Anal. Chem.
2010, 82, 3745–3750
Measurement of Urinary Total Desmosine and
Isodesmosine Using Isotope-Dilution Liquid
Chromatography-Tandem Mass Spectrometry
Osama Albarbarawi,† Alun Barton,† ZhaoSheng Lin,‡ Eddie Takahashi,‡ Ajay Buddharaju,†
Jeffrey Brady,†,‡ Douglas Miller,‡ Colin N. A. Palmer,§ and Jeffrey T.-J. Huang*,†,‡,§
Translational Medicine Research Collaboration, TMRC Laboratory, James Arrott Drive, Dundee, DD1 9SY,
Pfizer Inc., United Kingdom, and Biomedical Research Institute, Ninewells Hospital and Medical School,
University of Dundee
The current LC-MS based desmosine/isodesmosine (DES/
IDS) assays may be unsatisfactory for clinical use due to
lack of an appropriate internal standard or low through-
put. A fast and reliable LC-MS method using a D5-DES
as an internal standard for measuring urinary total
DES/IDS was developed and validated in this study.
The reportable range of this assay was 1.0 and 480.0
ng/mL. The intra- and interassay imprecision, ac-
curacy, and recovery for quality control samples were
within acceptable range (<25%). Urinary total DES/
IDS level was stable at room temperature or 4 °C for
20 h, and for three freeze/thaw cycles. The assay was
employed to measure urine samples from COPD
patients and demographically matched healthy volun-
teers. The total urinary DES/IDS levels were ap-
proximately 3-fold higher in COPD patients compared
to healthy volunteers. The suitability of using urinary
free DES to estimate elastin degradation was also
evaluated in a second cohort. Despite urinary free and
total DES/IDS levels being highly correlated, our data
suggest that urinary total DES/IDS level is a preferred
biomarker for elastin degradation. These results dem-
onstrate that the LC-MS/MS method provides sensi-
tive, reproducible and accurate quantification of uri-
nary total DES/IDS as a biomarker for monitoring
elastin degradation in diseases such as COPD.
Desmosine (DES) and isodesmosine (IDS) are isomeric
pyridinium-based amino acids resulting from the condensation of
four lysine residues by lysyl-oxidase,1,2 through which cross-
linking networks within elastin are formed. These special cross-
linkings are only found in mature elastin in human and the release
of these two compounds and/or DES- or IDS-containing peptides
to body fluids such as sputum, plasma, or urine have been used
* To whom correspondence should be addressed. E-mail: jeffrey.huang@
pfizer.com.
†
TMRC Laboratory.
‡
Pfizer Inc.
§
University of Dundee.
(1) Foster, J. A.; Rubin, L.; Kagan, H. M.; Franzblau, C.; Bruenger, E.; Sandberg,
L. B. J. Biol. Chem. 1974, 249, 6191–6196.
(2) Brown-Augsburger, P.; Tisdale, C.; Broekelmann, T.; Sloan, C.; Mecham,
R. P. J. Biol. Chem. 1995, 270, 17778–17783.
10.1021/ac100152f  2010 American Chemical Society
Published on Web 04/02/2010
to estimate levels of elastin degradation. This is particularly the
case for diseases such as chronic obstructive pulmonary disease
(COPD) and cystic fibrosis where degradation of elastin-containing
tissues is believed to occur due to an imbalance between proteases
and their inhibitors (for review, see refs 3 and 4).
Urine samples are the most frequently studied body fluid in
terms of DES/IDS measurements for respiratory diseases, prob-
ably due to the fact that total DES/IDS level is much higher in
urine than other body fluids such as plasma and sputum.5,6 Despite
being less directly linked to the lung pathophysiological processes,
urinary DES/IDS is known to be significantly elevated in patients
with cystic fibrosis,7,8 COPD,6,8-12 alpha-antitrypsin deficiency,6,8
and bronchiectasis.8 It has also been shown that elevation of
urinary DES and IDS can be caused by smoking in healthy
individuals.8,12 Given the importance of elastin degradation in lung
pathologies and the inhibition of proteases such as metallopro-
teases being one of the promising strategies for treating respira-
tory diseases;13 there is a strong need for a reliable method for
the quantification of urinary DES/IDS as a disease progression
biomarker or a pharmacodynamics biomarker for compound
efficacy.
