- Email: [email protected]
CSF enkephalins in diabetic neuropathy
Neuropeptides (1992) 22,125128 0 Longman Group UK Ltd 1992
CSF Enkephalins in Diabetic Neuropathy A. W. CHAN”, I. A. MACFARLANE”, E. A. MASSON”, D. BOWSHERt, J. S. MORLEYt and R. F. VENNt *Diabetes
Centre and tPain Research Institute,
Walton Hospital, Liverpool L9 lAE, U.K.
Abstract - CSF methionine and leucine enkephalins were measured by high performance liquid chromatography and radioimmunoassay in diabetic patients with painful neuropathy (n = 22) and painless neuropathy (n = 51, and non-diabetic subjects with low back pain (n = 11). Wide variations in CSF enkephalin levels were found and they were often below the limit of detection (< 0.1 pmol/l) in the diabetic and non-diabetic groups. The origin of CSF enkephalins is unknown and CSF levels may not reflect tissue concentrations. In conclusion, CSF enkephalin levels are difficult to interpret and do not provide useful information on the function of enkephalinergic pathways.
Introduction
Methods
Previous studies have suggested that chronically painful conditions may be associated with abnormally low concentrations of enkephalin-like compounds in the cerebrospinal fluid (1, 2, 3). No previous studies have examined CSF enkephalin levels in diabetic neuropathy, which may present with painful neuropathic symptoms or painless neuropathic foot ulceration. In this study, CSF methionine enkephalin and leucine enkephalin in diabetic patients with painful and painless varieties of peripheral neuropathy were compared with non-diabetic subjects with low back pain.
Subjects
Date received 2 December 199 1 Date accepted 5 February 1992 (Reprint request to AWC, Department of Geriatric Medicine, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 ZQQ, UK)
All diabetic patients had signs ofneuropathy, including impaired pinprick and light touch sensation. In 22 diabetic patients (17 men, 5 women: mean age 58.6 years; range: 28-83; SD 12.9) burning or shooting neuropathic lower limb pain had been present for at least 6 months (mean duration 4 years, SD 4.0, range 0.5-14 years). Other causes of neuropathic symptoms were excluded by normal screening investigations, including serum electrolytes, vitamin B-12, folate, liver and thyroid function tests. Peripheral vascular disease, which may mimic painful neuropathy, was excluded by measurement of systolic blood pressure at the ankle using Doppler ultrasound (4). Another 5 diabetic patients (all men: mean age 55.2 years, range 46-66, SD 8.2) with neuropathic foot ulceration, but without spontaneous 125
126 neuropathic pain, were also studied. The 11 non-diabetic subjects (7 men, 4 women: mean age 39.4 years, range 20-57, SD 9.2) had low back pain due to vertebral disk disease (mean duration of painful symptoms was 6 years, SD 11 .O, range 0.5 to 37 years). Pain scores The severity of painful symptoms was recorded on a visual analogue scale (VAS), represented by a horizontal 10 cm line and labelled at opposite ends by the words ‘no pain’ and ‘worst pain ever’ (5). Samples All CSF samples were obtained by lumbar punctures, using a standard aseptic technique and local skin analgesia (2% lignocaine). Approximately 5 ml CSF was collected into cold polystyrene tubes containing 20 pl formic acid/ml CSF, and extracted immediately or frozen at minus 25°C. Assays The CSF peptides were chromatographically separated from each other and from potentially crossreacting precursors or proteoly-tic products, and subsequently assayed by a specific RIA (6). The combination of high-performance liquid chromatography and radioimmunoassay (HPLC-RIA) is an improvement over other less specific methods which rely solely on radioimmunoassay or radioreceptor techniques in that unambiguous concentrations for each peptide can be determined. At the time of lumbar puncture, venous blood samples were collected from diabetic patients for measurements of glycosylated haemoglobin concentrations. Statistical calculations The HPLC-RIA system used in this study is able to detect very low concentrations of endogenous opiates in the CSF, with a lower limit of detection of 0.1 pmol/l. Previous studies with this method had shown that the CSF neuropeptide levels in non-diabetic individuals with chronic pain were not normally distributed and the values fell into a wide range (6). Therefore, the CSF data in the diabetic patients and control subjects were compared using Mann-Whitney U-tests. In some CSF samples, individual peptides were found to be below the limit of detection and these peptide levels were assigned values of zero in subsequent statistical calculations.
