Acta Diabetol (2004) 41:158–166

DOI 10.1007/s00592-004-0160-0

ORIGINAL

I. Mišur • K. Žarković • A. Barada • L. Batelja • Z. Miličević • Z. Turk

Advanced glycation endproducts in peripheral nerve in type 2 diabetes

with neuropathy

Received: 11 March 2003 / Accepted in revised form: 20 January 2004

Abstract Advanced glycation endproducts (AGE) accumu- nerves but not in the control specimen. The intensity of
late over proteins as a consequence of diabetic hyper- axonal AGE immunopositivity significantly correlated with
glycemia, and thus contribute to the pathogenesis of diabet- the severity of morphological alterations (p<0.005). AGE
ic complications. To improve the understanding of the localization, demonstrated by immunohistochemical meth-
pathology of diabetic neuropathy, AGE accumulation was ods, was also present in the endoneurium, perineurium and
analyzed in sural and/or femoral nerves obtained under microvessels. Morphometric analysis of the diabetic periph-
spinal anesthesia from 8 type 2 diabetic patients with both eral nerve showed perineurial thickening (diabetic vs. con-
distal symmetrical polyneuropathy and proximal neuropa- trol, 15.5±4.9 vs. 6.6±2.1 µm, p<0.001), narrowing of the
thy. Pronounced AGE immunoreactivity was detected on microvessel lumina (66.6±50.5 vs. 579.5±38.4 x103 µm2,
axons and myelin sheaths in 90% of diabetic peripheral p<0.001) and significant reduction in the number of pre-
served axons (3.6±3 vs. 8.9±2.3 per 105 µm2 per area,
p<0.037). The sera of diabetic patients contained epitope(s)
of AGE structure and soluble immune complexes containing
AGE moiety. In conclusion, to the best of our knowledge,
this is the first study providing evidence for excessive AGE
formation on peripheral nerve components, primarily axons,
and a significantly higher level of circulating AGE-immune
complexes in patients with both distal diabetic polyneuropa-
thy and proximal neuropathy. Humoral immune mecha-
nisms, including the production of anti-AGE autoantibody,
may potentially be involved in the development of structur-
al abnormalities described in this report.

Key words Diabetic polyneuropathy • Advanced glycation

endproducts • Peripheral nerve • Immunohistochemistry

I. Mišur • A. Barada • Z. Miličević • Z. Turk ()

Clinic for Diabetes, Endocrinology and Metabolic Diseases
Vuk Vrhovac University Introduction
Dugi dol 4A, HR-10000 Zagreb, Croatia
E-mail: [email protected] Diabetic neuropathy is one of the most common long-term
K. Žarković complications of diabetes, affecting as many as 60% of
Department of Neuropathology the diabetic population. In spite of ample research, the
Zagreb University Hospital Centre pathogenesis of diabetic neuropathy has not yet been fully
Zagreb, Croatia understood. The proposed mechanisms include cellular
L. Batelja damage in the peripheral nervous system, altered neuronal
School of Medicine metabolism, microvascular abnormalities, slowing of
Zagreb University, Croatia axonal transport mechanisms, deficiency of neurotrophic
I. Mišur et al.: AGE in peripheral nerve in type 2 diabetes 159