Several methodologies for monitoring urinary DES/IDS in-
cluding enzymatic immunoassay (ELISA),11,14 radioimmuno-
assay,15,16 capillary electrophoresis,8,9,17,18 HPLC isotope dilu-
tion,7,12,19,20 electrokinetic chromatography,21 and liquid chroma-
tography mass spectrometry (LC-MS/MS)5,6,22 have been devel-
oped in the last two decades. However, the urinary total DES/
IDS levels reported in the literature using these different
technologies have been inconsistent. In nonsmoking healthy
volunteers, for instance, average urinary total DES/IDS levels
ranged from 10-16 ng/mg creatinine using LC-MS/MS (with22
(3) Demedts, I. K.; Brusselle, G. G.; Bracke, K. R.; Vermaelen, K. Y.; Pauwels,
R. A. Curr. Opin. Pharmacol. 2005, 5, 257–263.
(4) Lagente, V.; Boichot, E. J. Mol. Cell Cardiol. 2010, 48 (3), 440–444.
(5) Ma, S.; Lieberman, S.; Turino, G. M.; Lin, Y. Y. Proc. Natl. Acad. Sci. U. S. A.
2003, 100, 12941–12943.
(6) Ma, S.; Lin, Y. Y.; Turino, G. M. Chest 2007, 131, 1363–1371.
(7) Stone, P. J.; Konstan, M. W.; Berger, M.; Dorkin, H. L.; Franzblau, C.;
Snider, G. L. Am. J. Respir. Crit. Care Med. 1995, 152, 157–162.
(8) Viglio, S.; Iadarola, P.; Lupi, A.; Trisolini, R.; Tinelli, C.; Balbi, B.; et al.
Eur. Respir. J. 2000, 15 (6), 1039–1045.
(9) Annovazzi, L.; Viglio, S.; Perani, E.; Luisetti, M.; Baraniuk, J.; Casado, B.;
et al. Electrophoresis 2004, 25, 683–691.
(10) Boschetto, P.; Quintavalle, S.; Zeni, E.; Leprotti, S.; Potena, A.; Ballerin, L.;
et al. Thorax 2006, 61, 1037–1042.
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010 3745
or without internal standard5,6) or using HPLC isotope dilution
methods (e.g., ref 12), 43-55 ng/mg creatinine using capillary
electrophoresis with laser induced fluorescence detection,9,18 and
102 ng/mg creatinine using ELISA.11 Among these methodolo-
gies, LC-MS/MS methods are believed to provide the best
specificity and its wide availability in clinical laboratories makes
it a potential “method of choice” for urinary DES/IDS measure-
ment. In addition, LC-MS/MS methods also provide superior
sensitivity, which is sufficient to quantify free DES/IDS in human
urine samples, or total DES/IDS in sputum.23 Despite better
sensitivity and specificity, one significant issue of the current LC-
MS/MS methods is the lack of internal standards for compensat-
ing potential matrix effects23 as such accuracy of quantification
is affected. Not until recently, Thibault et al.,22 developed a
sensitive nanoflow LC-MS/MS method with a significant improve-
ment by using a deuterated DES as internal standard. However,
the involvement of a derivatization step and the use of nanoflow
LC in this method make it difficult to be implemented in routine
clinical studies. In the present study, we describe the validation
of an LC-MS/MS method for the quantification of total urinary
DES/IDS using a deuterated DES as an internal standard.
EXPERIMENTAL DETAILS
Chemicals. DES (99.5%) was purchased from Elastin Products
Company Inc. (Owensville, MO). D5-DES Internal Standard was
obtained from Wyeth Pharmaceuticals, LC-MS grade acetonitrile
(ACN), and pentafluoropropionic acid (PFPA) were from Sigma-
Aldrich. (Steinheim, Germany). Acetic acid, butanol, 37% HCl and
tetrahydrofuran (THF) were from VWR. Finally, CC31 micro-
granular cellulose was purchased from Whatman International
Ltd., (Kent, UK).
