Peripheral Neuropathy after Chronic Endoneurial Ischemia John T. Sladky, MD," Raymond L. Tschoepe, BA,? Joel H. Greenberg, PhD," and Mark J. Brown, MD*
We have developed a method for producing chronic regional nerve ischemia in rats by creating proximal limb arteriovenous shunts. This procedure results in a 50 to 7590 reduction in endoneurial blood flow within the distal sciatic nerve as measured by the iodoantipyrine method. Nerve conduction velocities in sciatic nerves ipsilateral to the shunt fell by 25 to 30% within 2 weeks after creation of the shunt and did not recover for u p to 10 months after the procedure. Morphological studies of the ischemic nerves showed structural abnormalities at nodes of Ranvier and mild axonal atrophy. Neither segmental demyelination nor axonal degeneration were evident. These results indicate that reduced endoneurial blood flow, insufficient to cause infarction, may result in rneasureable functional and morphological abnormalities in peripheral nerves. Sladky JT, Tschoepe RL, Greenberg JH, Brown MJ. Peripheral neuropathy after chronic endoneurial ischemia. Ann Neurol 1991;29:272-278
Ischemia is important for the development of neuropathies associated with systemic vascular diseases [1-3) and may have a secondary role in the pathogenesis of chronic neuropathic disorders that are not usually considered to have a vascular etiology. Confirmatory evidence for the injurious effects of chronic hypoperfusion ha5 been elusive, in part because most patients at risk for nerve ischemia have other conditions that also may cause neuropathy. Conventional experimental models of neuropathies with chronic ischemia are associated with concomitant toxic or metabolic derangements that also may be injurious [4-7). Attempts to study the influence of isolated chronic ischemia on nerves have been thwarted by the rapid revascularization that occurs after experimental occlusion of regional feeding vessels [S-10). In this study, we report a method for producing focal chronic endoneurial ischemia in rats by creating a proximal arteriovenous (A-V) shunt, a procedure that leads to diversion of blood from the distal limb. By using this technique, we were able to demonstrate that sublethal chronic ischemia is sufficient to cause pronounced abnormalities of nerve conduction associated with pathological changes at nodes of Ranvier and mild reduction of axonal caliber.
Methods A series of experiments were performed to validate and characterize this new model of chronic ischemia. To measure
From the Departments of *Neurology and +Surgery, University of Pennsylvania School of hfedicine, Philadelphia, PA. Received Dcc 11, 1989, and in revised form Jun 28 and Aug 31, 1990. Accepted for publication Sep 2, 1990.
the severity and pattern of endoneurial hypoperfusion in our experimental model, we studied a group of 5 male Wistar rats 6 weeks after creation of the shunt. For correlation of physiological abnormalities with histological changes, a second series of 15 rats had nerve conduction measurements and then an A-V shunt was performed in 10. In 5 rats, the femoral vessels were exposed and clamped for 60 minutes, approximately the same duration as required for shunt surgery, but the anastomosis was not performed. These rats served as sham-operated controls. After 6 weeks, conduction studies were repeated and sciatic nerves were removed from 6 experimental rats and the 5 sham-operated rats for morphological analysis. To determine whether nerve conduction slowing was transient, 4 shunted rats were followed with serial nerve conduction studies at 6 weeks and subsequently at 3 , 5 , and 10 months after shunting. 'To examine the temporal evolution of physiological abnormalities in ischemic nerves, an independent group of 5 rats of similar age and weight underwent serial nerve conduction srudies before and at 1, 2, 3, 4 , and 6 weeks after vascular surgery.
Arterzovenozls Anastomosis Femoral A-V shunts were created in a microvascular research laboratory. Male Wistar rats (275-300 gm) were anesthetized with pentobarbital (50 mgikg i.p.) and placed in a supine position. Under sterile conditions, an anterior incision was made in the skin over the left groin and the femoral artery and vein were exposed, isolated, and held in apposition with vascular clamps. A side-to-side anastomosis was performed just distal to the level of the inguinal ligament, approximately 15 mm distal to the origin of the superficial
Address correspondence to Dr Sladky, Division of Neurology, Children's Hospital of Philadelphia, 34th St and Civic Center Blvd, Philade'phia, PA 19104.
272 Copyright 0 1991 by the American Neurological Association
circumflex iliac artery. The patency of the shunt was verified by inspection, the wound closed, and the rats allowed to awaken from anesthesia.
lateral to the A-V shunt were compared by using a paired t test. For comparison between experimental and sham-operaced rats, a nonpaired two-tailed t test was used.
Measurement of Endoneuriul Blood Flow
Morjhologicul Stndies Sciatic-posterior tibial nerves ipsilateral and contralateral to the A-V shunt or sham operation were fixed in sku and then after removal in 3.6% glutaraidehyde, postfixed in osmium ferrocyanide, dehydrated in alcohol, and embedded in epoxy. Cross- and longitudinal sections were obtained from sciatic nerves 1 cm proximal to the origin of the posterior tibial branch, at the origin, and from posterior tibial nerves 1 cm and 2 cm distal to the origin of the nerve. Longitudinal sections were made at three levels from intervening nerve segments between cross-sections. Cross- and longitudinal sections were examined by light microscopy for evidence of neuropathy. The myelinated fiber number and density, external circumference, and cross-sectional area were measured with a digitizer/computer from photomicrographs of nerves ipsilateral and convalaceral to the shunt at levels 1 cm proximal and 2 cm distal to the origin of the posterior tibial nerves in three rats. Data from these nerves were pooled and mean axonal area, circumference, and circumference to area ratios were compared by using a nonpaired t test. The myelinated fiber population profiles were displayed in standard size-frequency histograms. Osmicated segments from distal posterior tibial nerves (between 1 cm and 2 cm distal to the origin) were placed in glycerin and teased in creosocc onto glass slides. Thirty single teased fibers from each of 6 experimental and 6 control nerves were coded and examined by light microscopy for pathological changes. Longitudinal sections from ischemic and contralateral nerves from these 6 rats were examined by light and electron microscopy.
Nerve blood flow was measured 6 weeks after creation of the A-V anastomosis by using a previously reported method [ll].Rats were anesthetized with halothane to permit placement of carotid arterial and jugular venous catheters and a tracheostomy, and a ligature was loosely positioned around the abdominal aorta distal to the renal arteries. When vital signs were stable, {'4C)iodoantipyrine (200 pCi/kg) was infused intravenously over 1 minute. During the infusion, serial arterial blood samples were obtained at approximately i-second intervals. At the end of the infusion, the rats were killed by transection of the abdominal aorta. The sciatic-posterior tibial nerves were rapidly dissected free and frozen in liquid nitrogen. Autoradiographs were made from 20-km-thick frozen nerve cross-sections at five levels along the course of the nerves. Sections were taken 2 cm and 1 cm above, at the origin of the posterior tibial branch, and 1 cm and 2 cm below that landmark. The autoradiographs were analyzed by using a video processing system, and tissue '"C concentrations were calculated from autoradiographic densities calibrated to the standard curve. Endoneurial blood flow was calculated from the derived arterial I4C curve and the measured tissue '*C concentrations by using the equation of Kety and colleagues ClZ]. Because the A-V shunt may have affected perfusion of the contralateral nerve, regional blood flow values in ipsilateral sciatic nerves were compared with previously studied controls from this laboratory [l 11. Differences in regional blood flow along the course of ischemic and control nerves were compared by using a Wilcoxon rank sum test and the significance of differences between groups calculated with a nonpaired two-tailed t test.