NEUROPEPTIDES
Calculations were performed with an Amstat statistics programme (S. C. Coleman Ltd., Ashby-de-laZouch, Leicester, UK) on an Amstrad personal computer. The computer programme performed a correction for identical numerical data, or ‘ties’, such as zero peptide levels. The relationships between CSF peptide levels and other measured parameters (e.g. intensity of neuropathic pain and glycaemic control) were examined using Spearman’s rank correlation coefficients (r). The study protocol was approved by the local ethics committee. Informed consent was obtained from all patients, who understood that a lumbar puncture could be a painful procedure and may be associated with a post-lumbar puncture headache. Results The mean pain score in the 11 non-diabetic subjects was significantly greater than the mean of the pain scores in the 22 diabetic patients with painful neuropathy(7.4cm,SD 1.7~~4.3 cm, SD 1.6;t=5.05, p < 0.001). The CSF enkephalin concentrations from diabetic patients and non-diabetic subjects fell into a very wide range and were often below the limit of detection (0.1 pmol/l) (Table). The Mann-Whitney U-test failed to demonstrate any significant difference in the concentration of these enkephalins between diabetic patients and non-diabetic subjects. Table CFS Enkephalin levels in diabetic neuropathy and low back pain Painful diabetic neuropathy n = 22
Painless diabetic neuropathy n=5
Low backpain n =I1
Methionine Enkephalin pmolil 1st Quartile median 3rd Quartile (< (range)
co.1 co.1 1.39 0.1-62.4)
co.1 < 0.1 co.1 (~0.1)
<: 0.1 co.1 63.5 (< 0.1-325.0)
Leucine Enkephalin pmolil 1st Quartile median 3rd Quartile (< (range)
co.1 0.53 15.15 0.1-43.0)
co.1 < 0.1 < 0.1 (< 0.1)
CO.1 co.1 105.9 (<: 0.1-225.0)
TheupperrangesofCSFmethionineandleucineenkephalinwere greater in non-diabetic patients with low back pain, but median concentrations did not differ significantly in the three patient groups.
CSF ENKEPHALINS
127
IN DIABETIC NEUROPATHY
In the diabetic patients with painful and painless neuropathies, both methionine and leucine enkephalin concentrations correlated significantly with pain scores (ME, r = 0.51, p = 0.006; LE, r = 0.47, p = 0.01) (Figs. 1 and 2). Diabetic control represented by glycosylated haemoglobin (mean 10.6% SD 2.9, range 6.5% to 17.2%; non-diabetic range 5.0~8.0%) had no apparent effect on the CSF enkephalin concentrations. In non-diabetic subjects, there was no correlation between VAS pain scores and the CSF enkephalin levels.
log scale
50
. . .
.
. .
10 5 .
Discussion
.
.
1
There is no previous study on the concentrations of these neuropeptides in the CSF of patients with diabetes. The concentrations of enkephalins in our dia-
. . ..
log scale 0.1
.
__o-__---’
v,, , , * .
.
0 .
1
. . . . . .. . (
2 VAS
.
