The detection and quantitation of free desmosine and isodesmosine in human urine and their peptide-bound forms in sputum

Abstract

Desmosine (D) and isodesmosine (I), the intramolecular crosslinking amino acids that occur in chains of elastin, have now been found in free form in human urine. Until now, these amino acids (M_r = 526) were found to occur in urine only as higher molecular weight (M_r = 1,000–1,500) peptides. Thus, the previously used analytical methods required, as the first step, acid hydrolysis of the urine at elevated temperature to liberate D and I from their peptides. The analytical method described here uses HPLC followed by electrospray ionization MS for the detection and quantitation of free D and I in unhydrolyzed urine. Identities of both D and I were established by their retention times on LC and by their mass ion at 526 atomic mass units, characteristic of each compound. The sensitivity of the method is 0.10 ng. The average values of free D and I in the urine of seven healthy subjects were 1.42 ± 1.16 and 1.39 ± 1.04 μg/g of creatinine, respectively. After acid hydrolysis of the urine, the amounts of D and I were 8.67 ± 3.75 and 6.28 ± 2.87 μg/g of creatinine, respectively. The method was also successfully used to measure peptide-bound D and I levels in the sputum of patients with chronic obstructive pulmonary disease.

Two pyridinium amino acids, desmosine (D) and isodesmosine (I), are positional isomers that serve as crosslinking molecules binding the polymeric chains of amino acids into the 3D network of elastin (1–3). The degradation of elastin-containing tissues that occurs in several widely prevalent diseases, such as pulmonary emphysema, chronic obstructive pulmonary disease (COPD), cystic fibrosis, atherosclerosis, aortic aneurysm, etc., has been associated with the excretion in the urine of peptic derivatives of these two pyridinium compounds (4–13). Previously identified (14) urinary metabolites are a group of high molecular weight polypeptic derivatives of D and I ranging in M_r from 1,000 to 1,500. In this paper, we report that D and I are also excreted in urine in free form. Moreover, after hydrolysis, D and I have been found in sputum obtained from patients with COPD.

Previously used techniques for the analysis of urinary D and I have used RIA (15, 16), HPLC (17–19), and capillary zone electrophoresis (20, 21). In each case, acid hydrolysis (treatment of the urine with HCl at elevated temperature for 24 h) was used as the first step to liberate D and I from their peptidic conjugates. The analytical method described here is sufficiently sensitive and specific so that it is possible to analyze for free D and I even though they are less abundant in urine than their peptidic forms. The method involves HPLC followed by electrospray ionization MS (ESIMS). Besides, avoiding the drastic HCl hydrolysis (which severely complicates the subsequent purification process), this method makes possible the identification of D and I with high specificity and sensitivity, such that the analysis of only 1–3 ml of urine is possible. Moreover, other fluids of concern (sputum, lung lavage, etc.) where only small samples are available may also be analyzed by this method.

Tens of millions of people in this country alone suffer from various forms of lung disease, and consequently the method described here may find clinical use to determine elastin tissue destruction, to evaluate pathogenesis and severity of lung disease, and to estimate the efficacy of potential therapies.

Materials and Methods

Chemicals. HCl (37%, ACS reagent and 20% double distilled), methanol (HPLC grade), acetonitrile (HPLC grade), water (HPLC grade), heptafluorobutyric acid (99%), ammonium acetate (99.999%), and acetic acid (double distilled, 99.99+%) were purchased from Sigma–Aldrich. D and I standards were obtained from Elastin Products (Owensville, MI).

Creatinine and Protein Measurement. Urine creatinine was measured by the commercially available 555A creatinine kit (Sigma). Total protein in sputum samples was measured by the commercially available microprotein assay kit (Sigma), which is based on protein-dye (Coomassie blue) binding (22).

Sample Preparation for LC/MS Analysis. Urine and sputum samples were obtained from volunteers with informed consent at the James P. Mara Center for Lung Disease, St. Luke's/Roosevelt Hospital Center, New York.

Urine Samples. Twenty-four-hour urine samples were collected in 0.2% boric acid and stored at -20°C before analysis. Sputum from patients with previously diagnosed COPD was produced spontaneously and collected over a 3-h period in plastic containers before analysis.

