HAIKO SPROTT, ANDREAS MULLER, and HARTMUT HEINE. Objective. To determine if abnormal collagen metabolism is a characteristic of fibromyalgia.
ARTHRITIS & RHEUMATISM Vol. 40, No. 8, August 1997, pp 1450-1454 0 1997, American College of Rheumatology
1450
COLLAGEN CROSSLINKS IN FIBROMYALGIA HAIKO SPROTT, ANDREAS MULLER, and HARTMUT HEINE Objective. To determine if abnormal collagen metabolism is a characteristic of fibromyalgia. Methods. The diagnosis of fibromyalgia was made according to the American College of Rheumatology criteria. Skin biopsy samples were obtained from the trapezius region of 8 patients with fibromyalgia. Urine was collected under standardized conditions from 55 control subjects and 39 patients with fibromyalgia, and serum was obtained from 17 controls and 22 patients with fibromyalgia. Pyridinoline (Pyd), an indicator of connective tissue disease, and deoxypyridinoline (Dpyd), an indicator of bone degradation, both of which represent products of lysyl oxidase-mediated crosslinking in collagen, were analyzed by ion-paired and gradient high-performance liquid chromatography (HPLC) methods with fluorescence detection. Levels of hydroxyproline (Hyp), a collagen turnover marker, were also measured. The findings were related to creatinine levels, and the PydDpyd ratio was determined. Results. Highly ordered cuffs of collagen were observed around the terminal nerve fibers by electron microscopic examination of biopsy tissue from all 8 patients with fibromyalgia, but were not observed in any of the control skin samples. The Pyd:Dpyd ratios in the urine and serum and the Hyp levels in the urine were significantly lower in patients with fibromyalgia than in healthy controls. Conclusion. Decreased levels of collagen crosslinking in fibromyalgia may contribute to remodeling of the extracellular matrix and collagen deposition around the nerve fibers, and may contribute to the lower pain Dr. Miiller’s work was supported by the German Ministry for Research, Education and Technology (FKZ 01ZZ 9104/1). Dr. Heine’s work was supported by grants from the ERTOMIS Foundation, Familie-Ernst-Wendt Foundation, and E D E N Foundation (Germany). Haiko Sprott, MD, Andreas Miiller, PhD: University of Jena, Jena, Germany; Hartmut Heine, MD, PhD: University of Witteni Herdecke, Wilten, Germany. Address reprint requests to Haiko Sprott, MD, Department of Internal Medicine IV, D-07740 Jena, Germany. Submitted for publication November 7, 1996; accepted in revised form March 17, 1997.
threshold a t the tender points. Analysis of altered collagen metabolism either by histologic examination on biopsy, or preferably, by HPLC analysis of collagen metabolites in urine or serum may aid in understanding more about the pathogenesis of fibromyalgia.