substances, and impaired repair capabilities [1, 2]. tory of retinopathy (n=4), nephropathy (n=3) or neuropathy
Neuropathy has been described in patients with type 1 and (n=8). Simple or proliferative diabetic retinopathy was diagnosed
type 2 diabetes. The key role of hyperglycemia has ophthalmoscopically by an ophthalmologist, and graded accord-
received support from large prospective studies such as ing to the European protocol. The diagnosis of diabetic neuropa-
thy was made or ruled out by neurologic examination (subjective
DCCT and UKPDS [3–5]. The major metabolic changes
discomfort, clinical signs of disease, eletromyography [EMG]),
caused by hyperglycemia include an increased polyol
according to international classification. Renal impairment was
pathway flux, elevated oxygen free radical formation, and assessed on the basis of complete clinical examination. Normal
advanced glycation process [1, 6]. Although a number of renal function was defined as serum creatinine concentration
biochemical changes have been described in diabetic <120 µmol/l; urinary albumin excretion <30 mg/24 h, and creati-
nerve, the exact sequence of events leading from insulin nine clearance >0.90 ml/s. Microalbuminuria was considered at
deficiency and accompanying hyperglycemia to the func- values of ≥30 mg/24 h and ≤200 mg/g creatinine; macroalbumin-
tional and structural manifestations characterizing clinical uria was diagnosed at >200 mg/g creatinines, and overt protein-
neuropathy has not yet been fully elucidated. uria (>0.5 g/24 h) was defined as elevated serum creatinine (>130
Nonenzymatic glycation leading to advanced glyca- µmol/l) and creatinine clearance <0.80 ml/s. Macrovascular dis-
ease (n=4) was considered to be present if there was a history of
tion endproduct (AGE) formation is one of the important
myocardial infarction, coronary artery disease, stroke or periph-
biochemical pathways involved in the development of eral arterial disease.
long-term complications of diabetes [7], with a potential Specimens of sural and femoral nerve were obtained by biop-
role also in diabetic neuropathy [8–10]. Products of sy in spinal anaesthesia. Longitudinal and transverse 2- to 5- mm
advanced glycation are generated from series of reactions nerve specimens were embedded in tissue cryopreservative
involving the attachment of reducing sugars or oxoalde- (Cryomatrix) and frozen in liquid nitrogen. Thus prepared, the
hydes to protein [11]. The final reaction step, which is specimens were stored at -75° C until analysis. The Vuk Vrhovac
irreversible, gives rise to AGE products. AGEs are a het- University Clinic Ethics Committee approved the study protocol,
erogeneous group of molecules that accumulate in plasma and informed consent was obtained from each study patient
and tissues. These toxic macromolecules interact with before entering the study.
specific receptors and elicit pleiotropic response. There
are several possible pathways by which AGE formation
could be directly pathogenic. Extracellular AGEs induce Neurophysiology
cross-linking of proteins, modification of matrix compo-
nents, and may interfere with cellular adhesion and inter- Prior to biopsy, electrophysiological assessment of the lower
action. Intracellular AGE may induce alteration of DNA limb was performed with a Dantec Counterpoint EMG system
and nuclear proteins, and may alter protein transport and using surface electrodes. Nerve conduction velocity was meas-
function. ured proximally on the femoral nerve and distally on the sural
Extensive investigations of the advanced glycation nerve. Skin temperature was maintained over 32° C using a sur-
face heater.
process in human diabetic neuropathy are generally limit-
ed by analitical tissue unavailability. To improve the
understanding of the pathophysiology of diabetic neu-
ropathy, the human diabetic peripheral nerve was investi- Tissue sampling and histologic examination
gated by examining AGE modifications in situ. Biopsy
specimens of the sural and/or femoral nerve from type 2 Ten peripheral nerve cylindrical specimens, 3–5 mm in length
diabetic patients with distal symmetrical polyneuropathy and <1 mm in diameter, were analyzed. Ten peripheral nerve
and proximal neuropathy were examined. This report pro- specimens were obtained by biopsy of the sural and/or femoral
vides detailed immunohistologic description, however, nerves of type 2 diabetic patients. Control specimens were
with a limited body of relevant clinical data. obtained from a healthy, morphologically intact sural nerve from
a healthy individual, used as allograft to bypass a peripheral
nerve defect caused by gunshot wound; a part of the allograft was
referred for histologic analysis to assess the viability of its struc-
tures. Frozen transverse never sections previously made by a cry-
Patients and methods omicrotome were fixed in 10% formalin, then dehydrated in a
series of increasing ethanol concentrations (50%–100%), xylene,
The study included eight type 2 diabetic patients with both distal and embedded in paraffin blocks. The paraffin-embedded nerves
polyneuropathy and proximal neuropathy, mean age 62.1±6 were cut transversally, perpendicularly to longitudinal axis, into
years, diabetes duration of 1–15 years, HbA1c 7.9±1.2%, distal 5-µm sections. One section was stained with hematoxylin-eosin
polyneuropathy duration of 2.5 years, and proximal neuropathy (HE), and the other was immunohistochemically stained by use
duration of 4.2±1.8 months, pelvifemoral weakness MRC grade of PAP technique. Immunohistochemical staining of paraffin sec-
15.9±3.5/30, pain – visual analogue scale 7.8±1.6/10 above the tions with rabbit antibody against AGEs was done for more pre-
knee. Clinical data on the study patients are presented in Table 1. cise morphological analysis and positivity distribution in particu-
The presence of microangiopathy was defined as a positive his- lar parts of the nerve; precise analysis cannot be performed on
160 I. Mišur et al.: AGE in peripheral nerve in type 2 diabetes