For determination of free D and I, 1–3 ml of urine was mixed with an equal volume of 0.1 M HCl and applied to 3 ml of MCX cation-exchange column (Waters), which had been prewashed with 3 ml of methanol and 3 ml of H₂O. The column was then washed with 3 ml of H₂O/3 ml of methanol followed by 3 ml 0.1 M HCl. The D and I adsorbed on the column were finally eluted with 3 ml of 2 M HCl. This eluate was evaporated to dryness under vacuum at 45°C. The dry residue was dissolved in 0.1 ml of HPLC mobile phase used in the LC/ESIMS analysis.

For the determination of total (both free and peptide-bound) D and I, the urine sample was added to an equal volume of conc. HCl (12 M) in a glass vial; the air in the vial was displaced by a stream of nitrogen, and then the capped vial was heated at 110°C for 24 h. The acid hydrolyzed mixture was filtered, the filtrate was evaporated to dryness at 45°C under vacuum, and the residue was chromatographed on the cation-exchange column as described above. The appropriate elute was analyzed by LC/ESIMS.

Sputum Samples. Sputum samples of COPD were collected at the clinic. Sputum from subjects with no history or evidence of lung disease was produced by a sputum induction procedure by using inhalation of 3% sodium chloride aerosol over 20 min. One-half to 1 ml of sputum sample, either from normal or COPD patients, was added to an equal volume of conc. HCl and hydrolyzed at 110°C for 24 h. The acid hydrolyzed mixture was chromatographed on the cation-exchange column as described in the analysis of the urine samples. The content of D and I was analyzed by LC/ESIMS.

LC/ESIMS Analysis. The Alliance 2690 HPLC separation module with a 2487 UV and a ZQ 2000 ESIMS detector system (Waters) was used for the analysis. HPLC separation of D and I was achieved by a 150 × 2.1-mm IC-Pak (5 μm) cation-exchange column (Waters) and the temperature of the column set on 40°C. The mobile phase consisted of a mixture of 95% of solvent A (10 mM ammonium acetate solution containing 0.01% heptafluorobutyric acid and adjusted to pH 3.1 by acetic acid) and 5% of solvent B (acetonitrile). Isocratic elution was carried out at a flow rate of 0.3 ml/min.

The ESIMS was operated in the positive ion mode, and detection was achieved simultaneously both by scanning [200–600 atomic mass units (amu)] and selected ion recording (SIR) at 526 ± 1 amu. In this way, both identification and quantification of the compounds were established. The optimum mass spectrometric parameters were capillary voltage 3.20 kV, sample cone voltage 40 V, ion energy 0.5 eV, amplifier voltage 650 V, and temperatures of the desolvation and the source at 400°C and 120°C, respectively.

Results and Discussion

LC/ESIMS of D and I. The LC/ESIMS results on reference samples of D and I are shown in Fig. 1. Two isomers are well separated in the HPLC chromatogram (Fig. 1a) by using a 150 × 2.1-mm cation-exchange IC Pak column (Waters). The electrospray mass spectra of both D and I exhibited an intense molecular ion [M]⁺ at m/z 526 with little fragmentation. However, on raising the cone voltage from 40 to 80 V, four major fragmentation ions: m/z 481 [M-COOH]⁺, m/z 436 [M-2COOH]⁺, m/z 396 [M-CH₂CH₂CH₂CH₂CH(NH₂)COOH]⁺, and m/z 351 [M-CH₂CH₂CH₂CH₂CH(NH₂)COOH-COOH]⁺, appeared (Fig. 1 b and c). Although the fragmentation ions with both isomers are similar, differences in their intensity can be used to differentiate the two isomers; the more intense ion fragments of m/z 396 and 351 in I (the elimination of the amino acid substituent at the pyridinium nitrogen) can serve this purpose. The most abundant [M]⁺ at m/z 526 was chosen for the SIR mode for quantitative analysis of D and I.

Fig. 1.

LC/ESIMS of D and I. (a) Total ion chromatogram in a 150 × 2.1-mm IC Pak. (b) Mass spectrum of I at 80 V. (c) Mass spectrum of D at 80 V.

Quantitative Analysis of Free and Total D and I in Human Urine and Sputum. Fig. 2 shows the SIR analysis of standards of both D and I from amounts of 0.10–50.00 ng. These plots were used as the calibration curves for quantifying D and I in urine and sputum samples.

Fig. 2.

Linear regression calibration curve. D [0.10–50.00 ng (▴)], y = 926.015x - 113.437, r = 0.9984; I [0.10–50.00 ng (⋄)], y = 1,088.69x - 110.994, r = 0.9999.