3
4
PAIN
5
6
SCORE
,
(
,
,
7
8
9 10
(cm)
The relationship between CSF leucine enkephalin (LE) levels and the severity of diabetic neuropathic pain (10 cm visual analogue scale, VAS). (Key: painful diabetic neuropathy, closed circles 0; painless diabetic neuropathy, open circles C). Spearman’s rank correlation (r) between LE and VAS: r = 0.47, p = 0.01 (for all diabetic patients, n = 27); and r = 0.34, p = 0.12 (for diabetic patients with painful neuropathy. n = 22). Fig. 2
l
---
----
1
0
1
2
3 4
VAS
PAIN
5
6
SCORE
7
8
,
9 10
(pm)
Fig. 1 The relationship between CSF methionine enkephalin (ME) levels and the severity of diabetic neuropathic pain (10 cm visual analogue scale, VAS). (Key: painful diabetic neuropathy, closed circles 0; painless diabetic neuropathy, open circles 0). Spearman’s rank correlation (r) between ME and VAS: r = 0.5 1, p = 0.006 (for all diabetic patients, n = 27); and r = 0.47, p = 0.02 (for diabetic patients with painful neuropathy, n = 22). D
betic patients and non-diabetic subjects were considerably lower than those reported in previous human subjects. Hawkes et al (7) found CSF methionine enkephalin levels in patients undergoing dorsal column stimulation to be in the range of 20 to 60 pmoV1. Panerai et al (8), using HPLC-RIA techniques similar to ours, reported a mean CSF methionine enkephalin level of 5 pmol/l, and Kleine et al (9) reported a mean CSF methionine enkephalin level of 63 pmol/l in psychiatric patients. Our method uses HPLC to separate the high and low molecular weight peptides from each other, as well as separating the individual peptides one from another. The enkephalin antiserum used is specific for the pentapeptide sequences of methionine and
128 leucine enkephalin; thus it is not surprising that our results are lower than those found by others. All 5 diabetic patients with painless neuropathy had met- and leu-enkephalin levels of I 0.1 pmol/l. In the 22 diabetic patients with painful neuropathy, the range of met- and leu-enkephalin in the CSF was very large and many individuals had levels of less than 0.1 pmolfl. The highest methionine and leucine enkephalin levels were found in diabetic patients with the most severe neuropathic pain. These results are difficult to interpret for the following reasons: (1) the origin of CSF enkephalins is unknown; (2) there is no proof that CSF levels accurately reflect spinal tissue concentrations or the activity of enkephalinergic pathways (10). In conclusion, this HPLC-RIA method has shown that concentrations of enkephalins in the lumbar CSF of diabetic patients with peripheral neuropathy and non-diabetic subjects with low back pain were similar, usually less than 0.1 pmol/l, and lower than previously reported. In diabetic neuropathies, CSF enkephalin levels are difficult to interpret and their measurement do not provide useful information on the function of enkephalinergic pathways. Acknowledgements We are grateful toNovo Laboratories UK for their generous linancial support.
NEUROPEPTIDES
References 1. Almay, B. G. L., Johansson, F., von Knorring, L., Terenius and Wahlstrom, A. (1978). Differences in CSF endotphins between organic and psychogenic pain syndromes. Pain 5: 153-62. 2. Puig. M. M.. Laorden, M. L., Miralles. F. S. and Olaso, M. J. (1982). Endorphin levels in cerebrospinal fluid ofpatients with postoperative and chronic pain. Anesthiosiology 57: 1-4. 3. Simmonet, G., Taquet, H. Floras, P., Caille, J. M., Legrand, J. C., Vincent, J. D. and Cesselin, F. (1986). Simultaneous determination of radioimmunoassayable methionineenkephalin and radioreceptor-active opiate peptides in CSF of chronic pain suffering and non-suffering patients. Neuropeptides 7: 229-40. 4. Yao, S. T. (1970). Haemodynamic studies in peripheral vasculardisease. Br. J. Surg. 57: 761-66. 5. Huskisson. E. C. (1974). Measurement of oain. Lancet ii: 1127-113L ~ ’ 6. Venn, R. F. (1987). Combined high-performance liquid chromatographic-radioimmunoassay method for the analysis of endorphins, enkephalinsandotherneurotransmitterpeptides. J. Chromatography 423: 93-104. 7. Hawkes, C. H., Fawcett, D., Cookes, E. D., Emson, P. C., Paul, E. A. and Bowcock, S. A. (I 98 1). Dorsal column stimulation in multiple sclerosis: effects on bladder, leg blood flow and peptides. Appl. Neurophysiol. 44: 62-70. 8. Panerai, A. E., Martini, A., De Rosa, A., Sacerdote, P. and Fraioli, F. (1983). A HPLCRIA coupled technique for neuropeptides: its application to plasma and cerebrospinal fluid samples. Giom. It. Chim. Clin. 8 (supp): 39-47. 9. Kleine, T. O., Merten, B. and Singh, A. (1982). Determination of methionine enkephalin and leucine enkephalin in human cerebrosninal fluid. Protides Biol. Fluids 30: 243-246. 10. Morley, J. S. (1985). Peptides in nociceptive pathways. In: Lipton, S., Miles, J., eds. Persistent Pain, vol. 5. London/New York Gnme and Stratton. 65-92.
