- Email: [email protected]
Numbers of regenerating axons in parent and tributary peripheral nerves in the rat
Brain Research, 326 (1985) 27-40
27
Elsevier BRE 10482
Numbers of Regenerating Axons in Parent and Tributary Peripheral Nerves in the Rat C.-B. JENQ and R. E. COGGESHALL
Marine Biomedical Institute, Departments of Anatomy and Physiology & Biophysics, and the Neuroscience Graduate Program, University of Texas Medical Branch, Galveston, TX 77550-2772 (U.S.A.) (Accepted May 1st, 1984)
Key words: peripheral nerve - - regeneration
This study is concerned with numerical parameters of axonal regeneration in peripheral nerves. Our first finding is that the number of axons that regenerate into the distal stump of a somatic nerve at a particular time after transection is partially dependent on the type of lesion used to interrupt the axons. The second question concerns the proportion of axons that regenerate into the distal stump of a parent nerve compared to the proportions that regenerate into tributary nerves that arise from the parent. The proportions of regenerated myelinated axons in the nerve to the medial gastrocnemius muscle and myelinated and unmyelinated axons in the sural nerve are the same as the proportions of myelinated and unmyelinated axons that regenerate into the distal stump of the sciatic nerve for the crush, 0 and 4 mm gap transections. Proportionally fewer axons regenerate into the tributary nerves following the 8 mm gap transection, however. This implies that the length of the gap has an influence on whether or not axons in tributary nerves regenerate in concert with axons in the distal stump of the parent nerve. The unmyelinated fibers in the nerve to the medial gastrocnemius muscle are different because they do not regenerate in proportion to those in the distal stump of the sciatic nerve. We also provide evidence to indicate that myelinated axons branch whereas unmyelinated fibers end blindly when they enter the distal stump after crossing a sciatic nerve transection. Finally the normal arrangement of perineurial cells seems to be disrupted after the sciatic nerve regenerates across a gap. INTRODUCTION
providing the numbers of axons in the distal stump of the sciatic nerve, we also provide the numbers of ax-
The present study continues a series attempting to
ons at the site of transection to see how these num-
establish numerical parameters of regeneration in
bers compared with those in the distal stump. Finally,
mammalian peripheral nerve. In previous work we
in one sciatic nerve which r eg en er at ed across an 8
found that the numbers of axons that regenerate in
mm gap, the axons were counted in the proximal and
tributary nerves are a function of the type of lesion
distal stumps and at intervals throughout the length
used to transect the parent nerve 11,~2. Obvious ques-
of the transection to determine how axon numbers
tions that follow from this finding are: (1) whether
change throughout the length of the regenerated
the axonal numbers that appear in the distal stump of
gap.
the transected parent nerve are also a function of the type of transecting lesion; and (2) w h e t h e r there is
MATERIALS AND METHODS
any obvious relationship b e tw e e n the numbers of axons that regenerate into the distal stump and the
The experimental procedures are similar to those
numbers that regenerate into tributary nerves that
in our previous paper 12. Briefly, adult (200-300 g)
arise from the distal stump. T h e present study is con-
S p r a g u e - D a w l e y male rats were obtained from Tex-
cerned with rat sciatic nerve as the parent nerve and
as Inbred Mouse C o m p a n y , H o u s t o n , Texas. The
the sural nerve and the nerve to the medial gastroc-
rats were anesthetized with 35 mg/kg of pentobarbi-
nemius muscle as the tributary nerves. In addition to
tal (Nembutal), and when anesthesia was deep, one
Correspondence: R. E. Coggeshall, Marine Biomedical Institute, University of Texas Medical Branch, 200 University Boulevard, Galveston, TX 77550-2772 U.S.A. 0006-8993/85/$03.30 © 1985 Elsevier Science Publishers B.V.
2~ of the following procedures was done: {1) sciatic nerve crush: (2) sciatic nerve transection; (3) sciatic transection with a 4 mm length of nerve removed: and (4) sciatic transection with an 8 mm length of nerve removed. For the sciatic crush, the nerve was exposed in the thigh and pinched several times with fine forceps 2 mm distal to the ischial tuberosity. For transection, the nerve was exposed in the thigh and cut 2 mm distal to the ischial tuberosity. For transections with 4 or 8 mm gaps, the nerve was cut as above and then cut again 4 or 8 mm caudally, and the isolated segment of nerve was then removed. The cut ends of the transected nerves were placed in a silicone tube, and a single 9-0 silk suture was passed through the perineurium of the proximal and distal stumps 15.~. The cut ends of the nerve were o p p o s e d after simple transection and separated by 4 or 8 mm respectively when a 4 or an 8 mm segment of nerve was removed. Eight weeks following surgery, the animals were reanesthetized as above and when anesthesia was deep, the thorax was o p e n e d and a p p r o x i m a t e l y 25(I ml of 0.9% NaC1 containing 250 international units of heparin and 0.25 ml of 1% NaNO2 was perfused through the heart. The right auricle was o p e n e d and as soon as the effluent was free of blood, the perfusion fluid was changed to a mixture of 3% glutaraldehyde, 3% formaldehyde (made from p a r a f o r m a l d e hyde) and 0.1% picric acid in 0.1 M, p H 7.4, cacodylate buffer. A f t e r approximately 450 ml were perfused, the u n o p e r a t e d sciatic nerves, the distal stumps of the crushed and transected sciatic nerves and the portions of the regenerated nerves in the tubes were removed and placed in the same fixative overnight. The next day the nerves were rinsed in 3 changes of cacodylate buffer and then placed into a solution of 1% osmium tetroxide and 1.5% potassium ferricyanide 1~ in 0.1 M, pH 7.4, cacodylate buffer for 2 h. The tissue was again rinsed in 3 changes of cacodylate buffer and placed in 0.5% uranyl acetate in 0.1 M, pH 6, maleate buffer for 12-16 h. The tissue was then rinsed in p H 5.2 maleate buffer for 30 rain, d e h y d r a t e d in ascending concentrations of ethanol, and e m b e d d e d in a Pelco 110 e m b e d d i n g mold containing a mixture of epon and araldite. After hardening, thick sections were cut with glass knives and stained with 0.5% toluidine blue in a 1% sodium borate solution. Thin sections were cut and
placed on single-hole, formvar-coated grids. The sections were taken 3 mm distal to the site of crush and 3 mm distal to the proximal end of the distal stump in the transected nerves. The sections of regenerated nerves in the tube were taken at the site of transection for the 0 mm transection and midway between the proximal and distal stumps for the 4 and 8 mm gap transections. One nerve that regenerated across an 8 mm gap was sectioned in the proximal stump, at 2 mm intervals in the regenerated part and also in the distal stump. All thin sections were stained with 0.2% lead citrate and p h o t o g r a p h e d in a Philips 300 or 3()1 electron microscope. Montages were constructed. To lessen the labor of counting, each montage was divided into approximately 24 equal sized rectangles by drawing an a p p r o p r i a t e n u m b e r of horizontal and vertical lines on the montage. All axons in 3 montages were then counted. Then for these nerves, unmyelinated axons were calculated by a previously described formula:: TOT. UN. -
TOT. MY. S A M P . MY.
× S A M P . UN.
where TOT. UN is the estimated total number of unmyelinated fibers in the montage; TOT. MY. is the total number of myelinated fibers in the montage; SAMP. MY. is the n u m b e r of myelinated axons in a certain n u m b e r of rectangles; and S A M P . UN. is the number of unmyelinated fibers in the same rectangles. W h e n every other rectangle was counted, the estimated n u m b e r of unmyelinated axons deviated from the n u m b e r d e t e r m i n e d by complete counts by 1% or less, a variability that was acceptable. If every 3rd rectangle was counted, the variation from the complete counts rose to 5 - 1 0 % , which we flmnd unacceptable. Accordingly the counts in all nerves except the 3 completely counted montages were done by counting all myelinated axons and the unmyelinated axons in every other rectangle and then using the above formula. The statistical test was the Student's t-test, and a probability of P < 0.05 was chosen as the level of significance. RESULTS The n o r m a l sciatic n e r v e
The sciatic nerve of the rat is a typical somatic nerve. In a low power picture of a normal rat sciatic
29
®
ili~
100 p
Fig. 1. A section of a normal rat sciatic nerve 2 mm distal to the ischial tuberosity. The large clear areas in the nerve are blood vessels which are empty because of perfusion. The myelinated axons with their characteristic sheaths are also seen but the other cellular components of the nerve are not clearly visible at this magnification. Surrounding the axons is a line, the perineurium, which at higher magnification would be resolved into several sleeves made of apposed squamous cells. Surrounding the perineurium is the connective tissue of the epineurium. Calibration = 100~m.