Nerve Conduction Measurementf Rats were anesthetized with pentobarbital and placed in a prone position. Core temperature was maintained at 37°C with a rectal probe servocoupled to a heating blanket. Nearnerve temperature of operated and unoperated rats was approximately 34°C as monitored with a subcutaneous temperature probe. Sciatic motor nerve responses were studied by recording evoked intrinsic foot muscIe action potentials with bipolar subdermal electrodes after nerve stimulation at the ankle and the sciatic notch. Latency, amplitude, and duration of the evoked compound motor action potentials (CMAP) after proximal and distal stimulation were measured and sciatic motor conduction velocity (MNCV) calculated. The presence of dispersion or conduction block of the proximally evoked CMAP was assessed by comparing the amplitude, duration, and configuration of the responses. Sciatic mixed nerve conduction was measured with subdermal stimulating electrodes positioned adjacent to the posterior tibial nerve at the ankle and subdermal recording electrodes placed adjacent to the sciatic netve at the level of the sciatic notch. The latencies and amplitudes of the evoked nerve action potential (NAP) were measured and the mixed nerve conduction velocities calculated. Differences in conduction velocities and evoked potential amplitudes in nerves ipsilateral and contra-
Results Rats were clinically normal after recovering from surgery aside from 2 that had weakness of intrinsic foot muscles ipsilateral t o the A-V shunt for 24 hours after the procedure. Within 2 weeks after surgery, t h e shunted legs were visibly enlarged. Skin and subcutaneous tissues and t h e limbs and feet of all rats appeared normal at t h e time of electrophysiological and blood flow studies.
Meaurement of Endoneuriul Blood Flmu Endoneurial perfusion was normal in proximal sciatic nerves 6 weeks after creation of t h e shunt (12.1 5 4.4 m l / l 00 gmimin, control, 1I .g t 2.5 mll1OO gm/min) (Table 1). A zone of ischemia was detected distal t o this that began just proximal to t h e origin of the posterior tibial nerve. Endoneurial perfusion progressively declined in this ischemic z o n e and was lowest in the most distal nerve segments studied (3.6 t 1.7 ml/100 gmimin) (see Table 1).
Nerrie Conduction Measwements Sciatic nerve conduction velocities slowed t o ?4% of normal within 2 weeks after creation of the shunt and
Sladky et al: Chronic Ischemic Neuropathy
273
Table 1. Endoneurial Blood Flow along the Course of Sciatic-Posterior Tibzal N m e s 6 Weeks after a Proximal Femoral A-V Shunt
Site
Ischemic"
Controlb
2 cm Proximal 1 cm Proximal
12.1 2 4.4 8.7 k 4.0 8.2 2 4.4 4.4 2 2.2 3.6 1.7
11.9 t 2.5 9.7 ? 1.6 13.9 2 5.7 10.9 t 1.2
Origin
1 cm Distal 2 cm Distal
*
11.9
* 1.2
w-
P ,.
NS NS NS
< 0.001 < 0.001
Values (mean 2 1 SD) are expressed in (m11100 gmimin). 'Experimental Group 1 (n = 5).
0
1
1
3
4
5
6
Weeks after surgery ' in mlsec (mean
A-V = arteriovenous; NS = not significant.
+
I S E1
Fig 1. Evoltltion of slouiing of rat .rciatic motor nerve conduction velocity during experimental chronic ischemia (Group I. n = 5). Conduction decreared t o 74% of control values within 2 weeks of i.rchemia and remuined slow ocer the 6-week period of observation. Error ban are 1 SEM.
remained slow over the next 4 weeks (Fig 1).At postoperative week 6 , sciatic MNCV ipsilateral to the fistula was 32.6 2 4.3 m/sec (mean -t SD) compared with 47.8 6.2 m/sec in the contralateral nerve ( p < 0.005) (Table 2). The amplitudes of the evoked CMAPs were not significantly diminished in cxperimental nerves, and there was no evidence of dispersion of the proximally evoked responses, nor was conduction block evident (Table 3). Sciatic mixed nerve conduction velocity was similarly reduced to 32.8 k 8.6 mlsec on the ischemic side compared with 54.7 -+ 4.2 m/sec on the unoperated side (a < 0.003). The amplitudes of the sensory NAPS were similar on the control and ischemic sides without evidence of dispersion (see Table 3). Sciatic motor and mixed nerve conduction measurements in sham-operated rats were symmetrical and not different from those of unoperated nerves contralateral to the A-V shunt in experimental rats (see Table 2). Among the rats followed for up to 10 months after creation of the shunts (n = 4), the initial fall in nerve conduction velocities was similar to that in the shortcrterm rats. MNCVs were 34.9 k 2.8 after 6 weeks. In 2 rats studied, the disparity between sciatic MNCV in shunted compared with contralateral limbs persisted after 5 months (mean, 39 vs. 55 m/sec, respectively). This side-to-side difference remained after 10 months in 3 rats tested (mean, 33 ipsilateral vs. 48 misec con-
*
tralateral). One rat died from an anesthetic complication before electrophysiological studies could be performed. Morphological Analyses Dilated shunt vessels were prominent at the site of surgery in the right groin, but the superficial sciatic epineural vessels were not visibly enlarged ipsilateral to the shunt. Examination of cross- and longitudinal sections obtained along the proximal-distal extent of sciatic nerves failed to show signs of axonal breakdown or demyelination (Fig 2). Myelinated fiber numbers and densities were similar on the operated and unoperated sides at both prommal and distal levels. Mean sciatic axonal areas from segments taken 1 cm proximal to the origin of the posterior tibial branch were similar from nerves ipsilateral and contralateral to the shunt (47.8 t 33.9 vs. 48.1 34.2, respectively). Within the distal posterior tibial nerve at a point 2 cm below its origin, however, there was a left shift in the sizefrequency spectrum of axon profiles toward smaller diameter fibers; the mean axonal area was mildly de-
*
Table 2.Sciatic Nerve Conduction Velocities a&er 6 Weeh of Ischemia
Sham-operated Rats (n = 5 )
Shunted Rats (n = 10) Ipsilateral
32.6 32.8
Motor NCV (m/sec) Mixed NCV (m/sec) Values are mean
?
* 4.3" * 8.6'
Contralateral
49.8
* 6.2
54.7
?
4.2
1 SD.
"Significantly different from contralateral nerve and from sham-operated rats (p < 0.001).
NCV
=
nerve conduction velocity.
274 Annals of Neurology
Vol 29
N o 3 March 1991
Ipsilateral
*
4.2 49.6 58.4 2 7.8
Contralateral
48.8 57.7
k ?
4.9 9.1
Table 3. Sciatic Evoked Potential AmpLitudes after 6 Weeks of Ischemia Experimental Group (n
Proximal CMAP (mV) Distal CMAP (mV) Proximal:distal ratio Mixed NAP amplitude CpV)
=
Ischemic
Contralateral
12.0 2 5.1
10.3
13.2
?
5.7
0.92 t 0.04 24.4 2 11.4
10)
p
3.8
NS
11.8 t 4.7
NS NS NS
0.89 25.8
Values are mean 2 1 SD. CMAP = compound motor action potential; NAP potential; NS = not significant.
