the experiments at 344 and 402 days of degeneration, when a massive tumour-like schwannoma had formed, the explants were almost completely inactive. At 344 days the mean number of Schwann nuclei (twelve explants) was 1 % of that of the peripheral stump proper ; at 402 days no Schwann cells out-wandered (five explants). In these experiments the bulb had also only very slight activity.

It is notable that when activity is high Schwann cells often grow from the schwannoma explants in the form of isolated massive trunks, which dissociate into separate cells as outwandering proceeds. This is presumably because the schwannoma contains thick, cords of Schwann cells.

As we might expect, the latency (i.e. the period between the setting up of the cultures and the first appearance of outwandering) varies inversely, with the amount of activity of the explant. (Schwann cells wander out before or at the same time as fibroblasts in all regions except the bulb and schwannoma.) Changes of latency are not, however, the only cause of changes of activity with time of degeneration. There are also changes of the rate of outwandering when the period of latency is over. The outwandering per mm. of explant per day of the period of outwandering changes in a way generally similar to the activity for 4 days in vitro (Table 1, column 4) so far considered. In particular, there is a fall of rate similar to the fall of activity after a peak at 19–25 days of degeneration. In fact, latency is negatively correlated with rate of outwandering and changes in both together lead to the observed changes of amount of activity.

But the relationship between latency and rate of outwandering is not constant at all times of degeneration. We have compared explants of the same rate of outwandering but from nerves of different degeneration times. (Explants with a period of outwandering of 2 or 3 days’ duration were used; they had been in vitro 3–6 days, according to their period of latency.) We found that, in the later (post-peak) stages of degeneration, latency is, for the same rate of outwandering, greater than in early degeneration. For a rate of between 20 and 60 Schwann nuclei per mm. per day, nerves of 7–16 days of degeneration have a mean latency of 1·3 ± 0·13 days (nineteen explants) and of 33–402 days of degeneration a mean latency of 1·7 ± 0·07 days (fifty-seven explants), the difference being significant. Similarly, for a rate of outwandering of between 60 and 90 Schwann nuclei per mm. per day, nerves of 23–30 days of degeneration have a mean latency of 1·1 ± 0·04 (thirteen explants) and of 33–139 days’ degeneration a mean latency of 1·7 ± 0·14 (fifteen explants), the difference being significant. Thus both the pre-peak and the peak explants have a shorter latency for the same rate of outwandering than the post-peak explants. Our data do not conclusively show that there is-a progressive increase of latency during the post-peak decline of activity. With a rate of outwandering of 20–60 Schwarm nuclei per mm. per day, explants from nerves of 33–60 days of degeneration have a mean latency of 1·6 ± 0·10 (twenty-nine explants) and those from nerves of 139–402 days of degeneration a mean latency of 1·8 ±0·074 (twenty-eight explants). The difference is hardly significant (P= 0·05).

The fibroblasts are often too numerous to count, and the fact that they are intermixed with Schwann cells makes estimation by area of outwandering inaccurate. Only an approximate analysis of their outwandering can therefore be made, using an arbitrary scale of 4 indices, the index 4 representing the highest number of cells. The fibroblast activity of a nerve is expressed as the mean fibroblast index per explant.

The changes in the mean index of fibroblast activity with time of degeneration are shown in Table 1, column 5. Column 4 shows the Schwann cell activity of the same explants. The general trends of the Schwann cell and fibroblast figures are closely similar, though the fibroblast activity is a good deal more variable both between explants and between experiments. At different degeneration times the proportion of fibroblasts to Schwann cells shows no significant changes. With respect to the time of degeneration at which activity begins there is no discernible difference between fibroblasts and Schwann cells.

When traumatic and standard regions are considered separately we find that the changes of their fibroblast activity follow substantially those shown for the two regions jointly in Table 1. The difference between fibroblast activities of the traumatic and standard regions parallels the difference between their Schwann cell activities. At 2–4 days of degeneration the fibroblast activity of the traumatic region is considerably higher (7–50 times) than that of the normal region, while at later stages it is only about twice as high.