®
~..2J
I" :'i ;
"
i)
t'
.1
}
lOOp Fig. 2. A section of a distal stump 8 weeks after simple transection of the sciatic nerve. The blood vessels are large but the cellular components are not remarkably different from normal. Note that the general shape of the nerve is maintained. Calibration = 100,um
31
100~ Fig. 3. A section through the transection site of a sciatic nerve that had a simple transection 8 weeks previously. Note that the nerve is round. There are large numbers of blood vessels in the region of the perineurium and blood vessels are also more numerous than normal in the substance of the nerve. The cellular components of the nerve are not remarkable. Calibration = 100/~m.
nerve (Fig. 1), one can see the flattened oval cross section with one pole being slightly r o u n d e r and
areas, and the myelinated axons can be clearly seen (Fig. 1).
larger than the other. This reflects the fact that this nerve is shaped like a thick, somewhat assymetric ribbon in the adult. The cellular components of the nerve are myelinated and u n m y e l i n a t e d axons, Schwann cells, capillaries consisting of endothelial cells and pericytes, fibroblasts, and occasional mononuclear reticuloendothelial cells. At low magnification only the capillaries, which are the large clear
The cellular components of the nerve are embedded in a collagenous extracellular matrix that is referred to as the e n d o n e u r i u m . Surrounding the cellular components of the nerve and endoneurial connective tissue is the perineurium. This consists of several sleeves, each of which is made of squamous cells, but at low power the perineurium appears as a line surrounding the cellular components of the nerve
32
Fig. 4. A section through the middle of the regenerated gap of a sciatic nerve that was transected with the production of a 4 mm gap 8 weeks previously. Again note the round shape, thc more numerous blood vessels than normal and the smaller size of the nerve compared to that in Fig. 3. Calibration - 100 urn.
(Fig. 1). S u r r o u n d i n g the p e r i n e u r i u m is the e p i n e u -
n u m b e r and size of b l o o d vessels (Fig. 2). W h e n the
rium which is the c o n n e c t i v e tissue that holds the
n e r v e is t r a n s e c t e d and the two ends p l a c e d in a sili-
n e r v e in place in the b o d y (Fig. 1 ).
cone tube, h o w e v e r , t h e r e are significant c h a n g e s in the structure of the n e r v e at the site of t r a n s e c t i o n
The regenerated sciatic nerve
(Figs. 3 - 5 ) . First this part of the n e r v e b e c o m e s al-
t h e n r e g e n e r a t e s , the structure of the distal s t u m p is
most c o m p l e t e l y r o u n d (Figs. 3 - 5 ) , in contrast to the oval shape of the n o r m a l n e r v e or the r e g e n e r a t e d
similar to that of the n o r m a l n e r v e e x c e p t for different n u m b e r s and sizes of axons and for an increase in
distal stump. T h e r o u n d s h a p e is not i m p o s e d on the n e r v e bv the tube, h o w e v e r , for all of these n e r v e s
W h e n the sciatic n e r v e is c r u s h e d or t r a n s e c t e d and
33 float free in the tube and there is a considerable fluid
n u m b e r s and sizes of the axons are different. Also the
filled space between each nerve and the wall of its enclosing tube. Second, there is, as in the distal stump,
connective tissue of the nerve, with the exception of
a considerable increase in n u m b e r and size of the
tioned above and the presence of slightly more mo-
blood vessels (Figs. 3 - 5 ) . Third the perineurial cells
nonuclear reticuloendothelial cells, is not markedly different from normal.
do not form sleeves; especially in the 4 and 8 m m
the different arrangement of perineurial cells men-
gaps. Instead, there are large n u m b e r s of circularly
Differences between the pattern of regeneration
oriented squamous cells that do not seem to contact
when comparing the transections (the 0, 4 and 8 mm
one another. Thus one can trace many labyrinthine
gaps) with one another are primarily in the size of the regenerated nerves and in the n u m b e r s and sizes of
extracellular pathways from the external surface of the nerve to the endoneurial connective tissue. There is an outer lining of cells on the external surface of the
axons within the nerve (see below).
nerve, however, that might provide a diffusion bar-
As a final point, not all sciatic nerves regenerated when an 8 mm gap was made in the sciatic nerve. In
rier between nerve and fluid in the tube. Some features of the nerve are preserved in the
The latter 3 animals were not considered in this
gap. The axons, both myelinated and unmyelinated,
study, but in these animals there would be no regen-
our material, there was no regeneration in 3 cases.
and their supporting cells have the same cytologic
erated axons in the distal stump. The relation of the
characteristics as in the normal, even though the
length of removed nerve to the success of regenera-
Fig. 5. A section through the middle of the regenerated gap of a sciatic nerve that was transected with the production of an 8 mm gap 8 weeks previously. Note the features as before and also that this nerve is smaller than those in Figs. 3 and 4. The cross sectional area of our0 mm gap nerves is 0.76 + 0.33 mm2whereas for the 8 mm gap nerves, it is 0.15 + 0.03 mmz (mean +__S.D.). Calibration = 100~m.
34
Fig. f~. An electron micrograph of a myelinated ax(m and several unmyelinated axons trom a regenerated ncrvc in the m~ddlc ol an '4 mm gap. Note that the axons are cytologically similar to those from normal nerves, as illustrated in many studies, exccpl h:~r multiple unmyelinated axons m a single trough of Schwann cell cytoplasm. Calibration - 2 urn.
tion across the gap will be considered in future studies.
tubules, and their axoplasm is relatively electron h_lcent as compared to the Schwann cell cytoplasm
Myelinated and unmyelinated axons
(Fig. 6). The major difference from normal, and this does not interfere with accurate counting, is that mul-
The neural components of both normal and regenerated nerves are the myelinated and unmyelinated fibers. These axons have been described in many locations (e.g. ref. 18), and by us for the particular nerves under consideration here 11-13. The criteria for recognizing these fibers so they can be accurately counted are presented in the above papers. Thus for this paper the only point to make is that regenerated myelinated and unmyelinated axons are recognized by the same criteria as in the normal. In Fig. 6, one myelinated and a group of unmyelinated axons are shown. The myelin sheath is characteristic and easily recognizable. The unmyelinated axons occupy troughs of Schwann cell cytoplasm, they contain the characteristic axonal arrangement of filaments and
tiple unmyelinated axons often occupy a single trough of cytoplasm in the regenerate (Fig. (~) whereas this is relatively rare in the normal.
Axonal counts A major purpose of the present paper is to present axonal counts from both the distal stump and the region of the transection 8 weeks following proximal sciatic nerve crush, simple transection (0 mm gap), transection with a 4 m m gap (4 mm gap) and transection with an 8 mm gap (8 mm gap). For the last 3 procedures, the two ends of the nerve are placed in a tube and anchored in place by a stitch. The counts are presented in Table I, and means of these counts are shown graphically in Fig. 7. Normal sciatic nerves
35 TABLE I This table presents myelinated and unmyelinated axon counts from the sciatic nerve. The areas counted are in the tube for transections, in the distal stumps, and 2 mm distal to the ischial tuberosity of the unoperated side (Unop.). Lesion
Crush
Animal number
Axon numbers Myelinated Distal stump
Unop.
-
11007 10045 8842 10615 9752 10052 ± 834
8387 7378 8225 8001 8132 8025 ± 388
7 8 9 25
11805 13444 10807 17426 13371 ± 2914
12016 15986 10513 14589 13276± 2469
7886 8694 7948 7899 8107 ± 392
12 13 14 15 26
10355 2628 10260 10254 10892 8878 ± 3504
14243 10068 14258 15578 13781 13586± 2078
16 18 19 20 27
6334 2994 4499 4054 5328 4642 ± 1266
9644 4103 8398 5377 7708 7046 ± 2262
1 2 3 4 5
Mean ± S.D. 0 mm
Mean ± S.D. 4mm
Mean ± S.D. 8 mm
Mean ± S.D.
Unmyelinated
In tube
Distal stump
Unop.
12959 14566 9493 12979 13654 12730 ± 1925
17547 12660 17503 18444 14783 16187± 2402
17889 20063 17230 25288 20118 ± 3653
8905 9834 6941 13329 9752 ± 2672
13358 18066 15384 14922 15433± 1958
8283 8659 8966 8621 7451 8396± 581
14373 7436 15386 14427 18922 14109 ± 4168
11496 9304 7063 11258 13112 10447± 2325
14843 16805 15243 13786 13817 14899± 1241
8289 8106 7847 7379
13364 10387 10246 10025 8210 10446 ± 1853
12758 4038 6368 8544 6482 7638 ± 3277
17610 15964 14885 14128
7905 ± 395
In tube
15647± 1510
from the u n o p e r a t e d side at these levels contain an
parent nerve. Normal and r e g e n e r a t e d axon num-
average of 8210 myelinated axons and 15,542 unmye-
bers for tributary nerves are given in our previous pa-
linated axons. It is useful to normalize the axon counts from
per m. To obtain our comparisons, we divide the num-
these nerves. These ratios are diagrammed in Fig. 8.
side by the number in the parent sciatic nerve on the
ber of axons from each tributary on the u n o p e r a t e d
The dotted line represents mean axon numbers from
unoperated side of the same animal and co m p ar e this
normal nerves and the bars represent operated nerve
fraction to the n u m b e r of axons in the tributary nerve
numbers expressed in reference to these figures.
on the operated side divided by the n u m b e r of axons
Note that there are m o r e myelinated axons than nor-
in the distal stump of the sciatic nerve again in the
mal in the distal stump following crush, 0 and 4 mm
same animal. This gives a ratio, and these ratios are
gap transections but slightly fewer following the 8
presented in graphical form in Fig. 9. For the mye-
mm gap transection. Also note that there are fewer unmyelinated axons than normal no matter what sur-
linated fibers in the nerve to the medial gastrocnemi-
gical paradigm is used to interrupt the axons in the
bers in the sural nerve, note that the proportions that
parent nerve.