frozen sections by the immunofluorescence method. Prior to were 7.90% and 7.11%, respectively. Anti-AGE antibodies in
applying it onto the tissue, primary rabbit polyclonal antibody human serum were determined by use of blocking ELISA [13]. The
was diluted in phosphate-buffered saline (PBS) at a 1:30 ratio, presence of immune complexes containing AGE moiety was deter-
and incubated for 45 min. Secondary porcine IgG to rabbit IgG, mined by two independent criteria: (a) serum AGE immune com-
PAP and diethylaminobenzidine (DAB; Dako, Denmark) were plexes (AGE-IC) were detected by ELISA using an immunochem-
used according to the manufacturer’s instructions. ical bridge; and (b) soluble AGE-IC were precipitated from serum
by polyethylene glycol and analysed as previously reported [13].
The within-run C.V. was 8.80% and the between-run was 10.72%.
The products of advanced glycation were measured in the serum of
Immunohistochemistry diabetic patients (n=8) as well as in the serum of non-diabetic
healthy subjects (n=20, control group).
Nerve specimens were cut to 4–8 µm sections on a cryostar
(Leica, Germany) at a temperature of -25° C to -30° C. These sec-
tions were applied on microplates previously overlayered with a
fixation membrane (PAP pen). Before testing, the sections were HbA1c measurement
allowed to dry in cold air stream and rehydrated in PBS (pH 7.4)
for 10 min. The sections were preincubated with proteinase K Samples were prepared according to Jeppsson et al. [14]:
(0.5 mg/ml) at room temperature to allow for primary AGE anti- heparinized blood (600 µl) was incubated for 4 h at 37° C with
body binding. After washing with PBS for 30 min (in three fresh saline and centrifuged. The supernatant was discarded and the
portions), the sections were incubated for 30 min with normal erythrocyte pellet was hemolysed by the addition of H2O and
rabbit serum (1:10) to inhibit nonspecific binding. After careful CCl4. Separation of cell ghosts was accomplished by centrifuga-
washing with PBS at least 3 times for 10 min, the sections were tion; the clear supernatant was diluted with malonic buffer and
incubated with primary antibody (polyclonal rabbit antibody analysed. HbA1c was analyzed using Fast protein liquid chro-
against AGEs) for 2 h at room temperature. Primary antibody was matography system (Pharmacia, Uppsala, Sweden), on a Mono S
obtained by rabbit immunisation and prepared for testing by seri- HR 5/5 column at pH 5.7, and expressed as percentage of the
al dilution at 1:5 to 1:80 with PBS, pH 7.4. After washing with total amount of hemoglobin with normal range of 3.5% to 5.7%.
PBS, secondary antibody was added for 30 min at room temper-
ature. Tetramethylrhodamine isothiocyanate (TRITC)- or fluores-
cein isothiocyanate (FITC)-labelled goat anti-rabbit IgG antibody
was used as secondary antibody. The labelled antibody was used Statistical analysis
at a 1:80 dilution for rhodamine or at 1:100 and 1:150 dilutions
for fluorescein. After washing, the sections were observed under All morphometric analyses and evaluation of intensity of AGE
fluorescence microscope (Opton-Zeiss) with appropriate filters immunoreactivity were performed in a blinded manner. The analysis
for rhodamine (IF 545) and fluorescein (IF 490). of AGE-immunopositivity was semiquantitatively evaluated by the
following symbols (-no AGE positivity; ±, borderline AGE positivi-
ty; +, weak AGE-positivity; ++, strong AGE positivity), as shown in
Table 3. The intensity of immunostaining signals was also numeri-
Analysis of advanced glycation products cally scored (-as 0; ±as 1; + as 2 and ++ as 3). Morphological alter-
ations were classified as mild (*=1), moderate (**=2) and severe
The antibodies referred to in our previous reports [12, 13] describ- (***=3). The scores of total morphological changes correlated with
ing the preparation of AGE antigen, production of antibodies the graded AGE-immunostaining intensity. Quantitative analysis of
against AGE, and specificity assessment of anti-AGE antibody peripheral nerves was performed on transverse sections (Mallory
against a particular antigen were used in the present study. The three-chrome staining) with ISSA and SFORM programs (VAMS,
competitive-type enzyme-linked immunosorbent assay with poly- Zagreb, Croatia). The statistical methods employed included Mann-
clonal anti-AGE antibodies was used for detection of AGE content Whitney test for comparison of two groups of data, and Spearman’s
in serum samples. The assay utilizes AGE-human serum albumin rank correlation coefficients for bivariate analysis. A value of p<0.05
(AGE-HSA) as a standard. The immunoplate was coated with was considered to be statistically significant.
AGE-HSA antigen. The serum containing an unknown quantity of
antigens was incubated together with a constant, known quantity of
polyclonal anti-AGE antibody. Solid-phase antigen-antibody com-
plexes were generated based upon the competition between serum
AGEs and immobilized AGE antigen. AGE antigen-antibody com- Results
plexes were detected using (secondary antibody) alkaline phos-
phatase-conjugated anti-rabbit IgG. A colorimetric signal, which We evaluated the presence of advanced glycation endprod-
was inversely proportional to the amount of AGEs in serum, is ucts (AGEs) in 8 patients with type 2 diabetes and diabet-
obtained following the addition of chromogenic substrate.
ic neuropathy (Table 1). The patients had a mean age of 62
Competitive immunoreactivity of the samples was read from the
calibration curve and expressed relative to AGE-albumin standard
years (SD=6 years) and mean HbA1c concentration of
in µgEq/ml. The precision of the method was determined in serum 7.9% (SD=1.2%). Mean pelvifemoral weakness score was
pools from diabetic patients and from normoglycemic control indi- 15.9 (SD=3.5) and mean pain score (VAS) was 7.8
viduals. The within-run coefficients of variability of AGE-ELISA (SD=1.6). The patients had distal symmetrical diabetic
I. Mišur et al.: AGE in peripheral nerve in type 2 diabetes 161