2
* 0.07 f
=
11.1
nerve action
creased to 42.8 +- 24.3 versus 48.4 k 24.3 Fm2 on the unoperated side (mean ? 1 SD, p < 0.001) (see Figs 2, 3; Table 4). The incidence of axonal degeneration and segmental demyelination was not increased in teased fibers after 6 weeks of ischemia compared with unoperated contralateral nerves. Nodal morphology was abnormal within the ischemic nerves, however, especially in posterior tibial nerve segments between 1 and 2 cm distal to the origin. Juxtanodal portions of axons frequently were swollen and paranodal myelin was dysmorphic, with evidence of degeneration and remodeling, and there was symmetrical and asymmetrical paranodal myelin retraction with widened nodal gaps (Fig 4). These features were not present in similarly handled contralateral or sham-operated nerves. Electron microscopy of nodes within ischemic nerve segments showed simplification and degeneration of paranodal Schwann cell foot processes, with jutanodal myelin degeneration and nodal widening (Fig 5). With the exceptions of paranodal swelling and occasional herniation of axoplasm at nodes of Ranvier, axons appeared normal.
Discussion Our results demonstrate that chronic nerve ischemia can be produced experimentally despite peripheral nerve’s capacity to establish collateral circulation within minutes after occlusion of nutrient vessels. A state of chronic endoneurial ischemia was created by shunting or “stealing” blood away from the distal leg through a high-flow, low-resistance proximal A-V shunt. This is similar to the phenomenon that may occur in patients with fistulae created surgically to facilitate hemodialysis [13-151. Some of these individuals develop a monomelic ischemic neuropathy with arm pain and weakness distal to the shunt. The majority of reported patients also were diabetic and uremic, however, and therefore had preexisting metabolic and vascular disease that may have predisposed to neuropathy in an ischemic environmen t. In this study, moderately severe, noninfarctive chronic ischemia led to slowing of conduction along
Fig 2. Cross-section at the lad ofthe midposterior tibial nerve afier G uleeh of ischemia. There i.r occasional crenation of myelin profiles mgge.rtiny mild axonal atrophy. hut neither axonal degeneration nor demyelination are present. (originalmagniJication, x 410.) Despite its relathJelynormal appearance, sciatic motor neme condtlction velocity w ~ 15 s misec in this nerve, approximately 70% of controls.
motor and sensory axons. The pathophysiological basis for this slowing is not known. Although segmental demyelination is commonly invoked to explain conduction slowing of this degree, morphological analysis did not detect evidence of demyelination along affected nerves. Loss of large diameter, fast-conducting axons could slow conduction, but axonal degeneration was not evident. Block of the fastest conducting axons could explain NCV slowing. This is an unlikely explanation because CMAP amplitudes, CMAP proximal: distal amplitude ratios, and sciatic mixed nerve NAP amplitudes were similar in ischemic and control nerves. Axonal atrophy may also be associated with slowing of conduction velocity in a number of neuropathies. The degree of slowing that we observed in this model is difficult to ascribe to mild axonal atrophy that is confined to distal axon segments (see Figs 2, 3). A comparable degree of axonal atrophy from experimental protein-calorie deprivation in rats had no effect on motor or mixed sciatic nerve conduction velocities {16}. We speculate that conduction slowing was the consequence of ischemia-induced alterations at nodes of Ranvier, with increased nodal capacitance and slowing of saltatory conduction. Pathological changes in Schwann cell-axon relations at the nodes of Ranvier were striking within the ischemic zone. We found degeneration of Schwann cell foot processes and paranodal myelin with retraction of juxtanodal myelin and widening of nodal gaps. Previous reports have documented slowing of nerve conduction in association with paranodal demyelination { 171, and modeling studies predict that the most striking effects on saltatory conduction are related to changes in membrane capaciSladky et al: Chronic Ischemic Neuropathy
275
Ischemic Posterior Tbiol Nerve
Contmloteral Posterior Tibia1 Nerve
-..- pi 1
lo
0
m
I
7 6
5 4
3 2
1 0 0
Cross-sectional Axonal Area in square microns
10
20
50
40
70
UJ
W 100 110
la0 150 140 150
Cross-sectional Axonal Area in square microns
Contralateml Mid-Sciatic Nerve
la 91
n
1
c
c D
Q
o
20 s0 u1 50 (YI 70 w 90 im 110 120 im 140 Cross-sectional Axonol Area in square microns
10
150
Fig 3. Morphometric analysis of axon cross-sectionalureu pr0JEle.s at the leziel of the midsciatic und distal posterior tibia1 nerves obtainedfrom paived newes ip..tilatrraland contralateralto a femoral avterioven0u.r shunt. Data is pooled from 3 ruts.
tance at the juxtanodal axolemma [18). Similar nodal alterations have been reported in human and experimental diabetic neuropathy 119-2 1) and after experimental pyridoxine intoxication [22}. In both settings, slowing of nerve conduction occurs in parallel with pathological changes at nodes {23-251. In experimental diabetes, abnormalities in sodium conductance have been measured at the nodal axolemma {23}. Nerve blood flow may be transiently reduced after vascular occlusion to just 20 to 25% of normal without apparent functional or pathological consequences Ill}. Peripheral nerve lacks the capacity to autoregulate blood flow [24},but ischemic nerve may survive acute feeding vessel occlusion by developing enough collateral circulation to subsist with compensatory changes in glucose metabolism to maintain bioenergetic homeostasis [25}. In the present study, flow was diminished to a similar degree but maintained at this lower level. Within a few weeks, there was conduction slowing, nodal injury, and axonal atrophy. Functional and
276
Annals of Neurology
Vol 29
No 3
March 1991
0 o
20 m u) 50 M 70 w 90 im i i o 120 is0 iu) is Cross-sectional Axonal Area in square microns
1100
structural abnormalities persisted without appreciable change for up to 10 months. This failure to compensate may result from circulatory factors related to the mechanism for producing ischemia. An A-V shunt causes both diminished perfusion pressure in the endoneurial capillary bed and increased venous resistance. Both factors can diminish endoneurial perfusion {4,26f and, in concert, retard the reestablishment of sufficient collateral circulation to permit normal aerobic metabolism. It is not surprising that chronic endoneurial ischemia led to slowing of conduction and distal axonal atrophy. Systems in the peripheral nerve with the highest energy requirements are thought to be especially vulnerable to hypoxia. This includes the nodal transmembrane ion exchange, funclamental to axon potential generation and axoplasmic flow [27} necessary for the transport of filaments, the major determinant of axonal caliber [28]. It is likely that Schwann cell ischemia had deleterious consequences on myelin maintenance. This may prove to be the mechanism for paranodal myelin abnormalities in chronic ischemia. We conclude from these experimental observations that it is possible to produce chronic experimental nerve ischemia in the absence of concomitant toxic or
Table 4. Morphometric Analysis of Myelinated Fibers after 6 Weeks of Ischemia" Axonal Circumference (pm2) Left sciatic (n = 6,097) Right sciatic (n = 5,722) Left posterior tibial (n = 4,558) Right posterior tibial (n = 3,791) Values are mean
k
26.7 26.6 24.4 26.6
* 10.4 * 10.4 ? ?
7.9b 8.8
Axonal Area (pm2)
47.8 48.1 42.8 48.4
* 33.9 ? ?