regenerate into the tributary nerves are the same as
us muscle and the myelinated and unmyelinated fi-
O n e of the m a j o r goals of the present paper is to
the proportions that regenerate into the distal stump
determine wh et h er the numbers of axons that regen-
of the parent nerve for the crush, 0 and 4 m m gap
erate into tributary nerves are proportional to the
transections. For the 8 m m gap transection, propor-
numbers that regenerate into the distal stump of the
tionally fewer axons regenerate into the tributary
36
(!)
following crush ol a ,omatic nerve 5.~'-w-:'. ~tnd (,,tH data arc confirmatory. M:10052+- 834 U:12730_+ 1925
By contrast tO the myelinated fibers, we find thai
,..:..-..2.....................-
........... ..-.+....
unmyelinated fibers decrease in the distal stump following regeneration after sciatic nerve crush in the rat. The reason for the different response of unmye-
M~13371_+2914 M:13276 + 2469 U:20118+31183 U:9752 +_2672
linated axons is not known, but it is clear thai neither
SC,A,,C
o
iiiii ~."2325:8g?13-+3504 1410~g_4118 i :10447+ :13586+2078
myelinated nor unmyelinated axon numbers return to normal following sciatic nerve crush in our system. Nevertheless we provide data that is in partial agreement with G u t m a n n and Sanders: in that the
SCIATIC 4mm ===========i=i=.=i=.== i =.={==. i .::::::::::::::::::::::::::::::::: i:iiii~ ~z4642+1266 : 1044~_+18~3
M:7046+ 2262 U:7638+_3277
SCIATIC8mm ~iiiiiiiii!iiiiiii!iii!i!iii!iiiiiiiiiii~
numbers of axons are less deviant from normal following regeneration distal to a crush than following any of our transections.
®
Fig. 7. A schematic diagram of our axonal counts in this study. For each procedure, the proximal stump is indicated by horizontal lines, the distal stump by large dots and the surgical gaps by small dots. The myelinated axon numbers _+ the standard deviations are indicated by M and the unmyelinated axon numbers + the standard deviations by I"
MYELINATEO UNMYELiNATEO
nerves than into the distal stump of the parent nerve. By contrast, proportionally more unmyelinated axons regenerate into the nerve to the medial gastrocnemius muscle than appear in the distal stump of the parent nerve no matter what surgical procedure is
1.2
used. A final goal in the present paper is to present axon counts at various levels from one sciatic nerve that regenerated through an 8 mm gap. These data are pre-
i
sented in schematic form in Fig. 10. Note that the greatest n u m b e r of axons are found in the proximal stump, and that the myelinated and unmyelinated axon counts are proportionally different in the tube and in the distal stump.
N
DISCUSSION Recovery from nerve crush is reported to be better than recovery following any type of transection7.m. G u t m a n n and Sanders 7 hypothesized that functional return is maximal when axonal numbers and sizes return to normal distally, and these investigators reported that myelinated axon n u m b e r s did return to normal after crush but not after transection;. Others found, however, that the n u m b e r of myelinated axons was increased in the distal stump at various times
Sciatic Crush
Sciatic omrn
Sciatic 4ram
ciatic
8ram
Fig. 8. A bar diagram of the normalized axon numbers in the distal stumps in this study. The dotted line indicates the mean numbers of axons from the unoperated side and the bars indicate the deviations from these numbers following the surgical procedures used in this study. Note that the myelinated axons are more numerous than normal after the crush, 0 and 4 mm gap transections. These differences are significant (P < 0:02). Note that there are fewer myelinated axons than normal aftcr the 8 mm gap transections. This difference is not significant. The unmyelinated axons are less numerous than normal after all surgical procedures and these differences are significant (l' < 0.02).
37
(!) ~NMG,My NMG,Un F~SN,My ~SN,Un
3,¢
22
~. ,.; 2. ¢o co
1.¢
O.E
...... Sciatic Crush
i
i.--i .
Sciatic Ornm
Sciatic 4mm
.
.
.
Sciatic 8ram
Fig. 9. A diagram indicating the proportion of axons that regenerate in tributary nerves as compared to normal tributary axon numbers in relation to the numbers that regenerate into the distal stump of the transected parent nerve as compared to normal axon numbers in the parent. To obtain these ratios, the numbers of axons in a tributary nerve (Oper. Trib.) are divided by the numbers in the distal stump of the parent sciatic nerve (Oper. Pare.) and this fraction is divided by the numbers from the tributary nerve on the unoperated side of the animal (Unop. Trib.) divided by the number of axons in the sciatic nerve on the unoperated side (Unop. Pare.). The dotted line at 1.0 indicates the ratios that obtain when the same proportion of axons regenerate in the tributary and parent nerves. Ratios less than 1.0 indicate that proportionally fewer axons regenerate in the tributary nerves and greater than 1.0 that proportionally more axons regenerate in the tributary nerves. Note that proportionally the same number of myelinated axons in the nerve to the medial gastrocnemius muscle (NMG, My), and the sural nerve (SN, My) and unmyelinated axons in the sural nerve (SN, Un), regenerate as regenerate into the distal stump of the parent sciatic nerve for the crush, 0 and 4 mm gap transections, but that proportionally fewer of these fibers regenerate in the tributary nerves for the 8 mm gap transection. By contrast proportionally more unmyelinated axons in the nerve to the medial gastrocnemius muscle (NMG, Un) regenerate no matter what surgical procedure is used on the sciatic nerve.
For simple transection, it has been reported that more myelinated axons regenerate into the distal stumpS,10.20, fewer myelinated axons regenerate into
the distal stumpl,3,4,7, 9, o r that more, the same, or fewer regenerate depending on conditions8,17. A partial explanation for this difference may be our finding that the number of axons that regenerate is related to the amount of nerve removed when the gap is made. In this regard, there are the clinical observations that the longer the length of the nerve removed, the more trouble the patient is likely to have 21, and that the longer the gap, the less likely axons are to cross 15. The only previous work, to our knowledge, where axonal numbers are related to the length of nerve removed shows that the number of axons that regenerate into tributary nerves at a particular time is partially dependent on the length of parent nerve that is removed 12. The relation is not a simple linear correlation, however. Two complicating factors are: (1) that myelinated and unmyelinated axons respond differently; and (2) that the numbers of unmyelinated axons that regenerate are different in a cutaneous as opposed to a motor tributary nerveH.1L Thus the next question, which is considered in the present paper, is whether the numbers of regenerating axons in the distal stump of the parent nerve are related to the numbers of regenerating axons in the tributary nerves? To normalize axon counts so that comparisons can be made between the distal stump of parent on the one hand and tributary nerves on the other, it is necessary to determine the proportions of myelinated and unmyelinated axons that regenerate in the different situations and then to compare these proportions with one another. The axonal numbers in tributary nerves have been reported previouslyt2, and the numbers in the distal stump are given in this paper.
® I~10395 U:16396
M:4499 U:10246
~.'9366 835 U=9449
I:~ox.
~8368 D I ST.
Fig. 10. A schematic diagram showing myelinated (M) and unmyelinated (U) axon counts from different levels of a sciatic nerve whose axons regenerated across an 8 mm gap. As before, the proximal stump (PROX) is indicated by horizontal lines, the distal stump (DIST) by large dots and the region of the gap by small dots. The two stumps and central gap region are contained in a silicone tube.
