Peripheral Neuropathy in RatsInduced by Insulin TreatmentPER SIDENIUS AND JOHANNES JAKOBSEN

SUMMARYThe effect of sustained insulin-induced hypoglycemiaon peripheral nerve function and structure was exam-ined in rats. After a period of hypoglycemia (<2.5mmol/L) of at least 72 h, axonal degeneration and re-duction of the maximal amplitude of the evoked mus-cle action potential occurred, the two abnormalitiesbeing correlated negatively (r = -0.99, 2P = 0.00097).One of five rats developed paresis of both hindlegs aswell as nerve damage and perikaryal alterations oflower motor neurons. DIABETES 32:383-386, April 1983.

Since peripheral nerve metabolism is dependent onglucose utilization, a potential cause of peripheralneuropathy is sustained periods of hypoglycemia.1

Nonetheless, very little is known about hypogly-cemic nerve damage. A few case reports describe the oc-currence of paresis in patients with insulinomas, but onlyscanty information on histologic and neurophysiologic ab-normalities in such patients is available.2"5 Jaspan et al.described axonal degeneration in a sural biopsy from onepatient, signs of denervation at electromyography in all ofthe patients, and reduced motor nerve conduction velocityin three of four patients.45 However, there are no reports inthe literature on hypoglycemic damage of peripheral nervesinduced by insulin treatment.

Whether sustained periods of hypoglycemia exert a del-eterious effect on peripheral nerves has been an increasinglyimportant question after the insulin pump has been intro-duced in the treatment of diabetes. We are currently studyingthis question in an animal model and have observed that afew days of profound hypoglycemia produce classic axonaldegeneration (unpublished observation). We now for the firsttime report on the occurrence of muscle weakness and struc-

From the Second Clinic of Internal Medicine, Department of Neurology, andInstitute of Pathology, University of Aarhus, DK-8000 Aarhus C, Denmark.Address reprint requests to Per Sidenius, Institute of Anatomy, University ofAarhus, DK-8000 Aarhus C, Denmark.Received for publication 13 January 1983.

tural damage of peripheral nerve fibers and their cell bodiesinduced by insulin treatment.

MATERIALS AND METHODSFive 60-75-wk-old male Wistar rats from our own inbredcolony were treated once daily with a modified and very long-acting insulin (Ultralente, pH 5.5, NOVO, Denmark) during6-7 days and had free access to food and water. Six toseven days after the last dose of insulin was given, muscleaction potentials of the anterior tibial muscle on the right sidewere registered and subsequently the left sciatic nerve wasremoved for quantification of nerve fiber degeneration. Oneanimal developed a severe paresis of both hind limbs, andtissues from the lumbar enlargement of the spinal cord aswell as from the tibial nerve were taken out after whole bodyperfusion through the heart with a 2% glutaraldehyde fixative(0.1 M cacodylate, pH 7.35).Insulin treatment. To study the effect of one daily injectionof the applied insulin preparation on the variation of bloodglucose during a 24-h period in hypoglycemic animals, bloodglucose was measured every 2-3 h in an additional groupof seven rats. At the start of the experiment the mean bloodglucose value was 2.3 mmol/L (range: 1.7-2.8-mmol/L); atthe end it was 2.2 mmol/L (range: 2.0-2.5 mmol/L). Sinceblood glucose levels remained constantly depressed duringthe 24-h period, measurements in the experimental groupwere performed only once daily. Blood samples were takenfrom the tail veins at 4 p.m. and glucose levels were deter-mined with a reflectance meter (Glucometer, Ames, Elkhart,Indiana). Insulin (200-300 \L\, 1 IU/25 \L\) was dosed ac-cording to blood glucose determinations measured imme-diately before the subcutaneous injection. In all animals a72-h experimental period of severe hypoglycemia (<2.5 mmol/L) was intended.Muscle action potentials. To quantify the extent of motorweakness, evoked maximal compound muscle action po-tentials were registered. The rats were anesthetized withpentobarbitate, and a bare needle electrode (cathode) wasinserted into the vicinity of the sciatic nerve near the greatertrochanter of the femur. A second needle (anode) was placed