CSF Enkephalins in Diabetic Neuropathy A. W. CHAN”, I. A. MACFARLANE”, E. A. MASSON”, D. BOWSHERt, J. S. MORLEYt and R. F. VENNt *Diabetes
Centre and tPain Research Institute,
Walton Hospital, Liverpool L9 lAE, U.K.
Abstract - CSF methionine and leucine enkephalins were measured by high performance liquid chromatography and radioimmunoassay in diabetic patients with painful neuropathy (n = 22) and painless neuropathy (n = 51, and non-diabetic subjects with low back pain (n = 11). Wide variations in CSF enkephalin levels were found and they were often below the limit of detection (< 0.1 pmol/l) in the diabetic and non-diabetic groups. The origin of CSF enkephalins is unknown and CSF levels may not reflect tissue concentrations. In conclusion, CSF enkephalin levels are difficult to interpret and do not provide useful information on the function of enkephalinergic pathways.
Introduction
Methods
Previous studies have suggested that chronically painful conditions may be associated with abnormally low concentrations of enkephalin-like compounds in the cerebrospinal fluid (1, 2, 3). No previous studies have examined CSF enkephalin levels in diabetic neuropathy, which may present with painful neuropathic symptoms or painless neuropathic foot ulceration. In this study, CSF methionine enkephalin and leucine enkephalin in diabetic patients with painful and painless varieties of peripheral neuropathy were compared with non-diabetic subjects with low back pain.
Subjects
Date received 2 December 199 1 Date accepted 5 February 1992 (Reprint request to AWC, Department of Geriatric Medicine, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 ZQQ, UK)
All diabetic patients had signs ofneuropathy, including impaired pinprick and light touch sensation. In 22 diabetic patients (17 men, 5 women: mean age 58.6 years; range: 28-83; SD 12.9) burning or shooting neuropathic lower limb pain had been present for at least 6 months (mean duration 4 years, SD 4.0, range 0.5-14 years). Other causes of neuropathic symptoms were excluded by normal screening investigations, including serum electrolytes, vitamin B-12, folate, liver and thyroid function tests. Peripheral vascular disease, which may mimic painful neuropathy, was excluded by measurement of systolic blood pressure at the ankle using Doppler ultrasound (4). Another 5 diabetic patients (all men: mean age 55.2 years, range 46-66, SD 8.2) with neuropathic foot ulceration, but without spontaneous 125
126 neuropathic pain, were also studied. The 11 non-diabetic subjects (7 men, 4 women: mean age 39.4 years, range 20-57, SD 9.2) had low back pain due to vertebral disk disease (mean duration of painful symptoms was 6 years, SD 11 .O, range 0.5 to 37 years). Pain scores The severity of painful symptoms was recorded on a visual analogue scale (VAS), represented by a horizontal 10 cm line and labelled at opposite ends by the words ‘no pain’ and ‘worst pain ever’ (5). Samples All CSF samples were obtained by lumbar punctures, using a standard aseptic technique and local skin analgesia (2% lignocaine). Approximately 5 ml CSF was collected into cold polystyrene tubes containing 20 pl formic acid/ml CSF, and extracted immediately or frozen at minus 25°C. Assays The CSF peptides were chromatographically separated from each other and from potentially crossreacting precursors or proteoly-tic products, and subsequently assayed by a specific RIA (6). The combination of high-performance liquid chromatography and radioimmunoassay (HPLC-RIA) is an improvement over other less specific methods which rely solely on radioimmunoassay or radioreceptor techniques in that unambiguous concentrations for each peptide can be determined. At the time of lumbar puncture, venous blood samples were collected from diabetic patients for measurements of glycosylated haemoglobin concentrations. Statistical calculations The HPLC-RIA system used in this study is able to detect very low concentrations of endogenous opiates in the CSF, with a lower limit of detection of 0.1 pmol/l. Previous studies with this method had shown that the CSF neuropeptide levels in non-diabetic individuals with chronic pain were not normally distributed and the values fell into a wide range (6). Therefore, the CSF data in the diabetic patients and control subjects were compared using Mann-Whitney U-tests. In some CSF samples, individual peptides were found to be below the limit of detection and these peptide levels were assigned values of zero in subsequent statistical calculations.