Fibromyalgia is a noninflammatory disease of the soft tissue. Patients with fibromyalgia may experience pain over the entire body, but specific tender points are characteristic of the disease and are very painful (1). Fibromyalgia is generally considered to be associated with alterations in pain perception. Currently, the diagnosis of fibromyalgia is based solely on clinical findings (l),and there has been considerable controversy regarding the acceptance of fibromyalgia as a distinct clinical entity (2). The association of a specific marker of an underlying biochemical defect would aid in the resolution of these issues. Alterations in collagen metabolism may contribute to the pathophysiology of fibromyalgia. The tender points tend to be localized in areas of insertion of muscle in tendon and in tendon on bone. A higher order of matrix distribution involving regular layers of collagen has been observed histologically at the nerve endings in the tender points and may indicate alterations in collagen metabolism at these sites (3). Such alterations in collagen metabolism also might be detectable nonsurgically, since collagen degradation can be estimated from the measurement of collagen pyridinium crosslinks in urine or other bodily funds, including serum. F‘yridinoline (Pyd) and deoxypyridinoline (Dpyd) are trifunctional components of collagens and represent products of lysyl oxidase-mediated crosslinking (4). Pyd is abundant in cartilage and in connective tissue (including tendons, intramuscular collagen, and the aorta) (5). Dpyd has been detected in only a few different tissues, but the major pool is in bone collagen (6). Thus, the Pyd:Dpyd ratio indicates the extent of cartilage degradation. This ratio has been found to be elevated in diseases associated with increased collagen degradation,
1451
COLLAGEN CROSSLINKS IN FIBROMYALGIA
Table 1. Characteristics of the control subjects and patients with fibromydlgiti
Age, mean t SEM years Sex, fema1e:male
Healthy controls
Fibromyalgia patients
P
41.5 ? 1.7 8:1
48.1 t 1.6 18.5:l
0.35 0.11
such as rheumatoid arthritis (7). Hydroxyproline (Hyp) serves as a marker of collagen turnover (8). The current method of choice for estimation of collagen metabolism is simultaneous analysis of collagen crosslinks by high-performance liquid chromatography (HPLC) with detection based on the natural fluorescence of Pyd and Dpyd (9), and the determination of Hyp levels based on another HPLC method (10). We therefore analyzed collagen crosslinks in fibromyalgia to determine if these are potential biochemical markers for diagnosis of this disorder. PATIENTS AND METHODS Patients were diagnosed as having fibromyalgia according to the American College of Rheumatology criteria (1). Skin biopsy tissue was obtained from the trapezius region at the tender point (mean duration of pain 4.5 years) in 8 patients with fibromyalgia at the University of WitteniHerdecke. These samples were compared with skin biopsy tissue obtained from the corresponding area in healthy controls (n = 8). This anatomic area was the most painful region in all 8 patients with fibromyalgia. Two teams of observers from the University of WitteniHerdecke and the University of Hamburg, each comprising 2 readers who were blinded to the study design, examined the biopsy samples and showed 100% agreement regarding the characteristics of each sample. Urine and serum were collected under standardized conditions from 39 patients with fibromyalgia and from 55 healthy age- and sex-matched controls (Table 1) at the Friedrich Schiller University (Jena, Germany). Skin biopsy tissue (- 1 mm3) was prepared for electron microscopic analysis by immediate fixation in ice-cooled 1% osmium acid (0.1M phosphate buffer, pH 7.3) for 16 hours. Tissue was then washed in buffer solution, dehydrated through graded alcohol, and embedded in Epon. Sections were contrasted with uranyl acetate and lead citrate. A Zeiss EM9S2 microscope (Carl Zeiss, Jena, Germany) was used for electron microscopic analysis. Urine and serum samples were immediately centrifuged at 2,000g and frozen at -80°C. For analysis of Pyd and Dpyd levels, all urine and serum samples (V\amp,e= 250 p1 for urine and 0.5-1 ml for serum) were pretreated by hydrolysis (using an equal volume of 32% hydrochloric acid over 17 hours at 110°C) to obtain a peptide-free solution. Manual-partition chromatography of the hydrolysates, using CF1-cellulose (mobile phase, n-butano1:glacial acetic acid:water at 4:1:1), was performed according to an adaptation of the method of James and colleagues (9). The eluates were then lyophilized.