*
34.2 24.3b 29.2
Circumference/Area Ratio
0.84 L 0.81 1.1 0.85 0.73 2 0.34b 0.69 t 0.33
*
1 SD.
'Paired left (ischemic) and right (contralateral) sciatic nerves (1 cm proximal to the origin of the posterior tibial) and posterior tibial nerves (2 cm distal to the origin). Morphometric data pooled from 3 rats. bSignificantlydifferent from contralateral (j < 0.001).
Fig 4. Longitudinal section of rat posterior tibial nerve after 6 weeekr of ischemia. Nodal appearance rangesfrom normal (arrowhead) to markedly abnormal (arrow), with widening andparanodal swelling. A n equivocally widened node is also present (asterisk). (original magnification, x 650.)
metabolic insults by diverting the blood supply to the distal limb through a proximal arteriovenous shunt. In this model, chronic sublethal nerve ischemia slowed nerve conduction velocities to 50 to 75% of normal. There were alterations in nodal morphology and axon caliber but not segmental demyelination or axonal degeneration. These physiological and pathological consequences of chronic nerve ischemia are quite distinct from those of acute nerve ischemia where focal conduction block or nerve infarction are encountered ( 8 , 10, 29, 30). Future studies are necessary to determine whether the electrophysiological and morphological abnormalities that result from isolated chronic mild ischemia progress to demyelination, axonal degeneration, and clinical evidence of peripheral nerve dysfunction. This work was supported by CIDA Grant NS-01077 (to J. T. S.), the Muscular Dystrophy Association, and National Institutes of Health Grant NS-08075.
We thank Doreen Cashen for expert technical assistance and James Goin, PhD, for advice regarding statistical analyses.
References
Fig 5 . Electron micrograph of a node of Ranvier from the posterior tibial nerve after 6 weeks of chronic ischemia. There is simplz5cation and remodeling of Schwann cell foot processes with loss of Schwann cell-axolemma tight junctions, myelin retraction, and nodal widening. (original magnification, x 4,800.)
1. Eames RA, Iange LS. Clinical and pathological study of ischaemic neuropathy. J Neurol Neurosurg Psychiatry 1967;30: 215-226 2. Dyck PJ, Conn DL, Okazaki H . Necrotizing angiopathic neuropathy. Mayo Clin Proc 1972;47:461-475 3. Dyck PJ, Karnes JL, OBrien P, et al. The spatial distribution of fiber loss in diabetic polyneuropathy suggests ischemia. Ann Neurol 1986;19:450-45 7 4. Tuck RR, Schmelzer JD, Low PA. Endoneurial blood flow and oxygen tension in the sciatic nerve of rats with experimental diabetic neuropathy. Brain 1984;107:935-950 5. Low PA, Nukada H, Schmelzer JD, et al. Endoneurial oxygen tension and radial topography in nerve edema. Brain Res 1985;341:147-154 6. Meyers RR, Mizisin AP, Powell HC, Lampert WP. Reduced nerve blood flow in hexachlorophene neuropathy. Relationship to elevated endoneurial fluid pressure. J Neuropathol Exp Neurol 1982;41:391-399 7. Meyers RR, Powell HC. Galactose neuropathy: impact of chronic endoneurid edema on nerve blood flow. Ann Neurol 1984;16:587-594
Sladky et al: Chronic Ischemic Neuropathy 277
8. Parry GJ, Brown MJ. Arachidonate-induced experimental nerve infarction. J Neuroi Sci 1981;50:123-133 9. Korthals JK, Wisneiwski HM. Peripheral nerve ischemia: part 1. Experimental model. J Neurol Sci 1975;24:65-76 10. Nukada H, Dyck PJ. Microsphere embolization of nerve capillaries and fiber degeneration. Am J Pathol 1984;115:275-287 11. Sladky JT, Greenberg JH, Brown MJ. Regional perfusion in normal and ischemic rat sciatic nerve. Ann Neurol 1985: 17:191-195 12. Sakurada 0, Kennedy C, Jehle J, et al. Measurement of local cerebral blood flow with iodo-[14-C) antipyrine. Am J Physiol 1978;234:H59-H66 13. Wilbourn AJ, Furlan AJ, Hulley W, Ruschhaup TW. Ischemic monomelic neuropathy. Neurology 1983;33:447-45 1 14. Knezevic W, Mastaglia FL. Neuropathy associated with BresicaCimino arteriovenous fistulas. Arch Neurol 1984;41:11841192 15. Rqgs JE, Moss AH, Labosky DA, et al. Upper extremity ischemic monomelic neuropathy: a complication of vascular access procedures in uremic diabetic patients. Neurology 1989;39: 997-998 16. Cornblath D, Brown MJ. Influence of malnutrition on developing rat peripheral nerves. Exp Neurol 1988;99:403-411 17. LaFontaine S, Rasminsky M, Saida T, Sumner AJ. Conduction block in rat myelinated fibers following acute exposure to antigalactocerebroside serum. J Physiol 1982;323:287-306 18. Koles ZJ, Rasminsky M. A computer simulation of conduction in demyelinated nerve fibers. J Physiol 1972;227:351-364 19. Sima AAF, Bril V, Nathaniel V, et al. Regeneration and repair of myelinated fibers in sural-nerve biopsy specimens from patients with diabetic neuropathy treated with Sorbinil. N Engl J Med 1988;319:548-5 5 5 20. Sima AAF, Lattimer SA, Yagihashi S, Green DA. Axo-glial dys-
278 Annals of Neurology Vol 29 No 3 March 1991
junction: a novel structural lesion that accounts for poorly reversible slowing of nerve conduction in the spontaneously diabetic rat. J Clin Invest 1986;77:474-484 21. Brismar T, Sima AAF, Green DA. Reversible and irreversible nodal dysfunction in diabetic neuropathy. Ann Neurol 1987; 21:504-507 22. Sladky JT, Xu Y, Brown MJ. Transient ataxia after acute pyridoxine intoxication: electrophysiological and neuropathological correlations. Neurology 1987;37(suppl 1):376 23. Brismar T, Sima AAF. Changes in the nodal function in nerve fibers of the spontaneously diabetic BB-Wistar rat: potential clamp analysis. Acta Physiol Scand 1981;113:499-506 24. Low PA, Tuck RR. Effects of changes of blood pressure, respiratory acidosis, and hypoxia on blood flow in the sciatic nerve of the rat. J Physiol 1984;347:513-524 25. SladkyJT, GreenbergJH, Brown MJ. Enhanced 2-deoxyglucose incorporation in peripheral nerve during ischemia. Brain Res 1987;414:323-329 26. Simpson LO. Altered blood rheology in the pathogenesis of diabetic and other neuropathies. Muscle Nerve 1988;ll: 725-744 27. Hoffman PN, Griffin JW, Price DL. Control of axon caliber by. neurofilament transport. J Cell Biol 1984;99:705-714 28. Hoffman PN, Thompson GW, Griffin JW, Price DL. Changes in neurofilament transport coincide temporally with alterations in the caliber of axons in regenerating motor fibers. J Cell Biol 1985;101:1332-1340 29. Fowler CJ, Gilliat RW. Conduction velocity and conduction block after experimental ischemic nerve injury. J Neurol Sci 1981;52:221-228 30. Parry GJ, Linn DJ. Conduction block without demyelination following acute nerve infarction. J Neurol Sci 1988;84: 265-273
We have developed a method for producing chronic regional nerve ischemia in rats by creating proximal limb arteriovenous shunts. This procedure results in a 50 to 7590 reduction in endoneurial blood flow within the distal sciatic nerve as measured by the iodoantipyrine method. Nerve conduction velocities in sciatic nerves ipsilateral to the shunt fell by 25 to 30% within 2 weeks after creation of the shunt and did not recover for u p to 10 months after the procedure. Morphological studies of the ischemic nerves showed structural abnormalities at nodes of Ranvier and mild axonal atrophy. Neither segmental demyelination nor axonal degeneration were evident. These results indicate that reduced endoneurial blood flow, insufficient to cause infarction, may result in rneasureable functional and morphological abnormalities in peripheral nerves. Sladky JT, Tschoepe RL, Greenberg JH, Brown MJ. Peripheral neuropathy after chronic endoneurial ischemia. Ann Neurol 1991;29:272-278
Ischemia is important for the development of neuropathies associated with systemic vascular diseases [1-3) and may have a secondary role in the pathogenesis of chronic neuropathic disorders that are not usually considered to have a vascular etiology. Confirmatory evidence for the injurious effects of chronic hypoperfusion ha5 been elusive, in part because most patients at risk for nerve ischemia have other conditions that also may cause neuropathy. Conventional experimental models of neuropathies with chronic ischemia are associated with concomitant toxic or metabolic derangements that also may be injurious [4-7). Attempts to study the influence of isolated chronic ischemia on nerves have been thwarted by the rapid revascularization that occurs after experimental occlusion of regional feeding vessels [S-10). In this study, we report a method for producing focal chronic endoneurial ischemia in rats by creating a proximal arteriovenous (A-V) shunt, a procedure that leads to diversion of blood from the distal limb. By using this technique, we were able to demonstrate that sublethal chronic ischemia is sufficient to cause pronounced abnormalities of nerve conduction associated with pathological changes at nodes of Ranvier and mild reduction of axonal caliber.