DIABETES, VOL. 32, APRIL 1983 383

PERIPHERAL NEUROPATHY IN RATS INDUCED BY INSULIN TREATMENT

mmol/ i

16 H

4 -

1 -10 DAYS

FIGURE 1. Daily blood glucose values in five rats given insulin for6-7 days. In rat no. 1 (solid line) glucose values lower than 1.4 mmol/Lfor a 72-h period were obtained, whereas the lowest values for theother animals varied around 2 mmol/L.

subcutaneously on the anterior thigh. A supramaximal stim-ulus was delivered to the sciatic nerve and the compoundmuscle action potential evoked was recorded with a needleplaced subcutaneously under the loose skin transverselyover the anterior tibial muscle with a reference electrode atthe ankle. The nerve was stimulated at 0.5/s while the re-cording electrode was moved over the surface of the anteriortibial muscle to locate the maximal amplitude of the musclepotential. The maximal response (peak to peak) was pho-tographed and measured. The maximal muscle action po-tential is widely independent of age. Values obtained in ourlaboratory for 15-30-wk-old rats, therefore, are given as ref-erence. The animals were examined by one of the authorswho was unaware of the identity of the animals.Histology. Immediately after the animals were killed, 1£ cmof the sciatic nerve proximal to the knee was fixed for45 min in a 2% glutaraldehyde solution followed by osmifi-cation for 3-4 h in 1% OsO4. Subsequently, the nerves weremacerated in glycerine for 2 days. Under stereomicroscopea minimum of 100 fibers from each rat were teased consec-utively by means of two pairs of watchmakers' forceps. Clas-sification of fiber abnormalities was done blindly and in ac-cordance with Dyck's description.6 Teased fiber preparations

from the sciatic nerves of a group of 79-83-wk-old maleWistar rats from our own colony were used as controls in thepresent experiment.Statistics. Values are given in mean ± SD.

RESULTSDuring a 24-h period of hypoglycemia blood glucose valuesremained remarkably constant, the values for the individualrats being 2.0 ± 0.2,2.0 ± 0.3,2.1 ± 0.2,2.3 ± 0.3,2.4 ± 0.4,2.4 ± 0.3, and 2.5 ± 0.2 mmol/L; the ranges were 1.7-2.2,1.7-2.5,1.8-2.4,1.9-2.6,1.8-3.0,1.9-2.8, and 2.1-2.8 mmol/L, respectively.

The table gives the geometric mean blood glucose valuesfor the hypoglycemic period for each rat. Figure 1 illustratesthe blood glucose values for the first 10 days of the exper-iment. It appears that blood glucose levels in rat no. 1 variedbetween 1.0 and 1.4 mmol/L for a 72-h hypoglycemic period,whereas nadirs close to 2.0 mmol/L were obtained in theother four animals.

The rat with the most profound hypoglycemia (no. 1) de-veloped paresis of both hindlegs during the experimentalperiod, starting at the nadir of hypoglycemia. The muscleweakness was most pronounced distally with absence ofdorsiflexion of the feet during active movements. Knee flex-ors and extensors were also weak, whereas the movementsof thighs and forelimbs maintained their normal strength.Furthermore, two rats (nos. 2 and 3) developed a charac-teristic grunting 7 days after the first dose of insulin, butdisplayed no conspicuous signs of peripheral neuropathy.

The maximal muscle action potential of the anterior tibialmuscle is shown for each rat in Table 1. Two rats (nos. 1and 2) had reduced muscle action potentials (>2 SD fromthe control value), and the lowest value was obtained in therat with hindleg paresis.