In the bulb the activity of fibroblasts does not follow that of the Schwann cells. There is a rapid rise of fibroblast activity in the early stages of degeneration. In the experiments at 4 days of degeneration fibroblast activity was already half the maximum reached. Thereafter there is a slow rise accompanying the rapid rise in Schwann cell activity. In the post-peak period of Schwann cell activity there is no sharp fall in fibroblast activity which is well maintained for at least 4 months of degeneration. Cultures after 344 and 402 days of degeneration were, however, practically inactive in fibroblasts, as they were in Schwann cells. In five experiments for which satisfactory comparable data are available, the bulb is twice as active in fibroblasts as the traumatic region, and 4 times as active as the standard region. Thus the ratio of fibroblasts to Schwann cells is about $112$–2 times as big as in the traumatic or normal regions.

The fibroblasts of the knee region are like those of the standard and traumatic regions in that they follow the Schwann cells in their changes of activity with time of degeneration; but they differ in that (in eight out of ten experiments) they form a relatively higher proportion of the outwandering—roughly $112$ times as many fibroblasts per Schwann cell as in the traumatic and standard regions. The absolute level of fibroblast activity is about twice that of the standard region.

Data for the shank region are not sufficient to show anything but that the usual curve of activity with time of degeneration is followed, and that the proportion of fibroblasts to Schwann cells and the absolute level of fibroblast activity are probably higher than in the standard region.

Fibroblast outwandering in the schwannoma is variable like Schwann cell activity. Its changes with time of degeneration are rather similar to those of the bulb. At 4, 5, and 6 days after cutting a few explants already showed considerable fibroblast (and no Schwann cell) activity. Activity probably increased further up to about 20 days. The subsequent decline was much more marked than in the bulb, and activity was poor in experiments at 2 months after cutting. Almost complete inactivity was reached at 344 and 402 days after cutting. The activity we obtained from explants of the peripheral bulb suggests that many of the fibroblasts of the schwannoma will have wandered there from the cut surface of the stump.

The following regions of the central stump were cultured:

Bulb, which has the same position and origin as the bulb of the peripheral stump.

Distal region, extending for 3–4 mm. above the bulb.

Proximal region, central to the distal region, separated from it by $112$–2 cm.

Neuroma region, the connective tissue immediately distal to the bulb, into which cells and axons migrate from the bulb, forming eventually a tumour-like mass, corresponding to the schwannoma of the peripheral stump.

In the early stages after cutting, activity of the Schwann cells in the central stump is confined to the region immediately near the cut end. This part, consisting of the bulb and distal regions, is so small that the data have been grouped for comparison with the peripheral stump. The changes of activity with time after cutting are shown in Table 2 and Fig. 1. The explants which supplied these data were all 4 days in vitro with the exception of those at 344 and 402 days of degeneration, which were longer (their values at 4 days were approximately half those given).

Changes of activity with the course of degeneration are not nearly so marked as in the peripheral stump. Activity is first found at 2 days of degeneration. After 4 days of degeneration it has already reached about half its maximum. There does not appear to be a sharply defined peak of activity as in the peripheral stump, but rather a plateau of raised activity. A series of ‘paired’ experiments in which nerves of the same rabbit, cut at different times previously, were simultaneously cultured, together with some other cultures kept longer than 4 days in vitro and therefore not included in Table 2 and Fig. 1, suggest that the highest part of the plateau may be early, between 5 and 10 days after cutting. The results of the ‘paired’ experiments were as follows (the figures refer to days after cutting, the sign indicates which time had the greater outwandering); 1 <3, 2 <4, 7 >13, 16 >25, 23 >400, and 3 >344. Of the seven experiments in which the explants were kept in vitro for more than 4 days the two highest mean activities occurred at 6 and 7 days after cutting. But although the position of the highest point in the activity curve is uncertain (and it is quite possible that different nerves may differ widely in this), it is clear that the maximum activity of the central stump is less than 10% of the maximum activity reached by the peripheral stump. The data indicate that with increasing time after cutting there sooner or later sets in a very slow decline in activity. A year after cutting, activity is still present to the extent of about 10% of the maximum.