Methods A series of experiments were performed to validate and characterize this new model of chronic ischemia. To measure
From the Departments of *Neurology and +Surgery, University of Pennsylvania School of hfedicine, Philadelphia, PA. Received Dcc 11, 1989, and in revised form Jun 28 and Aug 31, 1990. Accepted for publication Sep 2, 1990.
the severity and pattern of endoneurial hypoperfusion in our experimental model, we studied a group of 5 male Wistar rats 6 weeks after creation of the shunt. For correlation of physiological abnormalities with histological changes, a second series of 15 rats had nerve conduction measurements and then an A-V shunt was performed in 10. In 5 rats, the femoral vessels were exposed and clamped for 60 minutes, approximately the same duration as required for shunt surgery, but the anastomosis was not performed. These rats served as sham-operated controls. After 6 weeks, conduction studies were repeated and sciatic nerves were removed from 6 experimental rats and the 5 sham-operated rats for morphological analysis. To determine whether nerve conduction slowing was transient, 4 shunted rats were followed with serial nerve conduction studies at 6 weeks and subsequently at 3 , 5 , and 10 months after shunting. 'To examine the temporal evolution of physiological abnormalities in ischemic nerves, an independent group of 5 rats of similar age and weight underwent serial nerve conduction srudies before and at 1, 2, 3, 4 , and 6 weeks after vascular surgery.
Arterzovenozls Anastomosis Femoral A-V shunts were created in a microvascular research laboratory. Male Wistar rats (275-300 gm) were anesthetized with pentobarbital (50 mgikg i.p.) and placed in a supine position. Under sterile conditions, an anterior incision was made in the skin over the left groin and the femoral artery and vein were exposed, isolated, and held in apposition with vascular clamps. A side-to-side anastomosis was performed just distal to the level of the inguinal ligament, approximately 15 mm distal to the origin of the superficial
Address correspondence to Dr Sladky, Division of Neurology, Children's Hospital of Philadelphia, 34th St and Civic Center Blvd, Philade'phia, PA 19104.
272 Copyright 0 1991 by the American Neurological Association
circumflex iliac artery. The patency of the shunt was verified by inspection, the wound closed, and the rats allowed to awaken from anesthesia.
lateral to the A-V shunt were compared by using a paired t test. For comparison between experimental and sham-operaced rats, a nonpaired two-tailed t test was used.
Measurement of Endoneuriul Blood Flow
Morjhologicul Stndies Sciatic-posterior tibial nerves ipsilateral and contralateral to the A-V shunt or sham operation were fixed in sku and then after removal in 3.6% glutaraidehyde, postfixed in osmium ferrocyanide, dehydrated in alcohol, and embedded in epoxy. Cross- and longitudinal sections were obtained from sciatic nerves 1 cm proximal to the origin of the posterior tibial branch, at the origin, and from posterior tibial nerves 1 cm and 2 cm distal to the origin of the nerve. Longitudinal sections were made at three levels from intervening nerve segments between cross-sections. Cross- and longitudinal sections were examined by light microscopy for evidence of neuropathy. The myelinated fiber number and density, external circumference, and cross-sectional area were measured with a digitizer/computer from photomicrographs of nerves ipsilateral and convalaceral to the shunt at levels 1 cm proximal and 2 cm distal to the origin of the posterior tibial nerves in three rats. Data from these nerves were pooled and mean axonal area, circumference, and circumference to area ratios were compared by using a nonpaired t test. The myelinated fiber population profiles were displayed in standard size-frequency histograms. Osmicated segments from distal posterior tibial nerves (between 1 cm and 2 cm distal to the origin) were placed in glycerin and teased in creosocc onto glass slides. Thirty single teased fibers from each of 6 experimental and 6 control nerves were coded and examined by light microscopy for pathological changes. Longitudinal sections from ischemic and contralateral nerves from these 6 rats were examined by light and electron microscopy.
Nerve blood flow was measured 6 weeks after creation of the A-V anastomosis by using a previously reported method [ll].Rats were anesthetized with halothane to permit placement of carotid arterial and jugular venous catheters and a tracheostomy, and a ligature was loosely positioned around the abdominal aorta distal to the renal arteries. When vital signs were stable, {'4C)iodoantipyrine (200 pCi/kg) was infused intravenously over 1 minute. During the infusion, serial arterial blood samples were obtained at approximately i-second intervals. At the end of the infusion, the rats were killed by transection of the abdominal aorta. The sciatic-posterior tibial nerves were rapidly dissected free and frozen in liquid nitrogen. Autoradiographs were made from 20-km-thick frozen nerve cross-sections at five levels along the course of the nerves. Sections were taken 2 cm and 1 cm above, at the origin of the posterior tibial branch, and 1 cm and 2 cm below that landmark. The autoradiographs were analyzed by using a video processing system, and tissue '"C concentrations were calculated from autoradiographic densities calibrated to the standard curve. Endoneurial blood flow was calculated from the derived arterial I4C curve and the measured tissue '*C concentrations by using the equation of Kety and colleagues ClZ]. Because the A-V shunt may have affected perfusion of the contralateral nerve, regional blood flow values in ipsilateral sciatic nerves were compared with previously studied controls from this laboratory [l 11. Differences in regional blood flow along the course of ischemic and control nerves were compared by using a Wilcoxon rank sum test and the significance of differences between groups calculated with a nonpaired two-tailed t test.