Table 1 Clinical characteristics of 8 patients with type 2 diabetes and diabetic neuropathy

Case Age, Sex DM Treatment HbA1c, Neuropathy Nephropathy Retinopathy Macrovascular

years duration, % diseasea
years

1 65 M 11 OHA 8.36 + – + –
2 65 M 8 OHA 7.18 + + – +
3 66 M 15 Insulin 8.57 + – + +
4 61 M 1 OHA 7.60 + – – –
5 57 M 13 Insulin 6.91 + + + –
6 64 F 10 Insulin 10.10 + – – +
7 50 F 3 Insulin 7.27 + – + –
8 69 F 11 Insulin 6.15 + + – +

DM, type 2 diabetes mellitus; HbA1c, hemoglobin A1c; OHA, oral hypoglycemic agents
a History of myocardial infarction, coronary artery disease or peripheral arterial disease

Table 2 Neurophysiological parameters of 8 patients with type 2 diabetes and diabetic neuropathy

Case PDN duration, months DSDP Nerve NCV, m/s

Duration, months Pain

1 6 48 No Sural 0.0
2 3 24 No Sural 35.8
Femoral 0.0
3 3 60 Yes Sural 0.0
Femoral 43.1
4 5 10 No Sural 40.2
5 5 36 Yes Sural 0.0
6 5 12 No Femoral 41.7
7 3 4 No Femoral 44.5
8 4 24 Yes Sural 0.0

PDN, proximal diabetic neuropathy; DSDP, distal symmetrical diabetic polyneuropathy; NCV, nerve conduction velocity (motor NCV in
femoral nerve, and sensory NCV in sural nerve)

Table 3 Semiquantitative analysis of advanced glycation endproduct (AGE) immunopositivity and morphological alteration in biopsied
diabetic peripheral nerves and control sural nerve

AGE immunopositivitya

Case Nerve Perineurium Endoneurium Myelin sheath Axons Blood vessels Morphological alteration

1 Sural – ± + + – Moderate
2 Femoral – – ± – – Mild
Sural ++ ++ ++ ++ + Severe
3 Sural – – – ± + Mild
Femoral + + + ++ + Moderate
4 Sural – + + + + Severe
5 Sural ± ± + ++ + Severe
6 Femoral + ± + – + Mild
7 Femoral – – + ± – Mild
8 Sural – ± ++ ++ + Moderate
Control Sural – – – – – None