Figure 2 shows the type of abnormalities found in teasedfiber preparations. It appears from Table 1 that axonal de-generation was present in all test animals, and that threerats (nos. 1, 2, and 3) had an increased number of degen-erated fibers (>2 SD from the control value). There was anegative statistically significant correlation between the num-ber of degenerated fibers and the logarithm of the amplitudeof the muscle action potential (Figure 3).

TABLE 1Age, weight at start and end of the experiment, total insulin dosage, blood glucose at start and end of the experiment, episodes ofhypoglycemia, evoked muscle action potential, amount of axonal degeneration, and demyelination for each individual rat of theexperiment as well as reference values for controls

Parameter

AgeWeight at startWeight at endTotal insulin dosageBlood glucose at startBlood glucose at endEpisodes of hypoglycemia (<2.5 mmol/L)

1st episode, geometric mean (duration in days)2nd episode, geometric mean (duration in days)

Evoked muscle action potentialAxonal degenerationDemyelination

Unit

wkggIU

mmol/Lmmol/L

mmol/Lmmol/L

mV%%

1

7549541049

4.96.4

1.4(4)

773.30

2

60515490

585.34.1

2.2 (2)2.0(1)

3316.80

Animal

3

60505435

474.74.2

2.0 (4)

5111.50

4

60505485

894.95.1

2.2 (5)

712.90

5

75545540

524.05.1

2.2 (2)

631.71.7

PpfprDnro wall i^ncioici IOC vciiuc

(mean ± SD)

470 ± 31

5.6 ± 0.4

60 ± 50.9 ±2.10.4 ± 0.7

384 DIABETES, VOL. 32, APRIL 1983

PER SIDENIUS AND JOHANNES JAKOBSEN

200 jumFIGURE 2. The type of changes observed in teased fiber preparationof the sciatic nerve 6-7 days after a 3-5-day period of hypoglycemia(<2.5 mmol/L). A normal fiber with a Ranver node in the middle isseen at top (1). Fiber 2 shows early signs of axonal degeneration withseparation of small myelin segments, and fibers 3 and 4 are laterstages in the same axonal process.

100/um

FIGURE 4. Cross-section of the tibia! nerve from the rat who devel-oped hindleg weakness during a 72-h period of profound hypoglyce-mia. Many dark or otherwise abnormal fiber profiles (about half of thepopulation) indicating axonal degeneration are seen. The insert showsa few degenerated fibers from the same section at a higher magnifica-tion.

Figure 4 shows a cross-section of the tibial nerve from thedisabled rat. Many abnormal profiles suggestive of axonaldegeneration were present. In the same rat the motor neu-rons of the lumbar enlargement of the spinal cord displayeduniform alterations. Figure 5 shows the type of abnormalitiesobserved. The chromophilic substance (rough endoplasmicreticulum) was retracted from the periphery of the perikaryonand concentrated centrally. No obvious changes of the nu-cleus or the nucleolus were recognized.

DISCUSSIONThe present article on the effect of sustained hypoglycemiaon peripheral nerve has shown that insulin treatment in ratscan cause muscle weakness, axonal degeneration, and lowermotor neuron changes. The fact that hindleg paresis was

mV

10-

1 -

r =-0.992p = 0.00097

0 50 100 %

FRACTION OF DEGENERATED NERVE FIBERS

FIGURE 3. The maximal amplitude (mV) of the evoked action potentialof the anterior tibial muscle in relation to the number (%) of teased fi-bers from the sciatic nerve showing axonal degeneration.

FIGURE 5. Lower motor neurons of the anterior horn of the lumbarenlargement from the rat who developed hindleg weakness during a72-h period of profound hypoglycemia (2 and 3) and from a control (1).The chromophilic substance of the perikaryon (Nissl substance orrough endoplasmic reticulum) concentrates around the nucleus andwithdraws from the periphery. A cell with severe changes (2) and agroup with typical alterations (3) are shown.