NEUROPEPTIDES
Calculations were performed with an Amstat statistics programme (S. C. Coleman Ltd., Ashby-de-laZouch, Leicester, UK) on an Amstrad personal computer. The computer programme performed a correction for identical numerical data, or ‘ties’, such as zero peptide levels. The relationships between CSF peptide levels and other measured parameters (e.g. intensity of neuropathic pain and glycaemic control) were examined using Spearman’s rank correlation coefficients (r). The study protocol was approved by the local ethics committee. Informed consent was obtained from all patients, who understood that a lumbar puncture could be a painful procedure and may be associated with a post-lumbar puncture headache. Results The mean pain score in the 11 non-diabetic subjects was significantly greater than the mean of the pain scores in the 22 diabetic patients with painful neuropathy(7.4cm,SD 1.7~~4.3 cm, SD 1.6;t=5.05, p < 0.001). The CSF enkephalin concentrations from diabetic patients and non-diabetic subjects fell into a very wide range and were often below the limit of detection (0.1 pmol/l) (Table). The Mann-Whitney U-test failed to demonstrate any significant difference in the concentration of these enkephalins between diabetic patients and non-diabetic subjects. Table CFS Enkephalin levels in diabetic neuropathy and low back pain Painful diabetic neuropathy n = 22
Painless diabetic neuropathy n=5
Low backpain n =I1
Methionine Enkephalin pmolil 1st Quartile median 3rd Quartile (< (range)
co.1 co.1 1.39 0.1-62.4)
co.1 < 0.1 co.1 (~0.1)
<: 0.1 co.1 63.5 (< 0.1-325.0)
Leucine Enkephalin pmolil 1st Quartile median 3rd Quartile (< (range)
co.1 0.53 15.15 0.1-43.0)
co.1 < 0.1 < 0.1 (< 0.1)
CO.1 co.1 105.9 (<: 0.1-225.0)
TheupperrangesofCSFmethionineandleucineenkephalinwere greater in non-diabetic patients with low back pain, but median concentrations did not differ significantly in the three patient groups.
CSF ENKEPHALINS
127
IN DIABETIC NEUROPATHY
In the diabetic patients with painful and painless neuropathies, both methionine and leucine enkephalin concentrations correlated significantly with pain scores (ME, r = 0.51, p = 0.006; LE, r = 0.47, p = 0.01) (Figs. 1 and 2). Diabetic control represented by glycosylated haemoglobin (mean 10.6% SD 2.9, range 6.5% to 17.2%; non-diabetic range 5.0~8.0%) had no apparent effect on the CSF enkephalin concentrations. In non-diabetic subjects, there was no correlation between VAS pain scores and the CSF enkephalin levels.
log scale
50
. . .
.
. .
10 5 .
Discussion
.
.
1
There is no previous study on the concentrations of these neuropeptides in the CSF of patients with diabetes. The concentrations of enkephalins in our dia-
. . ..
log scale 0.1
.
__o-__---’
v,, , , * .
.
0 .
1
. . . . . .. . (
2 VAS
.