Equipment and Materials. TSQ Quantum Ultra Mass Spec-
trometer (Thermo Scientific, Hemel Hempstead, UK) HPLC
Ultimate 3000 with autosampler (Dionex, Camberley, UK). Scan-
vac Coolsafe SpeedVac Model (Scanlaf, Denmark). Pursuit C18
HPLC Column, 3 µm, 50 × 2.0 mm and Bond Elute Reservoir with
two 20-µm frits were from (Varian, UK).
Clinical Sample Collection. Random spot urine samples from
patients with COPD and healthy volunteers (age and gender
(11) Cocci, F.; Miniati, M.; Monti, S.; Cavarra, E.; Gambelli, F.; Battolla, L.; et
al. Int. J. Biochem. Cell Biol. 2002, 34, 594–604.
(12) Stone, P. J.; Gottlieb, D. J.; O’Connor, G. T.; Ciccolella, D. E.; Breuer, R.;
Bryan-Rhadfi, J.; et al. Am. J. Respir. Crit. Care Med 1995, 151, 952–959.
(13) Hu, J.; Van den Steen, P. E.; Sang, Q. X.; Opdenakker, G. Nat. Rev. Drug
Discovery 2007, 6, 480–498.
(14) Luisetti, M.; Sturani, C.; Sella, D.; Madonini, E.; Galavotti, V.; Bruno, G.; et
al. Eur. Respir. J. 1996, 9, 1482–1486.
(15) McClintock, D. E.; Starcher, B.; Eisner, M. D.; Thompson, B. T.; Hayden,
D. L.; Church, G. D.; et al. Am J Physiol. Lung Cell Mol. Physiol. 2006,
291, L566–71.
(16) Starcher, B.; Green, M.; Scott, M. Respiration 1995, 62, 252–257.
(17) Fiorenza, D.; Viglio, S.; Lupi, A.; Baccheschi, J.; Tinelli, C.; Trisolini, R.; et
al. Respir. Med. 2002, 96, 110–114.
(18) Stolk, J.; Veldhuisen, B.; Annovazzi, L.; Zanone, C.; Versteeg, E.; van
Kuppevelt, T.; et al. Respir. Res. 2005, 6, 47–53.
(19) Stone, P. J.; Bryanrhadfi, J.; Lucey, E. C.; Ciccolella, D. E.; Crombie, G.;
Faris, B.; et al. Am. Rev. Respir. Dis. 1991, 144, 284–290.
(20) Cumiskey, W. R.; Pagani, E. D.; Bode, D. C. J. Chromatogr., B 1995, 668,
199–207.
(21) Viglio, S.; Zanaboni, G.; Luisetti, M.; Trisolini, R.; Grimm, R.; Cetta, G.;
Iadarola, P. J. Chromatogr., B 1998, 714, 87–98.
(22) Boutin, M.; Berthelette, C.; Gervais, F. G.; Scholand, M. B.; Hoidal, J.;
Leppert, M. F.; et al. Anal. Chem. 2009, 81, 1881–1887.
(23) Kaga, N.; Soma, S.; Fujimura, T.; Seyama, K.; Fukuchi, Y.; Murayama, K.
Anal. Biochem. 2003, 318, 25–29.
3746 Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
matched) were collected. All study participants gave their written
informed consent and the original study was approved by an
Institutional Review Board. Urine was collected in a polystyrene
urine collection bottle (ThermoFisher), aliquoted and frozen at
-80 °C until analysis.
Creatinine Analysis. The creatinine levels were measured for
each urine sample using a validated LC-MS/MS method as
published previously.24 Creatinine levels were 1.38 ± 1.00 mg/
mL in healthy volunteers and 1.22 ± 0.62 mg/mL in patients with
COPD in the first cohort. In the second cohort, creatinine levels
were 1.18 ± 0.64 mg/mL in healthy volunteers and 1.27 ± 0.90
mg/mL in patients with COPD in the first cohort.
Preparation of QC Samples. Approximately 500 mL urine
was pooled from a total of five volunteers. 150 mL urine was put
into three 250 mL plastic bottles for Low, Mid and High QCs.
For Mid and high QCs, 150 µL DES (10 µg/mL for Mid QC; 20
µg/mL for High QC in Milli-Q water) was added to the bottles in
such to give a final spiked-in concentration of (10 ng/mL +
endogenous for Mid QC; 20 ng/mL + endogenous for High QC).