Fig. 3 shows typical SIR chromatograms obtained from the urine of a normal subject. Fig. 3a illustrates the chromatographic analysis (by LC/ESIMS by using SIR of m/z 526 atomic mass units) of the two reference compounds. Fig. 3b shows the analysis of the total D and I content of normal urine after acid hydrolysis. Fig. 3c shows the analysis of the free D and I content of the same normal unhydrolyzed urine.

Fig. 3.

SIR chromatograms of D and I. (a) Reference D and I standards. (b) Total D and I in urine (after HCl hydrolysis). (c) Free D and I in urine (without HCl hydrolysis).

The commonly used HPLC methods for the analysis of D and I in urines all use their UV absorption at 268 nm for detection, which has a detection sensitivity of ≈1/100th that of the SIR method used here. The SIR method permits the analysis of 0.1 ng (0.2 pmol) of D or I. The sensitivity and specificity of the SIR analysis make unnecessary the acid hydrolysis of 20–30 ml of urine, and the subsequent time-consuming cleanup procedures, which are needed to remove the impurities interfering with the UV absorption. In fact, the D and I in the urine samples analyzed by the SIR method, shown in Fig. 3 b and c, could not be detected by their UV absorption due to the overwhelming intense background interferences (the chromatograms not shown).

Tables 1 and 2 show the results obtained for the free and total D and I in the urine samples from seven normal human subjects. Each sample was analyzed three times, and the average values are expressed as microgram per gram of creatinine. The recoveries of D and I from urine samples were >90%, which had been established by the parallel analysis of reference standard solutions of D and I.

Table 1. Free D and I in normal urine (micrograms per gram of creatinine).

Subjects (sex, age)	D	I
1 (M, 78)	0.42 ± 0.01	0.38 ± 0.01
2 (M, 70)	0.75 ± 0.16	0.35 ± 0.19
3 (M, 50)	0.51 ± 0.11	0.64 ± 0.17
4 (F, 53)	2.98 ± 0.33	2.75 ± 0.37
5 (M, 32)	1.55 ± 0.20	1.59 ± 0.26
6 (F, 28)	0.50 ± 0.15	0.43 ± 0.05
7 (M, 28)	3.33 ± 0.40	2.42 ± 0.31

M, male; F, female.

Table 2. Total D and I in normal urine (micrograms per gram of creatinine).

Subjects (sex, age)	D	I
1 (M, 78)	6.17 ± 0.32	4.67 ± 0.19
2 (M, 70)	5.16 ± 0.15	4.10 ± 0.02
3 (M, 50)	5.17 ± 0.13	3.96 ± 0.01
4 (F, 53)	11.28 ± 0.21	9.96 ± 0.56
5 (M, 32)	12.74 ± 0.83	11.44 ± 0.58
6 (F, 28)	5.67 ± 0.62	4.08 ± 0.34
7 (M, 28)	14.60 ± 1.75	5.77 ± 0.33

M, male; F, female.

The average of free D and I in the unhydrolyzed urines of the seven healthy subjects were determined to be 1.42 ± 1.16 and 1.39 ± 1.04 μg/g creatinine, respectively (Table 1); these values are ≈15% of those of the total D and I. The values of free D and I in normal urine has not been reported previously.

The average of the total D and I in the urine of the seven healthy subjects was 8.67 ± 3.75 and 6.28 ± 2.87 μg/g creatinine, respectively (Table 2). These values approximate those reported by HPLC analysis; 7.5 ± 1.7 and 6.8 ± 1.4 μg/g creatinine (8) and by capillary zone electrophoresis, 9.31 ± 2.75 and 7.34 ± 1.95 μg/g creatinine (9). On the other hand, values for total D and I excreted in the urine of healthy subjects determined by the RIAs (6, 16) were 5–10 times higher.

Table 3 shows the values obtained by the LC/ESIMS method for the total D and I content of sputum from patients diagnosed with COPD. Sputum samples obtained by sputum induction from three volunteers without lung disease revealed no detectable levels of free D and I. Further application of the method to the analysis of sputum samples and evaluation of their correlation to COPD will be described in subsequent reports. After completion of this paper, a report on LC/ESIMS analysis of D and I appeared (23); however, the method was applied to the analysis of D and I in rat lung after the conc. HCl hydrolysis of whole lung tissue.

Table 3. Total D and I in sputum of COPD patients, nanograms per milligram of protein.

Subjects (sex, age)	D	I
1 (M, 62)	2.07	1.27
2 (F, 64)	0.42	0.37
3 (M, 46)	0.37	0.31
4 (M, 52)	1.37	0.99
5 (M, 54)	0.29	0.25

M, male; F, female.