Pyd and Dpyd levels were measured by ion-paired chromatography and gradient separation on an RP-18 column with fluorescence detection (Ex 295 nm, Em 395 nm), using an intelligent pump (L6220), fluorescence detector (F-1080), and autosampler (AS 4000) (all from Merck, Darmstadt, Germany). The HPLC procedure was adapted from that of Colwell and colleagues (11). The injection volume for HPLC was 50 pl of loading buffer for urine and 200 pl of loading buffer for serum (1% n-heptafluorobutyric acid solution in water), in which the lyophilisates of urine and serum samples were dissolved. Inter-assay variation was assessed using pooled urine samples. The findings were related to creatinine levels (nmolesimmoles creatinine), and the Pyd:Dpyd ratio was determined. For the analysis of Hyp levels, urine samples (250 PI) from the same collectives were treated by resin-catalyzed hydrolysis of the urinary peptides overnight (16 hours) according to a modified method of Pang et a1 (12). We also used an internal standard for reproducible results. Hyp levels were measured by reverse-phase chromatography and gradient separation on an RP-18 column under fluorescence detection (Ex 470 nm, Em 530 nm). The modified HPLC method (10) included a fully automated pre-column derivatization with 7-chloro-4-nitrobenzofurazan using the robot function of the autosampler. Immediately after the derivatization step, the sample was analyzed. The injection volume for HPLC was 40 pl. The findings were related to creatinine levels (pmolesi mmoles creatinine). Statistical analysis was performed using SPSS for Windows (version 6.01 ; SPSS, Chicago, IL) and the Mann-Whitney U test. Results are reported as the median t SEM. Tests and P values were 2-tailed. A P value of 0.05 was considered significant.
RESULTS
In all biopsy tissues from patients with fibromyalgia, a collagen cuff was detected around the axons, that is, if any at all were found (Figure lb). No such changes were observed in any of the control tissues (Figure la). The differences in the Pyd:Dpyd ratio and in the Hyp levels between healthy controls and patients with fibromyalgia are shown in Figure 2 and Table 2. In urine, the ratios ranged between 1.507 and 11.272 for control subjects (mean 3.943, variance 4.094) and between 0.892 and 4.588 (mean 2.936, variance 0.525) for patients with fibromyalgia. Serum from 17 of the same control subjects (range 2.477-7.397, mean 4.359, variance 3.346) and from 22 of the same patients with fibromyalgia (range 0.360-5.565, mean 2.536, variance 2.449) were used in the HPLC examination. Serum from 38 controls and 17 patients with fibromyalgia were missing because these subjects declined the vein puncture. Dpyd was not measurable in the urine of 1 control subject (concentra-
SPROTT ET AL
1452
ratio. 7
p=O.O051
’
’
serum
I
I
6
5
’
urine
=0.0193
’
-
4 3 2 1
1
10%
Figure 2. Box plot of the pyridino1ine:deomyridinoline (PydiDpyd) ratio in urine and serum from healthy controls (clear box) and from patients with fibromyalgia (shaded box). * = statistically significant P values compared with controls. Bars show the median and the 25% and 75% values. The 10% and 90% values are also indicated.
5.808 and 74.974 kmoles/mmoles creatinine (mean 14.996, variance 142.29) for fibromyalgia patients. DISCUSSION Figure 1. Subepidermal connective tissue in representative skin samples from a healthy control subject and a patient with fibromyalgia. a, Uninvolved skin. Cross-section of 2 preterminal “naked” nerve fibers. Axons (arrows) are partly unfolded from Schwann’s cell sheath (S). A basal lamina (open arrow) surrounds the nerve fibers. G = ground substance (extracellular matrix); M = myofibroblast. b, Subepidermal connective tissue of fibromyalgia skin from a trapezius tender point. Cross-section of a nearly unmyelinated preterminal sensitive nerve with 3 axons. One of these is partly unfolded (arrow) from the Schwann’s cell sheath. A clearly visible collagen cuff (arrowheads) sheaths the preterminal nerve fibers. (Original magnification X 23,000.)
These studies indicate for the first time that there is a possible correlation between anatomic changes in subepidermal connective tissue (collagen cuffs around preterminal nerve fibers) and biochemical findings (decreased ratio in collagen crosslinks) in patients with fibromyalgia. An altered collagen metabolism could be a common biochemical defect in tender points of all Table 2. Ratio of pyridinoline to deoxypyridinoline (F‘yd:Dpyd) and levels of hydroxyproline (Hyp) in the urine and serum of control subjects and patients with fibromyalgia* Healthy controls ~~~
tion