Please note that after a 5 fold concentration step following SPE,
drying and re-suspension, the Mid QC and High QC concentra-
tions are (50 ng/mL + endogenous DES) and (100 ng/mL +
endogenous DES). After roller mixing, each of the QC samples
was aliquoted into 10 mL tubes and stored at -80 °C.
MS and MS/MS Analysis for D5-DES. 417 ng IS dissolved
in Milli-Q water was infused into a Thermo Quantum Ultra mass
spectrometer at the flow rate of 5 µL/min. For MS/MS analysis,
the precursor ion at m/z ) 531 was selected for fragmentation
with argon gas pressure set at 1.5 mTorr and collision energy set
at 33 V.
LC-MS/MS Analysis of D5-DES. DES (1 ng) together with
D5-DES (0.4 ng) was analyzed according to the method
described in LC-MS/MS analysis (below) with minor modifica-
tions in the LC condition in order to separate DES from IDS.
The LC condition used was as follows: Mobile phase A: 0.1%
PFPA/water, and Mobile phase B: 0.1% PFPA/ACN. HPLC was
performed using a 5 min linear gradient flow of the mobile
phase B from 0 to 8.5% for 5 min. The flow rate was 0.25 mL/
min. The temperature of the HPLC column was set at 30 °C.
Measurement of Urinary Total DES/IDS. The procedure
consisted of three steps: sample hydrolysis, solid phase extraction,
and LC-MS/MS analysis. The details are described as below.
Sample Hydrolysis. 250 µL of each urine sample plus 50 µL
of D5-DES 400 ng/mL was added to 250 µL of 11.8 N HCl (24
h, 108 °C) to liberate DES/IDS covalently bound to DES/IDS
containing peptides.
Solid Phase Extraction. A 3 mL Bond Elute reservoir was
packed with 100 mg CC31 microgranular cellulose and pressed
to get the cellulose into a uniform compact layer. Each hydrolyzed
sample (0.55 mL) was mixed with 2.5 mL (butanol/acetic acid )
4:1) before being applied to a SPE column previously conditioned
with 3.4 mL of Milli-Q water and 3.4 mL butanol solution (butanol/
acetic acid/water ) 4:1:1). The wash and elution of bound analytes
were performed using a Gilson Aspec GX-271 liquid handling
robot. First, the SPE cartridge was washed with 4 mL mobile
phase followed by a 2 mL air push. The mobile phase in the SPE
(24) Takahashi, N.; Boysen, G.; Li, F.; Li, Y.; Swenberg, J. A. Kidney Int. 2007,
71, 266–271.
Figure 1. Assay procedure for urinary total DES/IDS and examples of ion chromatograms from low-, mid-, and high-QCs. (A): The assay
procedure for measuring urinary total DES/IDS is comprised of five major steps including sample hydrolysis, solid phase extraction, drying/
resuspension, HPLC and MS/MS. A brief description of each step is listed each box. (B): Typical LC-MS/MS chromatograms of low-, mid-, and
high-QCs where IS (D5-DES) traces are shown in red and DES/IDS traces are in black.
was dried by applying 0.5 mL THF followed by 1 mL air push.
Finally, the analytes were eluted with 0.5 mL Milli-Q water
followed by 2 mL air push to ensure completion of extraction.
The eluates were dried in a SpeedVac at 45 °C then the residues
were resuspended in 50 µL of 2% PFPA in Milli-Q water.
LC-MS/MS Analysis. MRM was performed on a HPLC
coupled to a TSQ Quantum Ultra (Thermo Fisher Scientific) triple
quadrupole mass spectrometer in positive ion mode. For each
sample, 10 µL was injected onto a Pursuit C18 column (3 µm, 50
× 2.0 mm; Varian) using an Ultimate 3000 HPLC system
(Dionex), where the mobile phase A is aqueous with 0.1%
formic acid, and the mobile phase B is 0.1% formic acid in
acetonitrile. HPLC was performed using a 2 min isocratic
gradient at a flow rate of 0.35 mL/min. The temperature of
the HPLC column was set at 30 °C. Quantification was
performed by simultaneous monitoring of transitions (DES)
m/z 526.3f481.1 and (D5-DES) m/z 531.3 f 486.1 with
collision energy set at 33 eV for both transitions, collision gas
pressure was 1.5 mTorr, tube lens at 151 V, transfer capillary
temperature at 350 °C, sheath and auxiliary gas (nitrogen)
pressure at 60 and 10 arbitrary units and ion spray voltage at
4 kV. The scan time was set at 200 msec and both quadrupoles
(Q1 and Q3) were at 0.7 Da FWMH.