a AGE immunopositivity scores: - none; ± borderline; + weak; ++ strong

162 I. Mišur et al.: AGE in peripheral nerve in type 2 diabetes

polyneuropathy for a mean of 2.5 years and proximal dia- Borderline or weak positivity in the endoneurium and per-
betic neuropathy for 3–6 months. Clinical picture was pre- ineurium was only present in one case with mild patho-
dominated by pronounced pain with hyperesthesia and logic alterations (Fig. 1).
dysesthesia in upper legs, atrophy of upper leg muscula- Moderate pathologic alterations were detected in 3 of
ture, loss of pelvifemoral muscle strength, and weight loss. 10 peripheral nerves from diabetic patients. These were
Neurophysiological parameters are presented in Table 2. characterized by a reduced number of nerve fibers and
manifest myelin degeneration in more than half of the fas-
cicles of each nerve sample, thickening of the perineurium
and endoneurium, and thickening and hyalinization of the
Nerve pathohistology blood vessels. Either weak or strong AGE positivity was
present in the myelin sheath and axonal region of all 3
Transverse sections of ten peripheral nerves (6 sural and 4 nerve samples (Table 3). AGE immunopositivity of bor-
femoral) from 8 type 2 diabetic patients showed similar derline or weak intensity was also present in the
morphological changes of varying severity (mild, moder- endoneurium. A weak vascular immunopositivity was
ate or severe; Table 3). Some 4–6 fascicles were observed observed in 2 samples, and in only one perineurium of
in all the nerves examined. In all diabetic patients, the per- these 3 nerves (Table 3).
ineurium and vascular walls were markedly thickened and In the remaining 3 peripheral nerve samples, all fasci-
hyalinated, with occasional endothelial proliferation and cles on transverse nerve sections were diffuse, with a
microvascular thrombosis. Proliferation of the endoneur- severe loss of nerve fibers and myelin degeneration. The
ial connective tissue and a reduced number of nerve fibers perineurium and endoneurium as well as vascular walls
were observed. Degeneration in the myelin sheath region were markedly thickened and frequently hyalinated. AGE
and absence of axons varied from mild through moderate immunopositivity was present in the myelin sheath, axon-
and severe. al region, endoneurium and blood vessels of all samples
Mild pathological alterations were observed in 4 of 10 (Table 3). The intensity of AGE positivity in the axons and
peripheral nerves from diabetic patients (Table 3). These myelin was strong or weak, while microvascular
showed focal reduction of nerve fibers, myelin degenera- immunopositivity was weak. In the endoneurium, AGE
tion in one or two fascicles, thickening of the perineurium immunopositivity ranged from borderline to strong, while
and endoneurium, and thickening and hyalinization of in the perineurium it was absent, bordeline or strong
blood vessels. Immunhistochemical staining for AGE in (Table 3, Fig. 2).
these samples yielded weak or borderline myelin sheath In 9 of 10 peripheral nerves from diabetic subjects,
positivity in 3 of 4 cases. In 2 cases, weak AGE positivity AGE immunopositivity was present in the myelin sheath
was also present in the axonal region and blood vessels. and axons. The intensity of axonal positivity significantly

Fig. 1 Transverse section of diabetic sural nerve (case 5) shows Fig. 2 Transverse sections of sural nerve (case 2) shows thickened
markedly thickened and hyalinated perineurium with no AGE and hyalinated perineurium with strong AGE-immunopositivity.
immunopositivity. Vascular walls as well as some axons and Vascular walls and endoneurium are thickened and hyalinated with
myelin sheath are thickened and hyalinated with strong AGE posi- strong AGE positivity. In almost all preserved axons, AGE positiv-
tivity (PAP 200x) ity is strong, while in myelin sheath it is weaker (PAP 100x)
I. Mišur et al.: AGE in peripheral nerve in type 2 diabetes 163

performed on transverse sections (Mallory three-chrome

staining) (Fig. 4). Comparison of diabetic and control
nerves showed perineurial thickening (15.5±4.9 vs.
6.6±2.1 µm, p<0.001), narrowing of microvascular lumen
(66.6±50.5 vs. 579.5±38.4 x103 µm2, p<0.001), and sig-
nificant reduction in the number of intact axons in dia-
betes (3.6±3 vs. 8.9±2.3 per 105 µm2 area, p<0.037).
Femoral nerve samples contained a slightly greater num-
ber of preserved axons than sural nerve samples; howev-
er, the difference was not statistically significant.