When the activity of the central stump is compared with the activity of the corresponding region, cultured at the same time, of the peripheral stump (i.e. the peripheral bulb and the traumatic regions) of the same nerve a clear change with time after cutting is apparent. At 2 days after cutting in one experiment the central stump mean was 900% of the peripheral stump mean, in a second experiment 250%. At 3 days after cutting it was 143 % in one experiment, 5 % in another. At 4 days after cutting it was 186% in one, 95 % in another. At 5 days after cutting it was 23 %. Thereafter it never exceeded 12 % of the peripheral stump. Activity is therefore much greater in the central stump than in the peripheral Stump in the early stages after cutting. We have not shown that it starts earlier. Likewise the peripheral bulb and traumatic regions are much more active in this stage of degeneration than the standard region; the central stump is therefore far more active than the standard region of the peripheral stump. But when the peripheral stump starts its rapid rise at 4 days of degeneration, it quickly overtakes the central stump, and is thereafter more active: at its peak at 19–25 days, it is 20–100 times more active than the central stump; and after a year, about 10 times.

The central bulb and the stump immediately above it separately show the same behaviour as that described for them jointly. They differ in that the bulb usually has the higher activity. In thirteen out of nineteen experiments the bulb is ahead, averaging in all about twice the activity of the stump proper.

We found that the proximal region, more central than the distal region so far discussed, was inactive in the early stages after cutting but that later it acquires some activity. This region was cultured in 18 experiments comprising 125 explants. Of nine experiments, covering 1–19 days after cutting, only one showed a trace of activity: this consisted in the appearance of a few cytoplasmic processes. Of nine experiments, covering 23–402 days after cutting, only one shows no activity, and this was cultured 3 days after cutting. All experiments made after 23 days show activity in this region of the central stump. Normal (undegenerated) nerves were cultured simultaneously with four of the eight positive cultures. All normal explants were entirely inactive.

The activity of this proximal region of the stump is very small, however, and has a very long latent period. Adequate comparison with the more distal part of the central stump is not usually possible, because the latter was fixed after about 4 days in vitro, and at this time the former was still blank. In three experiments, at 23, 344 and 402 days of degeneration, in which all regions of the central stump were fixed at the same time, the proximal central stump averaged 35 % of the distal.

The neuroma region was cultured in nine experiments comprising sixty-four explants. In experiments at 15, 23, 23, 35, 53 and 139 days after cutting, the neuroma region showed always a moderate degree of Schwann cell activity, sometimes more, sometimes less than the central bulb. Activity is far less than that of the schwannoma region; actual figures are available only for one experiment, at 139 days, and in this there were 10 times as many Schwann cells in the schwannoma region. The other experiments showed a similar order of difference. Three massive neuromata, taken from nerves cut 298, 344 and 402 days previously, were quite inactive, showing that there is a decline in activity with long degeneration, as in the schwannoma.

In the central bulb the activity of fibroblasts at different times after cutting of the nerve is very similar to that of the peripheral bulb. The highest activity we found was at 4 days after cutting, and up to 2 months after cutting little decline was apparent. In central bulbs cultured a long time after cutting (139, 344 and 402 days) however no fibroblasts outwandered. In the central bulb there is, compared with the peripheral bulb, higher fibroblast activity, just as there is higher Schwann cell activity, between the second and fifth days after cutting. But after 5 days of degeneration the central bulb has, in nine out of ten experiments, a slightly lower fibroblast activity, averaging 80% of that of the peripheral bulb. Later than 5 days after cutting, the ratio of fibroblasts to Schwann cells is much higher in the central than in the peripheral bulb, for the fibroblasts in the central bulb are only slightly less active thair those in the peripheral bulb, while the Schwann cells are far less active. Up to 5 days after cutting the ratio is about the same in the two bulbs.

The active terminal part of the central stump (distal region) just above the bulb shows exactly the same trends as the central bulb. Fibroblast activity is very low or absent beyond 3 months after cutting. When compared with the corresponding region of the peripheral stump (traumatic region) its activity is found to be equal, or slightly higher in the first few days after cutting, but subsequently always lower, averaging 30% of that of the peripheral traumatic region.

The slight Schwann cell activity which appears late after cutting in the proximal part of the central stump is not associated with any fibroblast activity.