Nerve Conduction Measurementf Rats were anesthetized with pentobarbital and placed in a prone position. Core temperature was maintained at 37°C with a rectal probe servocoupled to a heating blanket. Nearnerve temperature of operated and unoperated rats was approximately 34°C as monitored with a subcutaneous temperature probe. Sciatic motor nerve responses were studied by recording evoked intrinsic foot muscIe action potentials with bipolar subdermal electrodes after nerve stimulation at the ankle and the sciatic notch. Latency, amplitude, and duration of the evoked compound motor action potentials (CMAP) after proximal and distal stimulation were measured and sciatic motor conduction velocity (MNCV) calculated. The presence of dispersion or conduction block of the proximally evoked CMAP was assessed by comparing the amplitude, duration, and configuration of the responses. Sciatic mixed nerve conduction was measured with subdermal stimulating electrodes positioned adjacent to the posterior tibial nerve at the ankle and subdermal recording electrodes placed adjacent to the sciatic netve at the level of the sciatic notch. The latencies and amplitudes of the evoked nerve action potential (NAP) were measured and the mixed nerve conduction velocities calculated. Differences in conduction velocities and evoked potential amplitudes in nerves ipsilateral and contra-
Results Rats were clinically normal after recovering from surgery aside from 2 that had weakness of intrinsic foot muscles ipsilateral t o the A-V shunt for 24 hours after the procedure. Within 2 weeks after surgery, t h e shunted legs were visibly enlarged. Skin and subcutaneous tissues and t h e limbs and feet of all rats appeared normal at t h e time of electrophysiological and blood flow studies.
Meaurement of Endoneuriul Blood Flmu Endoneurial perfusion was normal in proximal sciatic nerves 6 weeks after creation of t h e shunt (12.1 5 4.4 m l / l 00 gmimin, control, 1I .g t 2.5 mll1OO gm/min) (Table 1). A zone of ischemia was detected distal t o this that began just proximal to t h e origin of the posterior tibial nerve. Endoneurial perfusion progressively declined in this ischemic z o n e and was lowest in the most distal nerve segments studied (3.6 t 1.7 ml/100 gmimin) (see Table 1).
Nerrie Conduction Measwements Sciatic nerve conduction velocities slowed t o ?4% of normal within 2 weeks after creation of the shunt and
Sladky et al: Chronic Ischemic Neuropathy
273
Table 1. Endoneurial Blood Flow along the Course of Sciatic-Posterior Tibzal N m e s 6 Weeks after a Proximal Femoral A-V Shunt
Site
Ischemic"
Controlb
2 cm Proximal 1 cm Proximal
12.1 2 4.4 8.7 k 4.0 8.2 2 4.4 4.4 2 2.2 3.6 1.7
11.9 t 2.5 9.7 ? 1.6 13.9 2 5.7 10.9 t 1.2
Origin
1 cm Distal 2 cm Distal
*
11.9
* 1.2
w-
P ,.
NS NS NS
< 0.001 < 0.001
Values (mean 2 1 SD) are expressed in (m11100 gmimin). 'Experimental Group 1 (n = 5).
0
1
1
3
4
5
6
Weeks after surgery ' in mlsec (mean
A-V = arteriovenous; NS = not significant.
+
I S E1
Fig 1. Evoltltion of slouiing of rat .rciatic motor nerve conduction velocity during experimental chronic ischemia (Group I. n = 5). Conduction decreared t o 74% of control values within 2 weeks of i.rchemia and remuined slow ocer the 6-week period of observation. Error ban are 1 SEM.
remained slow over the next 4 weeks (Fig 1).At postoperative week 6 , sciatic MNCV ipsilateral to the fistula was 32.6 2 4.3 m/sec (mean -t SD) compared with 47.8 6.2 m/sec in the contralateral nerve ( p < 0.005) (Table 2). The amplitudes of the evoked CMAPs were not significantly diminished in cxperimental nerves, and there was no evidence of dispersion of the proximally evoked responses, nor was conduction block evident (Table 3). Sciatic mixed nerve conduction velocity was similarly reduced to 32.8 k 8.6 mlsec on the ischemic side compared with 54.7 -+ 4.2 m/sec on the unoperated side (a < 0.003). The amplitudes of the sensory NAPS were similar on the control and ischemic sides without evidence of dispersion (see Table 3). Sciatic motor and mixed nerve conduction measurements in sham-operated rats were symmetrical and not different from those of unoperated nerves contralateral to the A-V shunt in experimental rats (see Table 2). Among the rats followed for up to 10 months after creation of the shunts (n = 4), the initial fall in nerve conduction velocities was similar to that in the shortcrterm rats. MNCVs were 34.9 k 2.8 after 6 weeks. In 2 rats studied, the disparity between sciatic MNCV in shunted compared with contralateral limbs persisted after 5 months (mean, 39 vs. 55 m/sec, respectively). This side-to-side difference remained after 10 months in 3 rats tested (mean, 33 ipsilateral vs. 48 misec con-
*
tralateral). One rat died from an anesthetic complication before electrophysiological studies could be performed. Morphological Analyses Dilated shunt vessels were prominent at the site of surgery in the right groin, but the superficial sciatic epineural vessels were not visibly enlarged ipsilateral to the shunt. Examination of cross- and longitudinal sections obtained along the proximal-distal extent of sciatic nerves failed to show signs of axonal breakdown or demyelination (Fig 2). Myelinated fiber numbers and densities were similar on the operated and unoperated sides at both prommal and distal levels. Mean sciatic axonal areas from segments taken 1 cm proximal to the origin of the posterior tibial branch were similar from nerves ipsilateral and contralateral to the shunt (47.8 t 33.9 vs. 48.1 34.2, respectively). Within the distal posterior tibial nerve at a point 2 cm below its origin, however, there was a left shift in the sizefrequency spectrum of axon profiles toward smaller diameter fibers; the mean axonal area was mildly de-
*
Table 2.Sciatic Nerve Conduction Velocities a&er 6 Weeh of Ischemia
Sham-operated Rats (n = 5 )
Shunted Rats (n = 10) Ipsilateral
32.6 32.8
Motor NCV (m/sec) Mixed NCV (m/sec) Values are mean
?
* 4.3" * 8.6'
Contralateral
49.8
* 6.2
54.7
?
4.2
1 SD.
"Significantly different from contralateral nerve and from sham-operated rats (p < 0.001).
NCV
=
nerve conduction velocity.
274 Annals of Neurology
Vol 29
N o 3 March 1991
Ipsilateral
*
4.2 49.6 58.4 2 7.8
Contralateral
48.8 57.7
k ?
4.9 9.1
Table 3. Sciatic Evoked Potential AmpLitudes after 6 Weeks of Ischemia Experimental Group (n
Proximal CMAP (mV) Distal CMAP (mV) Proximal:distal ratio Mixed NAP amplitude CpV)
=
Ischemic
Contralateral
12.0 2 5.1
10.3
13.2
?
5.7
0.92 t 0.04 24.4 2 11.4
10)
p
3.8
NS
11.8 t 4.7
NS NS NS
0.89 25.8
Values are mean 2 1 SD. CMAP = compound motor action potential; NAP potential; NS = not significant.