DIABETES, VOL. 32, APRIL 1983 385

observed in the rat with the most severe depression of bloodglucose levels and developed during the hypoglycemic pe-riod makes it likely that this alteration is responsible for thenerve damage. Furthermore, we have never observed spon-taneously occurring muscle weakness in our colony of ratsinbred for more than 20 yr.

The distribution of muscle weakness in the disabled rat istypical of distal axonopathy. Recently, Sterman has ob-served a chromatolytic response of dorsal root ganglion cellsof rats during advanced distal axonopathy induced by 2,5-hexanedione.7 However, that this disease entity can be as-sociated with specific structural changes of the nerve cellbodies is a new observation of the present article.

Similar changes of nerve cell bodies have been observedby Agardh et al. in neurons of the cerebral cortices of ratsexposed to profound hypoglycemia with isoelectric EEG for1/2-1 h.89 The changes were considered characteristic forthe condition since they are unlike any other reaction of thenerve cell to injury. The structural alteration was visible im-mediately after the hypoglycemic period and was fully re-versible after 1 h of normoglycemia. In contrast, the presentstudy demonstrated the somal alteration after 1-wk recovery.

Further studies in our laboratory have shown that axonaldegeneration occurs to the same degree in the purely sen-sory sural nerve as in the mixed sensory motor tibial nerve(Jakobsen and Sidenius, to be published).

To elucidate whether insulin-induced neuropathy mightcontribute to the peripheral nerve damage present in long-term diabetics more studies on the relation between axonal

degeneration and duration and severity of hypoglycemia areneeded.

ACKNOWLEDGMENTSWe are grateful to Anette Larsen and Birtha Saugbjerg forskillful technical assistance and to Karin Wiedemann for carefulpreparation of the manuscript. This work was supported byNOVO.

REFERENCES1 Greene, D. A., and Winegrad, A. I.: Effects of acute experimental

diabetes on composite energy metabolism in peripheral nerve axons andSchwann cells. Diabetes 1981; 30:967-74.

2 Tom, M. I., and Richardson, J. C: Hypoglycaemia from islet cell tumorof pancreas with amyotrophy and cerebrospinal nerve cell changes. J. Neu-ropathol. Exp. Neural. 1951; 10:57-66.

3 Mulder, D. W., Bastron, J. A., and Lambert, E. H.: Hyperinsulin neu-ronopathy. Neurology 1956; 6:627-35.

4 Danta, G.: Hypoglycemic peripheral neuropathy. Arch. Neurol. 1969;21:121-32.

5Jaspan, J. B., Wollman, R. L, Bernstein, L, and Rubenstein, A. H.:Hypoglycemic peripheral neuropathy in association with insulinoma: impli-cation of glucopenia rather than hyperinsulinism. Medicine 1982; 61:33-44.

6 Dyck, P. J.: Pathologic alterations of the peripheral nervous systemof man. In Peripheral Neuropathy. Dyck, P. J., Thomas, P. K., and Lambert,E. H., Eds. Philadelphia, Saunders, 1975:296-336.

7 Sterman, A.: Cell body remodeling during dying-back axonopathy:DRG changes during advanced disease. J. Neuropathol. Exp. Neurol. 1982;41:400-11.

8 Agardh, C.-D., Kalimo, H., Olsson, Y., and Siesjo, B. K.: Hypoglycemicbrain injury. I. Metabolic and light microscopic findings in rat cerebral cortexduring profound insulin-induced hypoglycemia and in the recovery periodfollowing glucose administration. Acta Neuropathol. 1980; 50:31-41.

9 Kalimo, H., Agardh, C.-D., Olsson, Y, and Siesjo, B. K.: Hypoglycemicbrain injury. II. Electron-microscopic findings in rat cerebral cortical neuronsduring profound insulin-induced hypoglycemia and in the recovery periodfollowing glucose administration. Acta Neuropathol. 1980; 50:43-52.

386 DIABETES, VOL. 32, APRIL 1983