3~ I'he proportions are plotted in Fig. 9. Because they behave similarly, the myelinated fibers in the nerve to the medial gastrocnemius muscle and the myelinated and unmyelinated fibers in the sural nerve will be placed in one group. For these populations, the percentage of fibers, as c o m p a r e d to normal, that regenerate into the tributary nerves is the same as the percentage that regenerates into the distal stump of the parent nerve following crush and 0 and 4 mm gap transections. Thus in these cases, the normal tithe of axons that enters a tributary is maintained after regeneration, even though the numbers of regenerating axons in the distal stump of the parent and in the tributaries change considerably. A n o t h e r way of saying this is that the numbers of axons in tributary nerves change in concert with the numbers in the distal stump of the parent nerve for these particular surgical paradigms. For the 8 mm transection, however, the above conclusion does not hold. In this case proportionally fewer axons enter the tributary nerve than are found in the distal stump of the parent nerve. One explanation for this is that not enough time has been allowed for all axons to regenerate and thus if the samples were taken at a longer interval following surgery, there would not have been a discrepancy. This possibility will be checked. If this is not the explanation, it may be that when the length of the gap increases beyond a certain point, it becomes difficult for the organism to regulate the numbers of axons that regenerate into the tributary nerves. This would presumably imply either that some axons end blindly in the distal stump and so do not enter any tributary or if all axons in the distal stump enter a tributary, that some tributaries receive proportionally fewer axons and others receive more. We can determine experimentally which of these suggestions is correct. It will be important to do this for if there is a critical gap length beyond which the proportion of axons in parent and tributary nerves no longer are correlated, as seems to be the case, an understanding of this p h e n o m e n o n could have clinical impact. The unmyelinated fibers in the nerve to the medial gastrocnemius muscle clearly behave differently than the other fiber populations in this study. The data indicate that the r e g e n e r a t e d unmyelinated axons in the nerve to the medial gastrocnemius muscle are proportionally 2 - 3 times as numerous when compared to normal as the unmyelinated axon population
in the distal stump of the sciatic nerve no m~tter what type of surgery or how long the gap to interrupt the sciatic nerve. We have no explanation fo~ this but feel it may be important. Further data of interest wilt be to determine whether these fibers are sensory or postganglionic or both, and whether this is a characteristic of unmyelinated fibers in all nerves to muscles or only in the nerve to the medial gastocnemius muscle. A t present, the m a j o r value of the axonal counts in the tube is to allow a comparison with the numbers in the distal stump so as to get some indication of branching. The fact that axons branch at and slightly above the site of transection has been known since neuromas were investigated histologicallyl'L Furthermore the amount of branching presumably has a bearing on the.type of functional recovery m. The early histological work did not quantitate branching, but a combined anatomical and physiological method for quantitation of branching has recently been devised m. These investigators found that after transection without reapposition, each myelinated axon that regenerated through a transection site branched an average of 1.7 times. In this study, we find that after simple transection there are more axons than normal both in the tube and in the distal stump, but that the numbers of axons in these two locations are approximately equal. The simplest way to interpret these data would be to assume that there was considerable branching of myelinated axons proximal to the tube but that each axon in the tube passed into the distal stump without branching. A future task, therefore, is to determine the influence of various restorative procedures following transection on the amount of branching at or above the site of transection. Then the hypothesis that the procedure that results in the least amount of branching is associated with the best behavioral recovery will be tested. Since there are a greater number of myelinated axons in the distal stump than in the tube in the 4 and 8 mm gap transections, it seems necessary to postulate axonal branching in the tube m these conditions. If we divide the myelinated axon numbers in the distal stump by the number in the tube we get an average of 1.6. This is close to Horch and Lisney's branching factor of 1.7 m. We must repeat our work physiologically but if our anatomical data are accurate, it suggests that the length of the gap between the proximal
39 and distal stumps may be a factor in determining the amount of branching that the regenerating myelinated axons undergo at the site of transection. The unmyelinated axons regenerate differently than the myelinated axons in our paradigms. In our material, it is striking that the number of unmyelinated fibers in the tube is always much greater than in the distal stump. One interpretation of these data is that there is no branching of unmyelinated axons and that the excess axons end blindly in the tube. This is not definitive, of course, because it is possible that more axons than the 'excess' end blindly and that the remainder then branch. Physiological experiments will be necessary to choose between the possibilities. It is instructive in this case to examine the multilevel axon counts shown in Fig. 10. Here it can be seen that there are approximately 16,000 unmyelinated axons at the proximal end of the transection, which is essentially the normal number. In the tube, however, there are approximately 10,000 axons and in the distal stump there are approximately 6500 axons. Since the number at the proximal end is close to that obtained from the normal nerve, there is little evidence for branching at this site. The decrease in number in the tube then presumably indicates that 40% of the fibers in the proximal stump end when the axons leave the proximal stump and enter the tube. There is another significant loss of axons when the axons leave the tube and enter the distal stump which also suggests a site where a significant number of unmyelinated axons end blindly. This presumably indicates that the departure from the proximal stump and the entrance to the distal stump are two particularly important sites in that they both seem to be places where 'barriers' exist that prevent a siginificant number of unmyelinated axons from passing distally. The next point of this paper concerns the myelinated axon counts at different levels in the gap transection (Fig. 10). As for the unmyelinated axons,
REFERENCES 1 Cabaud, H. E., Rodkey, W. G., McCarroll, H. R., Jr., Mutz, S. C. and Niebauer, J. J., Epineurial and perineurial facicular nerve repairs: a critical comparison, J. Hand Surg., 1 (1976) 131-137. 2 Coggeshall, R. E., Coulter, J. D. and Willis, W. D., Jr., Unmyelinated axons in the ventral roots of the cat lumbosacral enlargement, J. comp. Neurol., 153 (1974) 39-58. 3 Davenport, H. A., Chor, H. and Cleveland, D. A., Fiber
there is a large reduction in axon numbers at the point where the axons leave the proximal stump and enter the tube. Then the numbers stay essentially constant down the length of the tube, although there is a low point distally. As the axons enter the distal stump, however, there is a sudden increase in axonal numbers. The easiest way to interpret these data is that many myelinated axons end blindly when they leave the proximal stump but then the remaining fibers branch when they enter the distal stump. Again the data need to be confirmed, but if they are, it is interesting that at the point of entry to the distal stump myelinated axons branch and unmyelinated axons drop out. As a final point, one of the noteworthy features of regeneration after transection is that the perineurium around the regenerated axons in the gap does not return to its normal configuration. There is an extensive literature on the morphology and barrier properties of perineurium in normal and regenerating nerves 22, but the perineurium where a nerve regenerates across a gap within a tube deserves more attention. One of the major difficulties in regeneration seems to be that transecting the perineurium exposes regenerating axons to the presumably alien environment of the extracellular fluid (e.g. ref. 10). It may be that placing two stumps of a transected nerve in a silicon tube makes up for any perineurium deficiencies. Further work on the morphology and barrier properties of the perineurium, and ionic content of the endoneurial fluid in regenerating nerves is thus in order. ACKNOWLEDGEMENTS This work is supported by N I H Grants NS 07377, NS 10161, NS 17039, NS 11255, and a support grant from the Moody Foundation.
ratios in regenerating nerves, II. The status of regrowth of the sciatic and nerves to the gastrocnemius in Macacus 3 months after section, J. comp. Neurol., 79 (1939) 153-159. 4 Davenport, H. A,, Chor, H. and Dolkart, R. E., The ratio of myelinated to unmyelinated fibers in regenerated sciatic nerves on Macacus rhesus, J. comp. Neurol., 67 (1937) 481-491. 5 Evans, D. H. L. and Murray, J. C., A study of regeneration in a motor nerve with a unimodal fiber diameter distribution, Anat. Rec., 126 (1956) 311-329.
4(t
7
8
9
10
11
12 13
14
Greenman, M. J., Studies on the peroneal nerve of the albino rat: number and sectional area of fibers: area relation ol axis to sheath, J. comp. Neurol., 23 (1913)479-513. Gutmann, E. and Sanders, F. K., Recovery of fiber num-. hers and diameters in the regeneration of peripheral nerves, J. Physiol. (Lond.), 101 (1943) 489-518. Guyon. L., Resultats anatomiques et fonctionnels observes au cours de la cicatrisation des nerfs chez la chien, Rev. neurol., 37 (1921) 937-949. Hausamen, J. E., Samii, M. and Schmidseder, R., Restoring sensation to the cut inferior alveolar nerve by direct anastomosis or by free autologous nerve grafting. Experimental study in rabbit. Hast. reconstr. Surg., 54 (1974) 83-87. Horch, K. W. and Lisney, S. J. W., On the number and nature of regenerating myelinated axons after lesions of cutaneous nerves in the cat, J. Physiol. (Lond.l, 313 (1981) 275-286. Jenq, C.-B. and Coggeshall, R. E., Effects of sciatic nerve regeneration on axonal populations in tributarv nerves, Brain Research, 295 (1984) 91-100. Jenq, C.-B. and Coggeshall, R. E., Regeneration of axons in tributary nerves, Brain Research, in press. Jenq, C.-B., Hulsebosch, C. E., Coggeshall, R. E. and Perez-Polo, R., The effect of nerve growth factor and its antibodies on axonal numbers in the medial gastrocnemius nerve of the rat, Brain Research, 299 (1984) 9-14. Langford, L. A. and Coggeshall, R. E., The use of potassium ferricyanide in neural fixation, Anat. Rec., 197 (1980)
297-303. 15 Lundberg, G., Dahlm, L. B., Danielsen, N., (iclberman. R. H.. Longo, F. M., Puwell, H ( . and Varon, S., Nerve regeneration in silicone chambers: influence o1 gap length and of distal stump components. Evp. Neurol.. 76 (1982) 361-375. 16 Lundberg, G., Gelberman, R. tI., Longo, R. M., Powell, H. C. and Varon, S., In vivo regeneration of cut nerve encased in silicone tubes, J. Neuropath. exp. Neurol., 41 (1982) 412-422. 17 Orgel, M. G. and Terzis, J. D., Epineuriat vs. perineurial repair: an ultrastructural and electrophysiological study of nerve regeneration, Plast. reconstr. Surg., 60 (1977)80-91. 18 Peters, A., Palay, S. L. and Webster, H., The Fine Struc-
ture of The Nervous System: The Neurons and Supporting Cells, W. B. Saunders Co., PhiLadelphia, 1976, pp. 1-4(16. 19 Ramon y Cajal, S., Degeneration and Regeneration of the Nervous System, Hafner Publishing, New York, 1959, 769 PP. 20 Shawe~ G. D. H., On the number o1 branches formed by 1egenerating nerve fibers, Brit. J. Surg., 42 (1955) 474-488. 21 Sunderland, Sir S., Nerves and Nerve Injuries, 2nd edn., Churchill Livingstone, Edinburgh, 1978, 1046 pp. 22 Thomas, P. K. and Olsson, Y., Microscopic anatomy and function of the connective tissue components of peripheral nerve. In P. J. Dyck, P. K. Thomas and E. H. Lambert (Eds.), Peripheral Neuropathy, W. B. Saunders Co., Philadelphia, 1975, pp. 168-189.