Our data on fibroblast activity of the neuroma are too incomplete for comparative purposes, but a decline to zero is fairly well established about a year after the cut. Between 15 and 53 days after cutting there were always considerable numbers of fibroblasts in the outwanderings (five experiments), at 139 days there were very few, and at 298, 344 and 402 days there were no fibroblasts, as there were no Schwann cells.

We have found that when explants from the stumps of a severed nerve are cultured in vitro, the number of Schwann cells and fibroblasts which wander out varies according to the region of the stumps from which the explants are taken and the length of time which has elapsed since the original section was made. Normal (i.e. not previously severed) nerve is usually completely inactive ; so, we found, was nerve taken 1 day after it had been cut, whatever the region cultured. On the second day after section a very slight amount of activity is apparent in both peripheral and central stumps, and up to the fourth day corresponding regions of the two stumps show a similar, usually small, increase in activity. With longer periods between section of the nerve and culture of the stumps further changes of activity occur, but differently in the two stumps. In the peripheral stump (none or very little reinnervation having occurred) all regions tested, including parts 8 cm. away from the original cut, show a very rapid rise in Schwann cell and fibroblast activity, beginning at the 4th-5th day (except for the bulb fibroblasts, which show an earlier rise) and reaching a high peak at about the 20th-25th day after section. This initial rise of activity is in agreement with the results of Ingebrigtsen (1916) who, investigating the percentage of explants showing any activity, found that this index increased from 17% on the 5th day after section to 82% on the 19th day. He found no outwandering before the 5th day. Ingebrigtsen probably used a medium of far less growth-promoting power than ours. He did not experiment on nerves which had undergone longer degeneration, nor on central stumps. In peripheral stumps taken more than 25 days since the nerve was cut we found a fall in activity with time since section, rapid at first but soon becoming slower. At 2 months after cutting the level of activity is back to that of the 10th-12th day after cutting (one-third of the peak activity) while during the next 10 months it falls another 50%, to the same activity as that of roughly the 7th day after cutting.

In the central stump on the contrary, activity is at first confined to the few mm. at the cut end, and shows no rapid changes with time since cutting. After an initial rise a maximum is reached probably between the 5th and 10th days, but at a level which, except in the case of bulb fibroblasts, is negligible compared with that of the peripheral stump (equivalent perhaps to the peripheral stump at the 5th day); activity then slowly falls, but is still not zero after one year. After 23 days since section and up to the longest time tested (402 days) we found that the central stump 1-2 cm. proximal to the cut, previously inactive, developed a very slight activity.

The spatial distribution of the activity corresponds to that of the degeneration of the nerve fibres, which as is well known starts in the whole length of the peripheral stump and in the tip (1–2 cm.) of the central stump soon after the nerve is severed (see Cajal, 1928). It must be assumed that activity is the direct result of this degeneration. The occurrence of a very small amount of activity occasionally in undegenerated nerve and consistently in the more central parts of the central stump during the later stages after cutting, may be independent of degeneration. But we believe that the occasional activity of normal (undegenerated) nerve is no more than can be accounted for by the occurrence of a few degenerate fibres (Duncan, 1930) and in one instance we verified the presence of degenerate fibres in an active normal nerve. As for the proximal part of the central Stump, it is said (see Spielmeyer, 1929) that when no reinnervation occurs a slow degeneration or atrophy of individual fibres sets in, which would perhaps account for the slight activity which we found to occur after the 23 rd day after cutting. Degeneration of Wallerian type is probably rare here, the changes being rather a diminution of fibre size, including the elimination of part of the myelin.

It is notable that in the peripheral stump (except in the bulb) where the changes of activity are pronounced, both Schwann cells and fibroblasts change in substantially the same way with time of degeneration (see Table 1). Both are presumably affected, directly or indirectly by the same changing stimulus. It is a reasonable hypothesis that this stimulus is a chemical one which is present during the early stages of Wallerian degeneration, contemporaneously in fact with the destruction of the major part of the nerve fibre. This occurs between the 4th and 20th days of degeneration, when the rise of activity is maximal. The stimulus presumably ceases to be present from about the 25th day onwards, and the result is a sharp fall in activity, not to zero, but, for reasons as yet unclear, to a moderate and slowly decreasing level. The persistence of this low level of activity is peculiar to the atmosphere of the nerve stump. The cells which out-wander from the cut end of the peripheral stump in vivo, and eventually form the schwannoma, show the same rise to a peak of activity during the early stages of degeneration; but the subsequent fall is more rapid and activity is zero after a year of degeneration. The almost complete absence of outwandering in these schwannomata of long standing is doubtless connected with the fact that they are extremely fibrous, with few cells, as sections of the explanted material showed. Holmes & Young (1942) suggest that in such schwannomata the Schwann cells which originally formed part of them may have atrophied.