Immunofluorescence analysis

Indirect immunofluorescence with the polyclonal AGE

antibody revealed formation of AGE products in tissue
sections of sural and femoral nerves from diabetic
Fig. 3 Transverse section of peripheral nerve shows regular mor- patients. Most specimens showed positive reaction to
phology with no AGE immunopositivity (PAP 200x) AGE, recorded in the endoneurium, perineurium and peri-
axonal area. The intensity of fluorescence, indicating the
severity of AGE lesions, differed among the specimens
tested; thus they were difficult to compare. Analysis of
each specimen showed the intensity of fluorescence to
vary among different cellular structures. Almost all speci-
mens of the sural and femoral nerves expressed AGE pos-
itivity, which was recorded in at least one of the above
mentioned cell segments. Out of 10 specimens analyzed,
the immunofluorescence was obscure in only one speci-
men of the sural nerve (specimen no. 4). The specimen
was found to have poorly preserved tissue structure,
which may have been the reason for unclear or negative
result. AGE accumulation in sural nerve samples was
observed to yield a more diverse pattern (Fig. 5a). In most
of these specimens, positive findings were recorded in the
endoneurium, perineurium and around axons, however,
differing according to the intensity of fluorescence from
specimen to specimen. In femoral nerve specimens, the
Fig. 4 Quantitative analysis of axon surface in transverse sections
distribution of AGE immunopositivity was more uniform
of peripheral nerve of diabetic patients (400x, Mallory three-
and was mostly recorded in the periaxonal area (Fig. 5b).
chrome staining, ISSA and SFORM programs)

correlated with the severity of morphological alteration in

the nerves (p<0.005). The same correlation was observed in Plasma AGE immunoreactivity
the endoneurium of 7 of 10 nerve samples with lower inten-
sity of AGE immunopositivity (p<0.008). Microvascular The total level of circulating AGEs determined by competi-
immunopositivity of a weak intensity, which yielded no tive ELISA was significantly higher in serum of 8 patients
correlation with the severity of morphological alteration in with diabetic neuropathy than of 20 nondiabetic healthy
the nerves of diabetic patients, was found in 7 of 10 nerve subjects (38.6±6.9 vs. 25.1±7.2 µEq/ml, p<0.0009). The
samples. Perineurial AGE immunopositivity was detected presence of antibodies against AGE epitope(s) was detected
in only 4 nerve samples, and the intensity of reaction did in all serum samples examined, with a higher AGE antibody
not correlate with the severity of morphological alteration. titer recorded in the group of diabetics (58.7%±20.2% vs.
The control sural nerve was free from pathologic changes, 17.4%±15.4% inhibition, p<0.0001). Higher serum levels of
and AGE immunopositivity was absent (Fig. 3). circulating AGE-IC were demonstrated in diabetic patients
Morphometric analysis of the peripheral nerves was as compared with control subjects (5.9±2.4 vs. 3.39±1.1
164 I. Mišur et al.: AGE in peripheral nerve in type 2 diabetes

a b

Fig. 5a, b Immunofluorescence analysis of AGE immunopositivity with rabbit polyclonal primary antibody and fluorescein-labelled sec-
ondary antibody. a AGE deposits on the transverse section of sural nerve (case 1). Positive AGE reaction was recorded all over the sec-
tion surface, but was unevenly distributed in different section segments. High AGE positivity was found in the perineurium and in the area
of myelin sheaths. A reaction of lower and uneven intensity was recorded in the endoneurium (x160). b Femoral nerve section (case 7)
revealed clearly positive AGE reaction in the area of myelin sheaths. AGE positivity was absent in the preserved axonal cylinders (x400)

µEq/ml, p<0.0004). There was no significant correlation patients but not in the control specimen.
between serum AGE content and AGE-IC with either actual The overall intensity of AGE immunopositivity as a total
HbA1c values or those assayed over the previous year. sum of the scores of tissue components showed that the most
intensive positive reaction was detected in the area of axons
and myelin sheaths, whereas lower scores of AGE
immunoreactivity were recorded in the endoneurium, per-
Discussion ineurium, and microvessel walls. The intensity of total mor-
phological changes correlated significantly with the intensi-
The patients with type 2 diabetes included in the study ty of AGE immunoreactivity in axonal components, sug-
exhibited clinical signs and symptoms of distal symmetri- gesting that structural changes of the axon may be related to
cal diabetic polyneuropathy of a mean duration of 2.5 the reduction of axonal conductivity and impulse conduc-
years. This complication of diabetes develops in some tion impairments, thus contributing to functional events.
50% of patients in the first ten years of disease. It is of a The neurophysiological findings in all study patients indi-
gradually progressive course that correlates with diabetes cated significant retardation of the motor and sensory con-
duration and glycemia control. In the study patients, distal ductivity of the nerves analyzed, with very low amplitudes
symmetrical diabetic polyneuropathy was superimposed of the muscular and neural responses. Excessive accumula-
by proximal diabetic neuropathy, a rare but extremely tion of AGE content in the endoneurium could be explained
troublesome syndrome distinct from the former [15]. As by the fact that endoneurium is rich in collagenous connec-
differentiated from it, proximal diabetic neuropathy has tive tissue. Collagens are abundant in lysine and hydroxyly-
an acute onset, frequently of a fulminant course, thus lead- sine NH2 residues, and therefore collagen is highly suscep-
ing abruptly to a significant neurologic disability. tible to glycation [16, 17]. Abnormally glycated collagen in
Upon neurological examination, the patients under- the endoneurium of the nerve trunks might act as a physical
went femoral and sural nerve electroneurography. Biopsy barrier to elongation of the axonal sprouts. Additionally, it is
specimens of these nerves were obtained under spinal difficult to separate the effects of glycation on the axonal
anesthesia. The nerve samples available for analysis were cytoskeleton from glycation of the extracellular matrix
small, thus they were frozen as a whole in a cutting medi- because they are interrelated. If connection between the
um. Because of prefreezing, the tissue was not suitable for axon and its end tissues were damaged by glycation, this
ultrastructural analysis. Morphometric analysis of the dia- could alter transport functions.
betic peripheral nerve showed thickening of the perineuri- There have been reports on glycation and AGE accu-
um and narrowing the microvessel lumen with change in mulation in the myelin fraction of peripheral nerve [8, 18,
the axon density. We demonstrated localization of AGE 19]. In the present study, we observed AGE immunoreac-
immunoreactivity in the peripheral nerve of diabetic tivity in myelin sheaths. However, we found no correlation
I. Mišur et al.: AGE in peripheral nerve in type 2 diabetes 165