27
Elsevier BRE 10482
Numbers of Regenerating Axons in Parent and Tributary Peripheral Nerves in the Rat C.-B. JENQ and R. E. COGGESHALL
Marine Biomedical Institute, Departments of Anatomy and Physiology & Biophysics, and the Neuroscience Graduate Program, University of Texas Medical Branch, Galveston, TX 77550-2772 (U.S.A.) (Accepted May 1st, 1984)
Key words: peripheral nerve - - regeneration
This study is concerned with numerical parameters of axonal regeneration in peripheral nerves. Our first finding is that the number of axons that regenerate into the distal stump of a somatic nerve at a particular time after transection is partially dependent on the type of lesion used to interrupt the axons. The second question concerns the proportion of axons that regenerate into the distal stump of a parent nerve compared to the proportions that regenerate into tributary nerves that arise from the parent. The proportions of regenerated myelinated axons in the nerve to the medial gastrocnemius muscle and myelinated and unmyelinated axons in the sural nerve are the same as the proportions of myelinated and unmyelinated axons that regenerate into the distal stump of the sciatic nerve for the crush, 0 and 4 mm gap transections. Proportionally fewer axons regenerate into the tributary nerves following the 8 mm gap transection, however. This implies that the length of the gap has an influence on whether or not axons in tributary nerves regenerate in concert with axons in the distal stump of the parent nerve. The unmyelinated fibers in the nerve to the medial gastrocnemius muscle are different because they do not regenerate in proportion to those in the distal stump of the sciatic nerve. We also provide evidence to indicate that myelinated axons branch whereas unmyelinated fibers end blindly when they enter the distal stump after crossing a sciatic nerve transection. Finally the normal arrangement of perineurial cells seems to be disrupted after the sciatic nerve regenerates across a gap. INTRODUCTION
providing the numbers of axons in the distal stump of the sciatic nerve, we also provide the numbers of ax-
The present study continues a series attempting to
ons at the site of transection to see how these num-
establish numerical parameters of regeneration in
bers compared with those in the distal stump. Finally,
mammalian peripheral nerve. In previous work we
in one sciatic nerve which r eg en er at ed across an 8
found that the numbers of axons that regenerate in
mm gap, the axons were counted in the proximal and
tributary nerves are a function of the type of lesion
distal stumps and at intervals throughout the length
used to transect the parent nerve 11,~2. Obvious ques-
of the transection to determine how axon numbers
tions that follow from this finding are: (1) whether
change throughout the length of the regenerated
the axonal numbers that appear in the distal stump of
gap.
the transected parent nerve are also a function of the type of transecting lesion; and (2) w h e t h e r there is
MATERIALS AND METHODS
any obvious relationship b e tw e e n the numbers of axons that regenerate into the distal stump and the
The experimental procedures are similar to those
numbers that regenerate into tributary nerves that
in our previous paper 12. Briefly, adult (200-300 g)
arise from the distal stump. T h e present study is con-
S p r a g u e - D a w l e y male rats were obtained from Tex-
cerned with rat sciatic nerve as the parent nerve and
as Inbred Mouse C o m p a n y , H o u s t o n , Texas. The
the sural nerve and the nerve to the medial gastroc-
rats were anesthetized with 35 mg/kg of pentobarbi-
nemius muscle as the tributary nerves. In addition to
tal (Nembutal), and when anesthesia was deep, one
Correspondence: R. E. Coggeshall, Marine Biomedical Institute, University of Texas Medical Branch, 200 University Boulevard, Galveston, TX 77550-2772 U.S.A. 0006-8993/85/$03.30 © 1985 Elsevier Science Publishers B.V.
2~ of the following procedures was done: {1) sciatic nerve crush: (2) sciatic nerve transection; (3) sciatic transection with a 4 mm length of nerve removed: and (4) sciatic transection with an 8 mm length of nerve removed. For the sciatic crush, the nerve was exposed in the thigh and pinched several times with fine forceps 2 mm distal to the ischial tuberosity. For transection, the nerve was exposed in the thigh and cut 2 mm distal to the ischial tuberosity. For transections with 4 or 8 mm gaps, the nerve was cut as above and then cut again 4 or 8 mm caudally, and the isolated segment of nerve was then removed. The cut ends of the transected nerves were placed in a silicone tube, and a single 9-0 silk suture was passed through the perineurium of the proximal and distal stumps 15.~. The cut ends of the nerve were o p p o s e d after simple transection and separated by 4 or 8 mm respectively when a 4 or an 8 mm segment of nerve was removed. Eight weeks following surgery, the animals were reanesthetized as above and when anesthesia was deep, the thorax was o p e n e d and a p p r o x i m a t e l y 25(I ml of 0.9% NaC1 containing 250 international units of heparin and 0.25 ml of 1% NaNO2 was perfused through the heart. The right auricle was o p e n e d and as soon as the effluent was free of blood, the perfusion fluid was changed to a mixture of 3% glutaraldehyde, 3% formaldehyde (made from p a r a f o r m a l d e hyde) and 0.1% picric acid in 0.1 M, p H 7.4, cacodylate buffer. A f t e r approximately 450 ml were perfused, the u n o p e r a t e d sciatic nerves, the distal stumps of the crushed and transected sciatic nerves and the portions of the regenerated nerves in the tubes were removed and placed in the same fixative overnight. The next day the nerves were rinsed in 3 changes of cacodylate buffer and then placed into a solution of 1% osmium tetroxide and 1.5% potassium ferricyanide 1~ in 0.1 M, pH 7.4, cacodylate buffer for 2 h. The tissue was again rinsed in 3 changes of cacodylate buffer and placed in 0.5% uranyl acetate in 0.1 M, pH 6, maleate buffer for 12-16 h. The tissue was then rinsed in p H 5.2 maleate buffer for 30 rain, d e h y d r a t e d in ascending concentrations of ethanol, and e m b e d d e d in a Pelco 110 e m b e d d i n g mold containing a mixture of epon and araldite. After hardening, thick sections were cut with glass knives and stained with 0.5% toluidine blue in a 1% sodium borate solution. Thin sections were cut and
placed on single-hole, formvar-coated grids. The sections were taken 3 mm distal to the site of crush and 3 mm distal to the proximal end of the distal stump in the transected nerves. The sections of regenerated nerves in the tube were taken at the site of transection for the 0 mm transection and midway between the proximal and distal stumps for the 4 and 8 mm gap transections. One nerve that regenerated across an 8 mm gap was sectioned in the proximal stump, at 2 mm intervals in the regenerated part and also in the distal stump. All thin sections were stained with 0.2% lead citrate and p h o t o g r a p h e d in a Philips 300 or 3()1 electron microscope. Montages were constructed. To lessen the labor of counting, each montage was divided into approximately 24 equal sized rectangles by drawing an a p p r o p r i a t e n u m b e r of horizontal and vertical lines on the montage. All axons in 3 montages were then counted. Then for these nerves, unmyelinated axons were calculated by a previously described formula:: TOT. UN. -
TOT. MY. S A M P . MY.
× S A M P . UN.
where TOT. UN is the estimated total number of unmyelinated fibers in the montage; TOT. MY. is the total number of myelinated fibers in the montage; SAMP. MY. is the n u m b e r of myelinated axons in a certain n u m b e r of rectangles; and S A M P . UN. is the number of unmyelinated fibers in the same rectangles. W h e n every other rectangle was counted, the estimated n u m b e r of unmyelinated axons deviated from the n u m b e r d e t e r m i n e d by complete counts by 1% or less, a variability that was acceptable. If every 3rd rectangle was counted, the variation from the complete counts rose to 5 - 1 0 % , which we flmnd unacceptable. Accordingly the counts in all nerves except the 3 completely counted montages were done by counting all myelinated axons and the unmyelinated axons in every other rectangle and then using the above formula. The statistical test was the Student's t-test, and a probability of P < 0.05 was chosen as the level of significance. RESULTS The n o r m a l sciatic n e r v e
The sciatic nerve of the rat is a typical somatic nerve. In a low power picture of a normal rat sciatic
29
®
ili~
100 p
Fig. 1. A section of a normal rat sciatic nerve 2 mm distal to the ischial tuberosity. The large clear areas in the nerve are blood vessels which are empty because of perfusion. The myelinated axons with their characteristic sheaths are also seen but the other cellular components of the nerve are not clearly visible at this magnification. Surrounding the axons is a line, the perineurium, which at higher magnification would be resolved into several sleeves made of apposed squamous cells. Surrounding the perineurium is the connective tissue of the epineurium. Calibration = 100~m.