2
* 0.07 f
=
11.1
nerve action
creased to 42.8 +- 24.3 versus 48.4 k 24.3 Fm2 on the unoperated side (mean ? 1 SD, p < 0.001) (see Figs 2, 3; Table 4). The incidence of axonal degeneration and segmental demyelination was not increased in teased fibers after 6 weeks of ischemia compared with unoperated contralateral nerves. Nodal morphology was abnormal within the ischemic nerves, however, especially in posterior tibial nerve segments between 1 and 2 cm distal to the origin. Juxtanodal portions of axons frequently were swollen and paranodal myelin was dysmorphic, with evidence of degeneration and remodeling, and there was symmetrical and asymmetrical paranodal myelin retraction with widened nodal gaps (Fig 4). These features were not present in similarly handled contralateral or sham-operated nerves. Electron microscopy of nodes within ischemic nerve segments showed simplification and degeneration of paranodal Schwann cell foot processes, with jutanodal myelin degeneration and nodal widening (Fig 5). With the exceptions of paranodal swelling and occasional herniation of axoplasm at nodes of Ranvier, axons appeared normal.
Discussion Our results demonstrate that chronic nerve ischemia can be produced experimentally despite peripheral nerve’s capacity to establish collateral circulation within minutes after occlusion of nutrient vessels. A state of chronic endoneurial ischemia was created by shunting or “stealing” blood away from the distal leg through a high-flow, low-resistance proximal A-V shunt. This is similar to the phenomenon that may occur in patients with fistulae created surgically to facilitate hemodialysis [13-151. Some of these individuals develop a monomelic ischemic neuropathy with arm pain and weakness distal to the shunt. The majority of reported patients also were diabetic and uremic, however, and therefore had preexisting metabolic and vascular disease that may have predisposed to neuropathy in an ischemic environmen t. In this study, moderately severe, noninfarctive chronic ischemia led to slowing of conduction along
Fig 2. Cross-section at the lad ofthe midposterior tibial nerve afier G uleeh of ischemia. There i.r occasional crenation of myelin profiles mgge.rtiny mild axonal atrophy. hut neither axonal degeneration nor demyelination are present. (originalmagniJication, x 410.) Despite its relathJelynormal appearance, sciatic motor neme condtlction velocity w ~ 15 s misec in this nerve, approximately 70% of controls.
motor and sensory axons. The pathophysiological basis for this slowing is not known. Although segmental demyelination is commonly invoked to explain conduction slowing of this degree, morphological analysis did not detect evidence of demyelination along affected nerves. Loss of large diameter, fast-conducting axons could slow conduction, but axonal degeneration was not evident. Block of the fastest conducting axons could explain NCV slowing. This is an unlikely explanation because CMAP amplitudes, CMAP proximal: distal amplitude ratios, and sciatic mixed nerve NAP amplitudes were similar in ischemic and control nerves. Axonal atrophy may also be associated with slowing of conduction velocity in a number of neuropathies. The degree of slowing that we observed in this model is difficult to ascribe to mild axonal atrophy that is confined to distal axon segments (see Figs 2, 3). A comparable degree of axonal atrophy from experimental protein-calorie deprivation in rats had no effect on motor or mixed sciatic nerve conduction velocities {16}. We speculate that conduction slowing was the consequence of ischemia-induced alterations at nodes of Ranvier, with increased nodal capacitance and slowing of saltatory conduction. Pathological changes in Schwann cell-axon relations at the nodes of Ranvier were striking within the ischemic zone. We found degeneration of Schwann cell foot processes and paranodal myelin with retraction of juxtanodal myelin and widening of nodal gaps. Previous reports have documented slowing of nerve conduction in association with paranodal demyelination { 171, and modeling studies predict that the most striking effects on saltatory conduction are related to changes in membrane capaciSladky et al: Chronic Ischemic Neuropathy
275
Ischemic Posterior Tbiol Nerve
Contmloteral Posterior Tibia1 Nerve
-..- pi 1
lo
0
m
I
7 6
5 4
3 2
1 0 0
Cross-sectional Axonal Area in square microns
10
20
50
40
70
UJ
W 100 110
la0 150 140 150
Cross-sectional Axonal Area in square microns
Contralateml Mid-Sciatic Nerve
la 91
n
1
c
c D
Q
o
20 s0 u1 50 (YI 70 w 90 im 110 120 im 140 Cross-sectional Axonol Area in square microns
10
150
Fig 3. Morphometric analysis of axon cross-sectionalureu pr0JEle.s at the leziel of the midsciatic und distal posterior tibia1 nerves obtainedfrom paived newes ip..tilatrraland contralateralto a femoral avterioven0u.r shunt. Data is pooled from 3 ruts.
tance at the juxtanodal axolemma [18). Similar nodal alterations have been reported in human and experimental diabetic neuropathy 119-2 1) and after experimental pyridoxine intoxication [22}. In both settings, slowing of nerve conduction occurs in parallel with pathological changes at nodes {23-251. In experimental diabetes, abnormalities in sodium conductance have been measured at the nodal axolemma {23}. Nerve blood flow may be transiently reduced after vascular occlusion to just 20 to 25% of normal without apparent functional or pathological consequences Ill}. Peripheral nerve lacks the capacity to autoregulate blood flow [24},but ischemic nerve may survive acute feeding vessel occlusion by developing enough collateral circulation to subsist with compensatory changes in glucose metabolism to maintain bioenergetic homeostasis [25}. In the present study, flow was diminished to a similar degree but maintained at this lower level. Within a few weeks, there was conduction slowing, nodal injury, and axonal atrophy. Functional and
276
Annals of Neurology
Vol 29
No 3
March 1991
0 o
20 m u) 50 M 70 w 90 im i i o 120 is0 iu) is Cross-sectional Axonal Area in square microns
1100
structural abnormalities persisted without appreciable change for up to 10 months. This failure to compensate may result from circulatory factors related to the mechanism for producing ischemia. An A-V shunt causes both diminished perfusion pressure in the endoneurial capillary bed and increased venous resistance. Both factors can diminish endoneurial perfusion {4,26f and, in concert, retard the reestablishment of sufficient collateral circulation to permit normal aerobic metabolism. It is not surprising that chronic endoneurial ischemia led to slowing of conduction and distal axonal atrophy. Systems in the peripheral nerve with the highest energy requirements are thought to be especially vulnerable to hypoxia. This includes the nodal transmembrane ion exchange, funclamental to axon potential generation and axoplasmic flow [27} necessary for the transport of filaments, the major determinant of axonal caliber [28]. It is likely that Schwann cell ischemia had deleterious consequences on myelin maintenance. This may prove to be the mechanism for paranodal myelin abnormalities in chronic ischemia. We conclude from these experimental observations that it is possible to produce chronic experimental nerve ischemia in the absence of concomitant toxic or
Table 4. Morphometric Analysis of Myelinated Fibers after 6 Weeks of Ischemia" Axonal Circumference (pm2) Left sciatic (n = 6,097) Right sciatic (n = 5,722) Left posterior tibial (n = 4,558) Right posterior tibial (n = 3,791) Values are mean
k
26.7 26.6 24.4 26.6
* 10.4 * 10.4 ? ?
7.9b 8.8
Axonal Area (pm2)
47.8 48.1 42.8 48.4
* 33.9 ? ?