Data Analysis. Peak area under curve for DES/IDS and D5-
DES, calibration curve, SD, CV% and %bias were generated/
calculated by LCquan (ThermoFisher, Hemel Hampstead. UK)
or manually using Microsoft Excel 2003 SP3.
RESULTS AND DISCUSSION
Technically it is feasible to quantify DES and IDS separately
using LC-MS/MS.5,6,22,25 However, as both isomeric compounds
(25) Ma, S.; Lin, Y. Y.; Tartell, L.; Turino, G. M. Respir. Res. 2009, 10, 12–19.
are derived from mature elastin, its degradation can be more
accurately estimated by the combined level of DES and IDS. As
such, not only does it eliminate technical difficulties in separating
these two isomers, the requirement for assay sensitivity also
becomes lower and thus easier to implement in a clinical setting.
In this report, we set out to develop a urinary total DES/IDS assay.
We adapted and modified the previously published procedure
for sample preprocessing.26 A simplified workflow is summarized
in Figure 1A. In brief, urine samples are spiked-in with the internal
standard and acid-hydrolyzed at 108 °C for 24 h. This duration of
acid hydrolysis was determined by a time course study where
we found that the level of total DES/IDS plateaued between 18-24
h of acid hydrolysis and longer acid hydrolysis duration increased
variability (Supporting Information (SI) Figure S-1). The hydro-
lyzed urine samples are then processed through cellulose SPE
cartridges in order to enrich DES/IDS and reconstituted in 2%
PFPA (an ion pairing reagent to help DES/IDS retain in a reverse
phase column) prior to LC-MS/MS analysis using an MRM mode.
The mobile phases in the LC step do not contain any ion pairing
reagent so DES and IDS are not resolved into separate peaks.
The transitions for DES/IDS and D5-DES are 526/481 and 531/
486, respectively. Examples of ion chromatograms are shown
in Figure 1B.
Characterization of DES Internal Standard. Deuterated
forms of DES were chemically synthesized and their structure,
physical/chemical properties and the level of D0-DES were
investigated. MS/MS analysis of DES and D5-DES showed that
these two compounds share the same MS/MS fragmentation
patterns (Figure 2A), confirming that these two compounds are
structurally identical with the exception of the replacement of
(26) Pratt, D. A.; Daniloff, Y.; Duncan, A.; Robins, S. P. Anal. Biochem. 1992,
207, 168–175.
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010 3747
Figure 2. Characterization of D5-DES (A). Both DES and D5-DES were infused to the mass spectrometer separately (see the Experimental
Section for the details) and MS/MS spectra were acquired for the precursor ions (m/z ) 526 for DES and 531 for D5-DES). The fragmentation
pattern of D5-DES was identical to that of DES except the mass shift due to deuterium atoms. (B). DES and D5-DES were coeluted in a LC
gradient. DES (1 ng) together with D5-DES (0.4 ng) was analyzed as described in (LC-MS/MS analysis of D5-DES). (C). A typical standard
curve ranges from 1 to 300 ng/mL.

deuterium atoms as indicated by the mass differences in fragment
peaks. The most abundant fragment ion at 481 for DES (and 486
for D5-DES) representing [M-COOH]+ ion were selected for
quantification using multiple reaction monitoring mode. LC-
MS/MS analysis indicated that D5-DES was coeluted with DES
using a gradient containing an ion pairing reagent PFPA in
LC mobile phase (where DES and IDS can be separated, Figure
2B), again suggesting that they share a similar chemical property.
There was virtually no deuterated IDS in the synthetic D5-DES.