®
~..2J
I" :'i ;
"
i)
t'
.1
}
lOOp Fig. 2. A section of a distal stump 8 weeks after simple transection of the sciatic nerve. The blood vessels are large but the cellular components are not remarkably different from normal. Note that the general shape of the nerve is maintained. Calibration = 100,um
31
100~ Fig. 3. A section through the transection site of a sciatic nerve that had a simple transection 8 weeks previously. Note that the nerve is round. There are large numbers of blood vessels in the region of the perineurium and blood vessels are also more numerous than normal in the substance of the nerve. The cellular components of the nerve are not remarkable. Calibration = 100/~m.
nerve (Fig. 1), one can see the flattened oval cross section with one pole being slightly r o u n d e r and
areas, and the myelinated axons can be clearly seen (Fig. 1).
larger than the other. This reflects the fact that this nerve is shaped like a thick, somewhat assymetric ribbon in the adult. The cellular components of the nerve are myelinated and u n m y e l i n a t e d axons, Schwann cells, capillaries consisting of endothelial cells and pericytes, fibroblasts, and occasional mononuclear reticuloendothelial cells. At low magnification only the capillaries, which are the large clear
The cellular components of the nerve are embedded in a collagenous extracellular matrix that is referred to as the e n d o n e u r i u m . Surrounding the cellular components of the nerve and endoneurial connective tissue is the perineurium. This consists of several sleeves, each of which is made of squamous cells, but at low power the perineurium appears as a line surrounding the cellular components of the nerve
32
Fig. 4. A section through the middle of the regenerated gap of a sciatic nerve that was transected with the production of a 4 mm gap 8 weeks previously. Again note the round shape, thc more numerous blood vessels than normal and the smaller size of the nerve compared to that in Fig. 3. Calibration - 100 urn.
(Fig. 1). S u r r o u n d i n g the p e r i n e u r i u m is the e p i n e u -
n u m b e r and size of b l o o d vessels (Fig. 2). W h e n the
rium which is the c o n n e c t i v e tissue that holds the
n e r v e is t r a n s e c t e d and the two ends p l a c e d in a sili-
n e r v e in place in the b o d y (Fig. 1 ).
cone tube, h o w e v e r , t h e r e are significant c h a n g e s in the structure of the n e r v e at the site of t r a n s e c t i o n
The regenerated sciatic nerve
(Figs. 3 - 5 ) . First this part of the n e r v e b e c o m e s al-
t h e n r e g e n e r a t e s , the structure of the distal s t u m p is
most c o m p l e t e l y r o u n d (Figs. 3 - 5 ) , in contrast to the oval shape of the n o r m a l n e r v e or the r e g e n e r a t e d
similar to that of the n o r m a l n e r v e e x c e p t for different n u m b e r s and sizes of axons and for an increase in
distal stump. T h e r o u n d s h a p e is not i m p o s e d on the n e r v e bv the tube, h o w e v e r , for all of these n e r v e s
W h e n the sciatic n e r v e is c r u s h e d or t r a n s e c t e d and
33 float free in the tube and there is a considerable fluid
n u m b e r s and sizes of the axons are different. Also the
filled space between each nerve and the wall of its enclosing tube. Second, there is, as in the distal stump,
connective tissue of the nerve, with the exception of
a considerable increase in n u m b e r and size of the
tioned above and the presence of slightly more mo-
blood vessels (Figs. 3 - 5 ) . Third the perineurial cells
nonuclear reticuloendothelial cells, is not markedly different from normal.
do not form sleeves; especially in the 4 and 8 m m
the different arrangement of perineurial cells men-
gaps. Instead, there are large n u m b e r s of circularly
Differences between the pattern of regeneration
oriented squamous cells that do not seem to contact
when comparing the transections (the 0, 4 and 8 mm
one another. Thus one can trace many labyrinthine
gaps) with one another are primarily in the size of the regenerated nerves and in the n u m b e r s and sizes of
extracellular pathways from the external surface of the nerve to the endoneurial connective tissue. There is an outer lining of cells on the external surface of the
axons within the nerve (see below).
nerve, however, that might provide a diffusion bar-
As a final point, not all sciatic nerves regenerated when an 8 mm gap was made in the sciatic nerve. In
rier between nerve and fluid in the tube. Some features of the nerve are preserved in the
The latter 3 animals were not considered in this
gap. The axons, both myelinated and unmyelinated,
study, but in these animals there would be no regen-
our material, there was no regeneration in 3 cases.
and their supporting cells have the same cytologic
erated axons in the distal stump. The relation of the
characteristics as in the normal, even though the
length of removed nerve to the success of regenera-
Fig. 5. A section through the middle of the regenerated gap of a sciatic nerve that was transected with the production of an 8 mm gap 8 weeks previously. Note the features as before and also that this nerve is smaller than those in Figs. 3 and 4. The cross sectional area of our0 mm gap nerves is 0.76 + 0.33 mm2whereas for the 8 mm gap nerves, it is 0.15 + 0.03 mmz (mean +__S.D.). Calibration = 100~m.
34
Fig. f~. An electron micrograph of a myelinated ax(m and several unmyelinated axons trom a regenerated ncrvc in the m~ddlc ol an '4 mm gap. Note that the axons are cytologically similar to those from normal nerves, as illustrated in many studies, exccpl h:~r multiple unmyelinated axons m a single trough of Schwann cell cytoplasm. Calibration - 2 urn.
tion across the gap will be considered in future studies.
tubules, and their axoplasm is relatively electron h_lcent as compared to the Schwann cell cytoplasm
Myelinated and unmyelinated axons
(Fig. 6). The major difference from normal, and this does not interfere with accurate counting, is that mul-
The neural components of both normal and regenerated nerves are the myelinated and unmyelinated fibers. These axons have been described in many locations (e.g. ref. 18), and by us for the particular nerves under consideration here 11-13. The criteria for recognizing these fibers so they can be accurately counted are presented in the above papers. Thus for this paper the only point to make is that regenerated myelinated and unmyelinated axons are recognized by the same criteria as in the normal. In Fig. 6, one myelinated and a group of unmyelinated axons are shown. The myelin sheath is characteristic and easily recognizable. The unmyelinated axons occupy troughs of Schwann cell cytoplasm, they contain the characteristic axonal arrangement of filaments and
tiple unmyelinated axons often occupy a single trough of cytoplasm in the regenerate (Fig. (~) whereas this is relatively rare in the normal.
Axonal counts A major purpose of the present paper is to present axonal counts from both the distal stump and the region of the transection 8 weeks following proximal sciatic nerve crush, simple transection (0 mm gap), transection with a 4 m m gap (4 mm gap) and transection with an 8 mm gap (8 mm gap). For the last 3 procedures, the two ends of the nerve are placed in a tube and anchored in place by a stitch. The counts are presented in Table I, and means of these counts are shown graphically in Fig. 7. Normal sciatic nerves
35 TABLE I This table presents myelinated and unmyelinated axon counts from the sciatic nerve. The areas counted are in the tube for transections, in the distal stumps, and 2 mm distal to the ischial tuberosity of the unoperated side (Unop.). Lesion
Crush
Animal number
Axon numbers Myelinated Distal stump
Unop.
-
11007 10045 8842 10615 9752 10052 ± 834
8387 7378 8225 8001 8132 8025 ± 388
7 8 9 25
11805 13444 10807 17426 13371 ± 2914
12016 15986 10513 14589 13276± 2469
7886 8694 7948 7899 8107 ± 392
12 13 14 15 26
10355 2628 10260 10254 10892 8878 ± 3504
14243 10068 14258 15578 13781 13586± 2078
16 18 19 20 27
6334 2994 4499 4054 5328 4642 ± 1266
9644 4103 8398 5377 7708 7046 ± 2262
1 2 3 4 5
Mean ± S.D. 0 mm
Mean ± S.D. 4mm
Mean ± S.D. 8 mm
Mean ± S.D.
Unmyelinated
In tube
Distal stump
Unop.
12959 14566 9493 12979 13654 12730 ± 1925
17547 12660 17503 18444 14783 16187± 2402
17889 20063 17230 25288 20118 ± 3653
8905 9834 6941 13329 9752 ± 2672
13358 18066 15384 14922 15433± 1958
8283 8659 8966 8621 7451 8396± 581
14373 7436 15386 14427 18922 14109 ± 4168
11496 9304 7063 11258 13112 10447± 2325
14843 16805 15243 13786 13817 14899± 1241
8289 8106 7847 7379
13364 10387 10246 10025 8210 10446 ± 1853
12758 4038 6368 8544 6482 7638 ± 3277
17610 15964 14885 14128
7905 ± 395
In tube
15647± 1510
from the u n o p e r a t e d side at these levels contain an
parent nerve. Normal and r e g e n e r a t e d axon num-
average of 8210 myelinated axons and 15,542 unmye-
bers for tributary nerves are given in our previous pa-
linated axons. It is useful to normalize the axon counts from
per m. To obtain our comparisons, we divide the num-
these nerves. These ratios are diagrammed in Fig. 8.
side by the number in the parent sciatic nerve on the
ber of axons from each tributary on the u n o p e r a t e d
The dotted line represents mean axon numbers from
unoperated side of the same animal and co m p ar e this
normal nerves and the bars represent operated nerve
fraction to the n u m b e r of axons in the tributary nerve
numbers expressed in reference to these figures.
on the operated side divided by the n u m b e r of axons
Note that there are m o r e myelinated axons than nor-
in the distal stump of the sciatic nerve again in the
mal in the distal stump following crush, 0 and 4 mm
same animal. This gives a ratio, and these ratios are
gap transections but slightly fewer following the 8
presented in graphical form in Fig. 9. For the mye-
mm gap transection. Also note that there are fewer unmyelinated axons than normal no matter what sur-
linated fibers in the nerve to the medial gastrocnemi-
gical paradigm is used to interrupt the axons in the
bers in the sural nerve, note that the proportions that
parent nerve.