*
34.2 24.3b 29.2
Circumference/Area Ratio
0.84 L 0.81 1.1 0.85 0.73 2 0.34b 0.69 t 0.33
*
1 SD.
'Paired left (ischemic) and right (contralateral) sciatic nerves (1 cm proximal to the origin of the posterior tibial) and posterior tibial nerves (2 cm distal to the origin). Morphometric data pooled from 3 rats. bSignificantlydifferent from contralateral (j < 0.001).
Fig 4. Longitudinal section of rat posterior tibial nerve after 6 weeekr of ischemia. Nodal appearance rangesfrom normal (arrowhead) to markedly abnormal (arrow), with widening andparanodal swelling. A n equivocally widened node is also present (asterisk). (original magnification, x 650.)
metabolic insults by diverting the blood supply to the distal limb through a proximal arteriovenous shunt. In this model, chronic sublethal nerve ischemia slowed nerve conduction velocities to 50 to 75% of normal. There were alterations in nodal morphology and axon caliber but not segmental demyelination or axonal degeneration. These physiological and pathological consequences of chronic nerve ischemia are quite distinct from those of acute nerve ischemia where focal conduction block or nerve infarction are encountered ( 8 , 10, 29, 30). Future studies are necessary to determine whether the electrophysiological and morphological abnormalities that result from isolated chronic mild ischemia progress to demyelination, axonal degeneration, and clinical evidence of peripheral nerve dysfunction. This work was supported by CIDA Grant NS-01077 (to J. T. S.), the Muscular Dystrophy Association, and National Institutes of Health Grant NS-08075.
We thank Doreen Cashen for expert technical assistance and James Goin, PhD, for advice regarding statistical analyses.
References
Fig 5 . Electron micrograph of a node of Ranvier from the posterior tibial nerve after 6 weeks of chronic ischemia. There is simplz5cation and remodeling of Schwann cell foot processes with loss of Schwann cell-axolemma tight junctions, myelin retraction, and nodal widening. (original magnification, x 4,800.)
1. Eames RA, Iange LS. Clinical and pathological study of ischaemic neuropathy. J Neurol Neurosurg Psychiatry 1967;30: 215-226 2. Dyck PJ, Conn DL, Okazaki H . Necrotizing angiopathic neuropathy. Mayo Clin Proc 1972;47:461-475 3. Dyck PJ, Karnes JL, OBrien P, et al. The spatial distribution of fiber loss in diabetic polyneuropathy suggests ischemia. Ann Neurol 1986;19:450-45 7 4. Tuck RR, Schmelzer JD, Low PA. Endoneurial blood flow and oxygen tension in the sciatic nerve of rats with experimental diabetic neuropathy. Brain 1984;107:935-950 5. Low PA, Nukada H, Schmelzer JD, et al. Endoneurial oxygen tension and radial topography in nerve edema. Brain Res 1985;341:147-154 6. Meyers RR, Mizisin AP, Powell HC, Lampert WP. Reduced nerve blood flow in hexachlorophene neuropathy. Relationship to elevated endoneurial fluid pressure. J Neuropathol Exp Neurol 1982;41:391-399 7. Meyers RR, Powell HC. Galactose neuropathy: impact of chronic endoneurid edema on nerve blood flow. Ann Neurol 1984;16:587-594
Sladky et al: Chronic Ischemic Neuropathy 277
8. Parry GJ, Brown MJ. Arachidonate-induced experimental nerve infarction. J Neuroi Sci 1981;50:123-133 9. Korthals JK, Wisneiwski HM. Peripheral nerve ischemia: part 1. Experimental model. J Neurol Sci 1975;24:65-76 10. Nukada H, Dyck PJ. Microsphere embolization of nerve capillaries and fiber degeneration. Am J Pathol 1984;115:275-287 11. Sladky JT, Greenberg JH, Brown MJ. Regional perfusion in normal and ischemic rat sciatic nerve. Ann Neurol 1985: 17:191-195 12. Sakurada 0, Kennedy C, Jehle J, et al. Measurement of local cerebral blood flow with iodo-[14-C) antipyrine. Am J Physiol 1978;234:H59-H66 13. Wilbourn AJ, Furlan AJ, Hulley W, Ruschhaup TW. Ischemic monomelic neuropathy. Neurology 1983;33:447-45 1 14. Knezevic W, Mastaglia FL. Neuropathy associated with BresicaCimino arteriovenous fistulas. Arch Neurol 1984;41:11841192 15. Rqgs JE, Moss AH, Labosky DA, et al. Upper extremity ischemic monomelic neuropathy: a complication of vascular access procedures in uremic diabetic patients. Neurology 1989;39: 997-998 16. Cornblath D, Brown MJ. Influence of malnutrition on developing rat peripheral nerves. Exp Neurol 1988;99:403-411 17. LaFontaine S, Rasminsky M, Saida T, Sumner AJ. Conduction block in rat myelinated fibers following acute exposure to antigalactocerebroside serum. J Physiol 1982;323:287-306 18. Koles ZJ, Rasminsky M. A computer simulation of conduction in demyelinated nerve fibers. J Physiol 1972;227:351-364 19. Sima AAF, Bril V, Nathaniel V, et al. Regeneration and repair of myelinated fibers in sural-nerve biopsy specimens from patients with diabetic neuropathy treated with Sorbinil. N Engl J Med 1988;319:548-5 5 5 20. Sima AAF, Lattimer SA, Yagihashi S, Green DA. Axo-glial dys-
278 Annals of Neurology Vol 29 No 3 March 1991
junction: a novel structural lesion that accounts for poorly reversible slowing of nerve conduction in the spontaneously diabetic rat. J Clin Invest 1986;77:474-484 21. Brismar T, Sima AAF, Green DA. Reversible and irreversible nodal dysfunction in diabetic neuropathy. Ann Neurol 1987; 21:504-507 22. Sladky JT, Xu Y, Brown MJ. Transient ataxia after acute pyridoxine intoxication: electrophysiological and neuropathological correlations. Neurology 1987;37(suppl 1):376 23. Brismar T, Sima AAF. Changes in the nodal function in nerve fibers of the spontaneously diabetic BB-Wistar rat: potential clamp analysis. Acta Physiol Scand 1981;113:499-506 24. Low PA, Tuck RR. Effects of changes of blood pressure, respiratory acidosis, and hypoxia on blood flow in the sciatic nerve of the rat. J Physiol 1984;347:513-524 25. SladkyJT, GreenbergJH, Brown MJ. Enhanced 2-deoxyglucose incorporation in peripheral nerve during ischemia. Brain Res 1987;414:323-329 26. Simpson LO. Altered blood rheology in the pathogenesis of diabetic and other neuropathies. Muscle Nerve 1988;ll: 725-744 27. Hoffman PN, Griffin JW, Price DL. Control of axon caliber by. neurofilament transport. J Cell Biol 1984;99:705-714 28. Hoffman PN, Thompson GW, Griffin JW, Price DL. Changes in neurofilament transport coincide temporally with alterations in the caliber of axons in regenerating motor fibers. J Cell Biol 1985;101:1332-1340 29. Fowler CJ, Gilliat RW. Conduction velocity and conduction block after experimental ischemic nerve injury. J Neurol Sci 1981;52:221-228 30. Parry GJ, Linn DJ. Conduction block without demyelination following acute nerve infarction. J Neurol Sci 1988;84: 265-273