3
4
PAIN
5
6
SCORE
,
(
,
,
7
8
9 10
(cm)
The relationship between CSF leucine enkephalin (LE) levels and the severity of diabetic neuropathic pain (10 cm visual analogue scale, VAS). (Key: painful diabetic neuropathy, closed circles 0; painless diabetic neuropathy, open circles C). Spearman’s rank correlation (r) between LE and VAS: r = 0.47, p = 0.01 (for all diabetic patients, n = 27); and r = 0.34, p = 0.12 (for diabetic patients with painful neuropathy. n = 22). Fig. 2
l
---
----
1
0
1
2
3 4
VAS
PAIN
5
6
SCORE
7
8
,
9 10
(pm)
Fig. 1 The relationship between CSF methionine enkephalin (ME) levels and the severity of diabetic neuropathic pain (10 cm visual analogue scale, VAS). (Key: painful diabetic neuropathy, closed circles 0; painless diabetic neuropathy, open circles 0). Spearman’s rank correlation (r) between ME and VAS: r = 0.5 1, p = 0.006 (for all diabetic patients, n = 27); and r = 0.47, p = 0.02 (for diabetic patients with painful neuropathy, n = 22). D
betic patients and non-diabetic subjects were considerably lower than those reported in previous human subjects. Hawkes et al (7) found CSF methionine enkephalin levels in patients undergoing dorsal column stimulation to be in the range of 20 to 60 pmoV1. Panerai et al (8), using HPLC-RIA techniques similar to ours, reported a mean CSF methionine enkephalin level of 5 pmol/l, and Kleine et al (9) reported a mean CSF methionine enkephalin level of 63 pmol/l in psychiatric patients. Our method uses HPLC to separate the high and low molecular weight peptides from each other, as well as separating the individual peptides one from another. The enkephalin antiserum used is specific for the pentapeptide sequences of methionine and
128 leucine enkephalin; thus it is not surprising that our results are lower than those found by others. All 5 diabetic patients with painless neuropathy had met- and leu-enkephalin levels of I 0.1 pmol/l. In the 22 diabetic patients with painful neuropathy, the range of met- and leu-enkephalin in the CSF was very large and many individuals had levels of less than 0.1 pmolfl. The highest methionine and leucine enkephalin levels were found in diabetic patients with the most severe neuropathic pain. These results are difficult to interpret for the following reasons: (1) the origin of CSF enkephalins is unknown; (2) there is no proof that CSF levels accurately reflect spinal tissue concentrations or the activity of enkephalinergic pathways (10). In conclusion, this HPLC-RIA method has shown that concentrations of enkephalins in the lumbar CSF of diabetic patients with peripheral neuropathy and non-diabetic subjects with low back pain were similar, usually less than 0.1 pmol/l, and lower than previously reported. In diabetic neuropathies, CSF enkephalin levels are difficult to interpret and their measurement do not provide useful information on the function of enkephalinergic pathways. Acknowledgements We are grateful toNovo Laboratories UK for their generous linancial support.
NEUROPEPTIDES
References 1. Almay, B. G. L., Johansson, F., von Knorring, L., Terenius and Wahlstrom, A. (1978). Differences in CSF endotphins between organic and psychogenic pain syndromes. Pain 5: 153-62. 2. Puig. M. M.. Laorden, M. L., Miralles. F. S. and Olaso, M. J. (1982). Endorphin levels in cerebrospinal fluid ofpatients with postoperative and chronic pain. Anesthiosiology 57: 1-4. 3. Simmonet, G., Taquet, H. Floras, P., Caille, J. M., Legrand, J. C., Vincent, J. D. and Cesselin, F. (1986). Simultaneous determination of radioimmunoassayable methionineenkephalin and radioreceptor-active opiate peptides in CSF of chronic pain suffering and non-suffering patients. Neuropeptides 7: 229-40. 4. Yao, S. T. (1970). Haemodynamic studies in peripheral vasculardisease. Br. J. Surg. 57: 761-66. 5. Huskisson. E. C. (1974). Measurement of oain. Lancet ii: 1127-113L ~ ’ 6. Venn, R. F. (1987). Combined high-performance liquid chromatographic-radioimmunoassay method for the analysis of endorphins, enkephalinsandotherneurotransmitterpeptides. J. Chromatography 423: 93-104. 7. Hawkes, C. H., Fawcett, D., Cookes, E. D., Emson, P. C., Paul, E. A. and Bowcock, S. A. (I 98 1). Dorsal column stimulation in multiple sclerosis: effects on bladder, leg blood flow and peptides. Appl. Neurophysiol. 44: 62-70. 8. Panerai, A. E., Martini, A., De Rosa, A., Sacerdote, P. and Fraioli, F. (1983). A HPLCRIA coupled technique for neuropeptides: its application to plasma and cerebrospinal fluid samples. Giom. It. Chim. Clin. 8 (supp): 39-47. 9. Kleine, T. O., Merten, B. and Singh, A. (1982). Determination of methionine enkephalin and leucine enkephalin in human cerebrosninal fluid. Protides Biol. Fluids 30: 243-246. 10. Morley, J. S. (1985). Peptides in nociceptive pathways. In: Lipton, S., Miles, J., eds. Persistent Pain, vol. 5. London/New York Gnme and Stratton. 65-92.