between the index of total morphological changes and elucidate the mechanisms by which the process of
index of myelin AGE immunoreactivity. This observation is advanced glycation could be involved in the pathogenesis
open to speculation. Namely, the myelin AGE products can of diabetic neuropathy.
act as ligands for the uptake and degradation by
macrophages via scavenger or AGE receptors [20]. The
interaction between peripheral nerve AGE myelin and AGE
cell surface receptor might contribute to the segmental References
demyelinization by a variety of mechanisms. However, the
presence of AGE-modified myelin in experimental as well 1. King RH (2001) The role of glycation in the pathogenesis of
as in human diabetes demonstrated in our study, appears to diabetic polyneuropathy. J Clin Pathol Mol Pathol
be less relevant for distal diabetic neuropathy than the gly- 54:400–408
cation of axonal cytoskeletal proteins. Diabetic polyneu- 2. Cameron NE, Eaton SEM, Cotter MA, Tesfaye S (2001)
Vascular factors and metabolic interactions in the pathogene-
ropathy is a distal, length-related axonal neuropathy where
sis of diabetic neuropathy. Diabetologia 44:1973–1988
axonal damage probably plays a major role. AGE formation 3. The Diabetes Control and Complications Trial Research
on axonal proteins may influence their turnover, alter phos- Group (1993) The effect of intensive treatment of diabetes on
phorylation and dephosphorylation, and affect molecular the development and progression of long-term complications
architecture and transport mechanisms, which may con- in insulin-dependent diabetes mellitus. N Engl J Med
tribute to axonal atrophy and degeneration. 329:977–986
In the present study, the serum of both diabetic 4. The Diabetes Control and Complications Trial Research
patients and nondiabetic individuals contained autoanti- Group (1998) The effect of intensive diabetes therapy on
bodies to epitope(s) of AGE structure. However, a higher measures of autonomic nervous system function in the
Diabetes Control and Complications Trial (DCCT).
AGE-antibody titer was observed in diabetic subjects.
Diabetologia 41:416–423
Additionally, we demonstrated the presence of circulating 5. Stratton IM, Adler AI, Neil HAW, Matthews DR, Manley SE,
immune complexes containing AGE moiety, predominant- Cull CA, Hadden D, Turner RC, Holman RR (2000)
ly in diabetic serum. Immune complexes may form conse- Association of glycaemia with macrovascular and microva-
quentially to the association of AGE-modified proteins scular complications of type 2 diabetes (UKPDS 35): pro-
and antibodies against AGE epitope(s). As many body spective observational study. BMJ 321:405–412
proteins are AGE modified, they may become a substrate 6. Carrington AL, Litchfield JE (1999) The aldose reductase
for the formation of insoluble immune complexes formed pathway and nonenzymatic glycation in the pathogenesis of
at the tissue level. Therefore, we speculate that excessive diabetic neuropathy: a critical review for the end of the 20th
glycated components of diabetic peripheral nerve are a century. Diabetes Rev 4:275–299
7. Vlassara H, Bucala R, Striker L (1994) Pathogenic effects of
substrate for the association with anti-AGE antibodies in
AGEs: biochemical, biologic, and clinical implications for
situ [13, 21, 22]. All patients included in the present study diabetes and aging. Lab Invest 70:138–151
had proximal diabetic neuropathy. There is a growing 8. Ryle C, Donaghy M (1995) Non-enzymatic glycation of peri-
body of evidence suggesting the potential involvement of pheral nerve proteins in human diabetics. J Neurol Sci
humoral immune mechanisms in the generation of neural 129:62–68
structural abnormalities, or of some of them in this partic- 9. Sugimoto K, Nishizawa Y, Horiuchi S, Yagihashi S (1997)
ular type of neuropathy. Localization in human diabetic peripheral nerve of N (epsi-
In conclusion, the results presented provide evidence lon)-carboxymethyl lysine-protein adducts, an advanced
that nonenzymatic glycation leading to AGE formation glycation endproduct. Diabetologia 40:1380–1387
10. Mišur I, Turk Z, Barada A, Miličević Z, Metelko Z (2001)
primarily involves axonal proteins and endoneurium. We
Advanced glycation endproducts in human diabetic periphe-
are inclined to believe that the lack of correlation between ral nerve. Diabetologia 44[Suppl 1]:A318
the index of total morphological alterations and index of 11. Thornalley PJ (1999) The clinical significance of glycation.
AGE immunoreactivity on the myelin protein may have Clin Lab 45:263–273
resulted from the fact that AGE myelin probably was par- 12. Turk Z, Mesić R, Benko B (1998) Comparison of advanced
tially internalized via AGE receptor. Also, humoral glycation endproducts on haemoglobin (Hb-AGE) and hae-
immune mechanisms, including the production of anti- moglobin A1c for the assessment of diabetic control. Clin
AGE autoantibody, may potentially be involved in the Chim Acta 277:159–170
development of structural abnormalities described in this 13. Turk Z, Ljubic S, Turk N, Benko B (2001) Detection of
autoantibodies against advanced glycation endproducts and
report. However, because of the small number of speci-
AGE-immune complexes in serum of patients with diabetes
mens and diversity of lesions, no regularity related to mellitus. Clin Chim Acta 303:105–115
morphological changes and AGE immunopositivity could 14. Jeppsson JO, Jerntorp P, Sundkvist G, Englund H, Nylund V
be definitely established. It can only be emphasized that (1986) Measurement of haemoglobin A1c by a new liquid-
we confirmed the peripheral nerve structure involvement chromatographic assay: methodology, clinical utility, and
by AGE modification. Additional studies are necessary to relation to glucose tolerance evaluated. Clin Chem
166 I. Mišur et al.: AGE in peripheral nerve in type 2 diabetes