regenerate into the tributary nerves are the same as
us muscle and the myelinated and unmyelinated fi-
O n e of the m a j o r goals of the present paper is to
the proportions that regenerate into the distal stump
determine wh et h er the numbers of axons that regen-
of the parent nerve for the crush, 0 and 4 m m gap
erate into tributary nerves are proportional to the
transections. For the 8 m m gap transection, propor-
numbers that regenerate into the distal stump of the
tionally fewer axons regenerate into the tributary
36
(!)
following crush ol a ,omatic nerve 5.~'-w-:'. ~tnd (,,tH data arc confirmatory. M:10052+- 834 U:12730_+ 1925
By contrast tO the myelinated fibers, we find thai
,..:..-..2.....................-
........... ..-.+....
unmyelinated fibers decrease in the distal stump following regeneration after sciatic nerve crush in the rat. The reason for the different response of unmye-
M~13371_+2914 M:13276 + 2469 U:20118+31183 U:9752 +_2672
linated axons is not known, but it is clear thai neither
SC,A,,C
o
iiiii ~."2325:8g?13-+3504 1410~g_4118 i :10447+ :13586+2078
myelinated nor unmyelinated axon numbers return to normal following sciatic nerve crush in our system. Nevertheless we provide data that is in partial agreement with G u t m a n n and Sanders: in that the
SCIATIC 4mm ===========i=i=.=i=.== i =.={==. i .::::::::::::::::::::::::::::::::: i:iiii~ ~z4642+1266 : 1044~_+18~3
M:7046+ 2262 U:7638+_3277
SCIATIC8mm ~iiiiiiiii!iiiiiii!iii!i!iii!iiiiiiiiiii~
numbers of axons are less deviant from normal following regeneration distal to a crush than following any of our transections.
®
Fig. 7. A schematic diagram of our axonal counts in this study. For each procedure, the proximal stump is indicated by horizontal lines, the distal stump by large dots and the surgical gaps by small dots. The myelinated axon numbers _+ the standard deviations are indicated by M and the unmyelinated axon numbers + the standard deviations by I"
MYELINATEO UNMYELiNATEO
nerves than into the distal stump of the parent nerve. By contrast, proportionally more unmyelinated axons regenerate into the nerve to the medial gastrocnemius muscle than appear in the distal stump of the parent nerve no matter what surgical procedure is
1.2
used. A final goal in the present paper is to present axon counts at various levels from one sciatic nerve that regenerated through an 8 mm gap. These data are pre-
i
sented in schematic form in Fig. 10. Note that the greatest n u m b e r of axons are found in the proximal stump, and that the myelinated and unmyelinated axon counts are proportionally different in the tube and in the distal stump.
N
DISCUSSION Recovery from nerve crush is reported to be better than recovery following any type of transection7.m. G u t m a n n and Sanders 7 hypothesized that functional return is maximal when axonal numbers and sizes return to normal distally, and these investigators reported that myelinated axon n u m b e r s did return to normal after crush but not after transection;. Others found, however, that the n u m b e r of myelinated axons was increased in the distal stump at various times
Sciatic Crush
Sciatic omrn
Sciatic 4ram
ciatic
8ram
Fig. 8. A bar diagram of the normalized axon numbers in the distal stumps in this study. The dotted line indicates the mean numbers of axons from the unoperated side and the bars indicate the deviations from these numbers following the surgical procedures used in this study. Note that the myelinated axons are more numerous than normal after the crush, 0 and 4 mm gap transections. These differences are significant (P < 0:02). Note that there are fewer myelinated axons than normal aftcr the 8 mm gap transections. This difference is not significant. The unmyelinated axons are less numerous than normal after all surgical procedures and these differences are significant (l' < 0.02).
37
(!) ~NMG,My NMG,Un F~SN,My ~SN,Un
3,¢
22
~. ,.; 2. ¢o co
1.¢
O.E
...... Sciatic Crush
i
i.--i .
Sciatic Ornm
Sciatic 4mm
.
.
.
Sciatic 8ram
Fig. 9. A diagram indicating the proportion of axons that regenerate in tributary nerves as compared to normal tributary axon numbers in relation to the numbers that regenerate into the distal stump of the transected parent nerve as compared to normal axon numbers in the parent. To obtain these ratios, the numbers of axons in a tributary nerve (Oper. Trib.) are divided by the numbers in the distal stump of the parent sciatic nerve (Oper. Pare.) and this fraction is divided by the numbers from the tributary nerve on the unoperated side of the animal (Unop. Trib.) divided by the number of axons in the sciatic nerve on the unoperated side (Unop. Pare.). The dotted line at 1.0 indicates the ratios that obtain when the same proportion of axons regenerate in the tributary and parent nerves. Ratios less than 1.0 indicate that proportionally fewer axons regenerate in the tributary nerves and greater than 1.0 that proportionally more axons regenerate in the tributary nerves. Note that proportionally the same number of myelinated axons in the nerve to the medial gastrocnemius muscle (NMG, My), and the sural nerve (SN, My) and unmyelinated axons in the sural nerve (SN, Un), regenerate as regenerate into the distal stump of the parent sciatic nerve for the crush, 0 and 4 mm gap transections, but that proportionally fewer of these fibers regenerate in the tributary nerves for the 8 mm gap transection. By contrast proportionally more unmyelinated axons in the nerve to the medial gastrocnemius muscle (NMG, Un) regenerate no matter what surgical procedure is used on the sciatic nerve.
For simple transection, it has been reported that more myelinated axons regenerate into the distal stumpS,10.20, fewer myelinated axons regenerate into
the distal stumpl,3,4,7, 9, o r that more, the same, or fewer regenerate depending on conditions8,17. A partial explanation for this difference may be our finding that the number of axons that regenerate is related to the amount of nerve removed when the gap is made. In this regard, there are the clinical observations that the longer the length of the nerve removed, the more trouble the patient is likely to have 21, and that the longer the gap, the less likely axons are to cross 15. The only previous work, to our knowledge, where axonal numbers are related to the length of nerve removed shows that the number of axons that regenerate into tributary nerves at a particular time is partially dependent on the length of parent nerve that is removed 12. The relation is not a simple linear correlation, however. Two complicating factors are: (1) that myelinated and unmyelinated axons respond differently; and (2) that the numbers of unmyelinated axons that regenerate are different in a cutaneous as opposed to a motor tributary nerveH.1L Thus the next question, which is considered in the present paper, is whether the numbers of regenerating axons in the distal stump of the parent nerve are related to the numbers of regenerating axons in the tributary nerves? To normalize axon counts so that comparisons can be made between the distal stump of parent on the one hand and tributary nerves on the other, it is necessary to determine the proportions of myelinated and unmyelinated axons that regenerate in the different situations and then to compare these proportions with one another. The axonal numbers in tributary nerves have been reported previouslyt2, and the numbers in the distal stump are given in this paper.
® I~10395 U:16396
M:4499 U:10246
~.'9366 835 U=9449
I:~ox.
~8368 D I ST.
Fig. 10. A schematic diagram showing myelinated (M) and unmyelinated (U) axon counts from different levels of a sciatic nerve whose axons regenerated across an 8 mm gap. As before, the proximal stump (PROX) is indicated by horizontal lines, the distal stump (DIST) by large dots and the region of the gap by small dots. The two stumps and central gap region are contained in a silicone tube.