Since the terminal 1–2 mm. of the central stump undergoes a degeneration which is quite similar to the Wallerian degeneration of the peripheral stump, it might be expected to show the same changes of activity with time since cutting as the peripheral stump. That it does not do so can probably be ascribed to the presence of growing axons, which is the most important difference between central and peripheral stumps. But although the growing axons perhaps prevent the activity in the central stump from reaching a high level, they do not reduce it to zero even a year after cutting. The persistence of activity in the central stump recalls that in the peripheral stump, although it is at a much lower level; and as in the schwannoma, the cells which have wandered out from the cut end of the stump to form the neuroma do not maintain their activity after long degeneration.

The $12$ cm. of the peripheral stump immediately adjacent to the cut has a higher activity than the next $12$ cm. during at least the first 97 days of degeneration. It is known that degenerative changes begin earlier in the first few mm. of the peripheral stump, and in the corresponding region of the central stump, than in the rest of the peripheral stump. Correspondingly, activity is precociously high in these regions (particularly so, for an unknown reason, in the central stump) ; and, as a result, during the rise in activity in the peripheral stump, the first $12$ cm. is more active than the second. The fact, however, that activity remains higher in the first $12$ cm. at and after the peak of activity, instead of undergoing a precocious decline, must mean that the stimulus given to the cells in this region is not only precocious but also greater. This is also indicated by the fact that stimulation was obtained in the neighbourhood of the wound when an already degenerated nerve was recut (though we have only done two experiments to test this) There appears to be therefore a distinct and additional traumatic stimulus, similal perhaps to that which activates fibroblasts when connective tissue is wounded, which is superimposed on the Wallerian activation near the cut end of the peripheral stump. It must be assumed, since the trauma affects the ends of the two stumps equally, that the condition is the same in the central stump.

The traumatic activation affects fibroblast and Schwann cells equally in the peripheral stump just below the terminal bulb. But in the bulb itself the number of fibroblasts is increased proportionately to the number of Schwann cells, especially in the early stages of degeneration (znd-5th days). This is also true of the central bulb, where the number of fibroblasts is almost the same as in the peripheral bulb, although the number of Schwann cells (after the first 3 days of degeneration) is far smaller. The simplest explanation is that there is an early invasion of fibroblasts, stimulated by the wound, from the perineurium or other adjacent connective tissue, through the cut surface of the stumps into the bulbs. Later the direction of the migration is of course reversed, and cells leave the bulbs to form the neuroma and schwannoma. The schwannoma, like the bulb, shows a high fibroblast activity in the first few (4–6) days of degeneration, with no Schwann cell activity until after the 6th day. Whether these are the fibroblasts which invade the bulb is unknown.

We found that after degeneration of a year, the peripheral bulb had lost far more of its activity than had the rest of the peripheral stump. This is no doubt connected with the extensive fibrosis of the Schwann bundles which occurs in late degeneration close to the lesion (see Holmes & Young, 1942). The fibrosis in its turn may be related to the intense fibroblast activity characteristic of the bulb region in the earlier stages of degeneration. The loss of activity in the schwannoma is perhaps to be similarly explained.

Clearly the changes in activity which the Schwann cells and fibroblasts show in vitro with increasing time of degeneration is a result of changes in the physiology of the nerve, which result from degeneration. We do not yet know anything further about these physiological changes, and it is not possible to correlate them clearly with the known changes of these cells during degeneration in vivo. Some correlation of mitotic activity and outwandering would be expected, since they have important features in common (pseudopodial activity, response to same stimuli in tissue culture). The literature suggests that the temporal correspondence in the changes of the two activities is not closes mitosis starting in the Schwann cells and probably in the endoneurial fibroblasts at the 4th day of degeneration, reaching its maximum at about the 15th-20th days and ceasing after the ijth-2oth day (Cajal, 1928). Roughly speaking, however, the rise to the peak of outwandering activity coincides with the total duration of mitotic activity. It is probable that some migration of the Schwann cells takes place in the degenerating nerve in vivo, for instance during the formation of the Büngner bands (Holmes & Young, 1942).