32:1867–1872 32:670–674
15. Pascoe MK, Low PA, Windebank AJ, Litchy WJ (1997) 19. Ryle C, Leow CK, Donaghy M (1997) Nonenzymatic glyca-
Subacute diabetic proximal neuropathy. Mayo Clin Proc tion of peripheral and central nervous system proteins in
72:1123–1132 experimental diabetes mellitus. Muscle Nerve 20:577–584
16. Monnier VM, Sell DR, Nagaraj RH, Miyata S, Grandhee S, 20. Vlassara H, Brownlee M, Cerami A (1985) Recognition and
Odetti P, Ibrahim SA (1992) Maillard reaction-mediated mole- uptake of human diabetic peripheral nerve myelin by macro-
cular damage to extracellular matrix and other tissue proteins phages. Diabetes 34:553–557
in diabetes, aging, and uremia. Diabetes 41[Suppl 2]:36–41 21. Morano S, Tiberti C, Cristina G, Sensi M, Cipriani R,
17. Turk Z, Mišur I, Turk N, Benko B (1999) Rat tissue collagen Guidobaldi L, Torresi P, Medici F, Anastasi E, DiMario U
modified by advanced glycation: correlation with duration of (1999) Autoimmune markers and neurological complications
diabetes and glycemic control. Clin Chem Lab Med 37: in non-insulin-dependent diabetes mellitus. Hum Immunol
813–820 60:848–854
18. Vlassara H, Brownlee M, Cerami A (1983) Excessive non- 22. Turk Z, Šesto M, Skodlar J, Ferenčak G, Pokupec R, Turk N,
enzymatic glycosylation of peripheral and central nervous Stavljenić-Rukavina A (2003) The products of advanced
system myelin components in diabetic rats. Diabetes glycation in patients with type 2 diabetes and vascular disea-
se. Ann Clin Biochem 40:552–559