3~ I'he proportions are plotted in Fig. 9. Because they behave similarly, the myelinated fibers in the nerve to the medial gastrocnemius muscle and the myelinated and unmyelinated fibers in the sural nerve will be placed in one group. For these populations, the percentage of fibers, as c o m p a r e d to normal, that regenerate into the tributary nerves is the same as the percentage that regenerates into the distal stump of the parent nerve following crush and 0 and 4 mm gap transections. Thus in these cases, the normal tithe of axons that enters a tributary is maintained after regeneration, even though the numbers of regenerating axons in the distal stump of the parent and in the tributaries change considerably. A n o t h e r way of saying this is that the numbers of axons in tributary nerves change in concert with the numbers in the distal stump of the parent nerve for these particular surgical paradigms. For the 8 mm transection, however, the above conclusion does not hold. In this case proportionally fewer axons enter the tributary nerve than are found in the distal stump of the parent nerve. One explanation for this is that not enough time has been allowed for all axons to regenerate and thus if the samples were taken at a longer interval following surgery, there would not have been a discrepancy. This possibility will be checked. If this is not the explanation, it may be that when the length of the gap increases beyond a certain point, it becomes difficult for the organism to regulate the numbers of axons that regenerate into the tributary nerves. This would presumably imply either that some axons end blindly in the distal stump and so do not enter any tributary or if all axons in the distal stump enter a tributary, that some tributaries receive proportionally fewer axons and others receive more. We can determine experimentally which of these suggestions is correct. It will be important to do this for if there is a critical gap length beyond which the proportion of axons in parent and tributary nerves no longer are correlated, as seems to be the case, an understanding of this p h e n o m e n o n could have clinical impact. The unmyelinated fibers in the nerve to the medial gastrocnemius muscle clearly behave differently than the other fiber populations in this study. The data indicate that the r e g e n e r a t e d unmyelinated axons in the nerve to the medial gastrocnemius muscle are proportionally 2 - 3 times as numerous when compared to normal as the unmyelinated axon population
in the distal stump of the sciatic nerve no m~tter what type of surgery or how long the gap to interrupt the sciatic nerve. We have no explanation fo~ this but feel it may be important. Further data of interest wilt be to determine whether these fibers are sensory or postganglionic or both, and whether this is a characteristic of unmyelinated fibers in all nerves to muscles or only in the nerve to the medial gastocnemius muscle. A t present, the m a j o r value of the axonal counts in the tube is to allow a comparison with the numbers in the distal stump so as to get some indication of branching. The fact that axons branch at and slightly above the site of transection has been known since neuromas were investigated histologicallyl'L Furthermore the amount of branching presumably has a bearing on the.type of functional recovery m. The early histological work did not quantitate branching, but a combined anatomical and physiological method for quantitation of branching has recently been devised m. These investigators found that after transection without reapposition, each myelinated axon that regenerated through a transection site branched an average of 1.7 times. In this study, we find that after simple transection there are more axons than normal both in the tube and in the distal stump, but that the numbers of axons in these two locations are approximately equal. The simplest way to interpret these data would be to assume that there was considerable branching of myelinated axons proximal to the tube but that each axon in the tube passed into the distal stump without branching. A future task, therefore, is to determine the influence of various restorative procedures following transection on the amount of branching at or above the site of transection. Then the hypothesis that the procedure that results in the least amount of branching is associated with the best behavioral recovery will be tested. Since there are a greater number of myelinated axons in the distal stump than in the tube in the 4 and 8 mm gap transections, it seems necessary to postulate axonal branching in the tube m these conditions. If we divide the myelinated axon numbers in the distal stump by the number in the tube we get an average of 1.6. This is close to Horch and Lisney's branching factor of 1.7 m. We must repeat our work physiologically but if our anatomical data are accurate, it suggests that the length of the gap between the proximal
39 and distal stumps may be a factor in determining the amount of branching that the regenerating myelinated axons undergo at the site of transection. The unmyelinated axons regenerate differently than the myelinated axons in our paradigms. In our material, it is striking that the number of unmyelinated fibers in the tube is always much greater than in the distal stump. One interpretation of these data is that there is no branching of unmyelinated axons and that the excess axons end blindly in the tube. This is not definitive, of course, because it is possible that more axons than the 'excess' end blindly and that the remainder then branch. Physiological experiments will be necessary to choose between the possibilities. It is instructive in this case to examine the multilevel axon counts shown in Fig. 10. Here it can be seen that there are approximately 16,000 unmyelinated axons at the proximal end of the transection, which is essentially the normal number. In the tube, however, there are approximately 10,000 axons and in the distal stump there are approximately 6500 axons. Since the number at the proximal end is close to that obtained from the normal nerve, there is little evidence for branching at this site. The decrease in number in the tube then presumably indicates that 40% of the fibers in the proximal stump end when the axons leave the proximal stump and enter the tube. There is another significant loss of axons when the axons leave the tube and enter the distal stump which also suggests a site where a significant number of unmyelinated axons end blindly. This presumably indicates that the departure from the proximal stump and the entrance to the distal stump are two particularly important sites in that they both seem to be places where 'barriers' exist that prevent a siginificant number of unmyelinated axons from passing distally. The next point of this paper concerns the myelinated axon counts at different levels in the gap transection (Fig. 10). As for the unmyelinated axons,
REFERENCES 1 Cabaud, H. E., Rodkey, W. G., McCarroll, H. R., Jr., Mutz, S. C. and Niebauer, J. J., Epineurial and perineurial facicular nerve repairs: a critical comparison, J. Hand Surg., 1 (1976) 131-137. 2 Coggeshall, R. E., Coulter, J. D. and Willis, W. D., Jr., Unmyelinated axons in the ventral roots of the cat lumbosacral enlargement, J. comp. Neurol., 153 (1974) 39-58. 3 Davenport, H. A., Chor, H. and Cleveland, D. A., Fiber
there is a large reduction in axon numbers at the point where the axons leave the proximal stump and enter the tube. Then the numbers stay essentially constant down the length of the tube, although there is a low point distally. As the axons enter the distal stump, however, there is a sudden increase in axonal numbers. The easiest way to interpret these data is that many myelinated axons end blindly when they leave the proximal stump but then the remaining fibers branch when they enter the distal stump. Again the data need to be confirmed, but if they are, it is interesting that at the point of entry to the distal stump myelinated axons branch and unmyelinated axons drop out. As a final point, one of the noteworthy features of regeneration after transection is that the perineurium around the regenerated axons in the gap does not return to its normal configuration. There is an extensive literature on the morphology and barrier properties of perineurium in normal and regenerating nerves 22, but the perineurium where a nerve regenerates across a gap within a tube deserves more attention. One of the major difficulties in regeneration seems to be that transecting the perineurium exposes regenerating axons to the presumably alien environment of the extracellular fluid (e.g. ref. 10). It may be that placing two stumps of a transected nerve in a silicon tube makes up for any perineurium deficiencies. Further work on the morphology and barrier properties of the perineurium, and ionic content of the endoneurial fluid in regenerating nerves is thus in order. ACKNOWLEDGEMENTS This work is supported by N I H Grants NS 07377, NS 10161, NS 17039, NS 11255, and a support grant from the Moody Foundation.
ratios in regenerating nerves, II. The status of regrowth of the sciatic and nerves to the gastrocnemius in Macacus 3 months after section, J. comp. Neurol., 79 (1939) 153-159. 4 Davenport, H. A,, Chor, H. and Dolkart, R. E., The ratio of myelinated to unmyelinated fibers in regenerated sciatic nerves on Macacus rhesus, J. comp. Neurol., 67 (1937) 481-491. 5 Evans, D. H. L. and Murray, J. C., A study of regeneration in a motor nerve with a unimodal fiber diameter distribution, Anat. Rec., 126 (1956) 311-329.
4(t
7
8
9
10
11
12 13
14
Greenman, M. J., Studies on the peroneal nerve of the albino rat: number and sectional area of fibers: area relation ol axis to sheath, J. comp. Neurol., 23 (1913)479-513. Gutmann, E. and Sanders, F. K., Recovery of fiber num-. hers and diameters in the regeneration of peripheral nerves, J. Physiol. (Lond.), 101 (1943) 489-518. Guyon. L., Resultats anatomiques et fonctionnels observes au cours de la cicatrisation des nerfs chez la chien, Rev. neurol., 37 (1921) 937-949. Hausamen, J. E., Samii, M. and Schmidseder, R., Restoring sensation to the cut inferior alveolar nerve by direct anastomosis or by free autologous nerve grafting. Experimental study in rabbit. Hast. reconstr. Surg., 54 (1974) 83-87. Horch, K. W. and Lisney, S. J. W., On the number and nature of regenerating myelinated axons after lesions of cutaneous nerves in the cat, J. Physiol. (Lond.l, 313 (1981) 275-286. Jenq, C.-B. and Coggeshall, R. E., Effects of sciatic nerve regeneration on axonal populations in tributarv nerves, Brain Research, 295 (1984) 91-100. Jenq, C.-B. and Coggeshall, R. E., Regeneration of axons in tributary nerves, Brain Research, in press. Jenq, C.-B., Hulsebosch, C. E., Coggeshall, R. E. and Perez-Polo, R., The effect of nerve growth factor and its antibodies on axonal numbers in the medial gastrocnemius nerve of the rat, Brain Research, 299 (1984) 9-14. Langford, L. A. and Coggeshall, R. E., The use of potassium ferricyanide in neural fixation, Anat. Rec., 197 (1980)
297-303. 15 Lundberg, G., Dahlm, L. B., Danielsen, N., (iclberman. R. H.. Longo, F. M., Puwell, H ( . and Varon, S., Nerve regeneration in silicone chambers: influence o1 gap length and of distal stump components. Evp. Neurol.. 76 (1982) 361-375. 16 Lundberg, G., Gelberman, R. tI., Longo, R. M., Powell, H. C. and Varon, S., In vivo regeneration of cut nerve encased in silicone tubes, J. Neuropath. exp. Neurol., 41 (1982) 412-422. 17 Orgel, M. G. and Terzis, J. D., Epineuriat vs. perineurial repair: an ultrastructural and electrophysiological study of nerve regeneration, Plast. reconstr. Surg., 60 (1977)80-91. 18 Peters, A., Palay, S. L. and Webster, H., The Fine Struc-
ture of The Nervous System: The Neurons and Supporting Cells, W. B. Saunders Co., PhiLadelphia, 1976, pp. 1-4(16. 19 Ramon y Cajal, S., Degeneration and Regeneration of the Nervous System, Hafner Publishing, New York, 1959, 769 PP. 20 Shawe~ G. D. H., On the number o1 branches formed by 1egenerating nerve fibers, Brit. J. Surg., 42 (1955) 474-488. 21 Sunderland, Sir S., Nerves and Nerve Injuries, 2nd edn., Churchill Livingstone, Edinburgh, 1978, 1046 pp. 22 Thomas, P. K. and Olsson, Y., Microscopic anatomy and function of the connective tissue components of peripheral nerve. In P. J. Dyck, P. K. Thomas and E. H. Lambert (Eds.), Peripheral Neuropathy, W. B. Saunders Co., Philadelphia, 1975, pp. 168-189.