However, it is highly probable that the changes in activity found in vitro will be directly reflected in one aspect of the behaviour of the cells in vivo: the outwandering from the cut surface of the nerve stumps into the surrounding connective tissue. The changes with time of this outwandering in vivo during the building up of the neuroma and schwannoma cannot be directly inferred from our experiments. But if a degenerated peripheral stump is recut, the new in vivo outwandering from the cut surface will, it is highly probable, vary according to the curve of activity with degeneration time which we have obtained in vitro. The actual amount of outwandering will not necessarily correspond, owing to the difference of medium in which the cells grow; and the proportionate changes may not be exactly the same, owing to possible differential effects of the media; but it is unlikely that the general trend will differ significantly. Holmes and Young (1942), by measuring the length of the schwannoma produced from a cut nerve in vivo, after various degeneration times, have in fact obtained direct evidence on this point which is in general concordance with our conclusions.

Does the activity shown during the time spent in vitro represent the physiological state of the nerve at the moment of explanation : or do the processes of Wallerian degeneration proceed in vitro as they would in vivo, correspondingly increasing the activity of the cells? It is clear that if activation does proceed in vitro, it does not do so at the in vivo rate. Normal, i.e. undegenerated nerve (with a few exceptions) and nerve on the first day of degeneration, do not develop activity even though kept in vitro quite long enough (8 days and more) to do So if they changed at the rate they do in vivo. Deterioration of the medium is not the cause of this, since subculturing during this period also failed to elicit any outwandering. An experiment in which explants of an early stage of degeneration were subcultured after 4 days in vitro, and then grown another 4 days, showed that there was a slight rise in activity during the second period, but only a fraction of what would have occurred in vivo. We conclude therefore that, at least in the early stages of degeneration, activity in vitro represents fairly closely the physiological state of the nerve at the time of explanation.

It has been pointed out that Schwann cell activity in vitro may be expected to reflect the Schwann cell outwandering from the cut end of the nerve in vivo. In the repair of nerves by suture or graft a successful junction is probably formed by a vigorous out-wandering of Schwann cells from the peripheral stump or from the graft (Young, 1942; Holmes & Young, 1942). We find that, except for a short initial period, the longer a nerve is left before surgical repair is undertaken, the less active the Schwann cells of the peripheral stump are likely to be in forming a junction. In the rabbit the optimum time for repair from the point of view of Schwann cell outwandering is not later than 25 days after the initial lesion. Further, our results suggest that immediate suture would not be as favourable as suture delayed for a few (say 10-20) days, in order to allow the development of a fairly high Schwann cell activity at the time of suture. In this way there would be less likelihood that the Schwann cell junction will be hindered by the prior development in the suture-line of serious fibrosis. Such fibrosis is a likely consequence of the numerous active fibroblasts which we found in the bulb and nearby connective tissue from 4 days after the nerve was cut. The same argument indicates the use of predegenerated grafts, provided predegeneration is short (10–20 days).

These conclusions are in entire agreement with those of Holmes & Young (1942), who found from experimental suture that long delay after the initial lesion before surgical interference is inimical to good repair, but that immediate suture is not so effective as somewhat delayed suture. Further, Sanders & Young (1942), comparing autografts of fresh nerve with autografts predegenerated 6–9, 14–16 or 25–28 days, found a small delay of the growing axons at the junction of graft and central stump with predegenerated than with fresh grafts; a difference which they regard as suggestive though not statistically significant.

It may prove possible to apply our results with reasonable assurance to the human without experimental analysis by a study of the histological correlations of the activity curve we have obtained in the rabbit.

The expenses of this research were defrayed by a grant from the Rockefeller Foundation to Professor Lancelot Hogben for research work in the Department of Zoology in the University of Birmingham.