As the assay procedure for urinary total DES/IDS contains
an acid hydrolysis step, it is critical to test if the deuterium atoms
in D5-DES are stable under the hydrolysis condition. D5-DES
was treated at RT or 108 °C in the presence or absence of 6 M
HCl. Our results show that treatment of D5-DES on its own at
108 °C in the presence of 6 M HCl caused a significant decrease
in the signal (Figure S-2A), probably due to free radical induced
chlorination.27 Similar results were observed on DES with the
same experiment setup (data not shown). However, such a
decrease was not observed when D5-DES was added into urine
samples before acid hydrolysis. This was demonstrated in a
recovery study where D5-DES signals of the “before hydrolysis
spiked-in” QC samples did not differ significantly from those
of the “after hydrolysis spiked-in” QC samples (SI Figure S-2C).
The acid/high temperature effects on DES and D5-DES is likely
to be compensated by the existence of other proteins and/or
(27) Davidson I. Protein Sequencing Protocols; Springer: London, 1996; Vol. 64,
pp 119-129.
3748 Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
metabolites and we did not observe a significant effect to the
accuracy of measurement in the assay.
To quantify the level of D0-DES/IDS in the internal standard,
0.1 µg IS (in a volume of 10 µL) was injected to LC-MS/MS
six times and the content of D0-DES/IDS was estimated to be
1.2 ng/µg IS (SI Figure S-3). The contamination of D0-DES from
the IS is much lower than the endogenous DES/IDS level in
urine and therefore is not considered to affect the results.
Limits of Quantification. The calibration curve for DES was
performed between 1.0 and 300.0 ng/mL. The intra-assay %CV
and %bias for the lower end of calibrators (i.e., 1 and 2 ng/mL)
failed to achieve the acceptable specification limit (25%). The intra-
assay precision (%CV) and accuracy (%bias) of the three analytical
runs (n ) 3) for each of the other seven calibrators (5-300 ng/
mL) are shown in SI Table S-1 with %CV ranging from 1.4 to 19.2%
and %bias from -3.5 to 12.5%. The interassay precision (%CV) and
accuracy (%bias) of the calibrators evaluated over three different
days are shown in Table 1. The %CV of 5.0-300 ng/mL calibrators
ranged from 0.6 to 5.3% and %bias from -1.3 to 9.3%. A typical
example of calibration curve over the range of calibrators is shown
in Figure 2C.
Sample Dilution Linearity and Reportable Range. Sample
dilution recovery and linearity was assessed using three urine
samples from two healthy individuals and a patient with COPD
over a dilution ratio from 1:2 to 1:32. Dilution of urine at ratios of
1:2, 1:4, and 1:8 produced good linearity between measured DES/
Table 1. Inter-Assay Precision and Accuracy of the
Validation Samples

LOW QC MID QC HIGH QC

(endogenous (endogenous + (endogenous +
only) 50 ng/mL) 100 ng/mL)
mean measured 41 88 140
concentration
(ng/mL)
expected N/A 91 141
(ng/mL)a
SD 2 1 4
%CV 4.9 1.1 2.9
%bias N/A -3.3 -0.7
% recovery N/A 96.7 99.3
a
Expected concentrations in MID or HIGH ) mean measured
concentration of LOW + spiked concentration.

Table 2. Demographic Details of the First and Second

Cohorts of Healthy Volunteers and Patients with COPD
for Urinary Free and Total DES Measurements

first cohort second cohort

healthy COPD healthy COPD
volunteers patients volunteers patients
age 52 ± 8 57 ± 7 31 ± 16 67 ± 10
gender (male/female) 8/10 7/11 8/12 12/8
smoking statusa 2/1/15/0 14/2/2/0 4/0/12/4 10/4/2/4
BMI 25 ± 4 26 ± 8 27 ± 5 28 ± 9
FEV [% predicted] 102 ± 12 37 ± 13 107 ± 15 43 ± 20
a
Smoker/ex-smoker/nonsmoker/not available.

IDS concentrations and calculated DES/IDS concentrations (SI

Figure S-4; r2 ) 0.9993). Accuracy of measured concentrations
in diluted samples was also assessed. The %bias of measured
DES/IDS concentrations in diluted samples (compared with
calculated DES/IDS concentrations relative to undiluted con-
centration) ranges from -19.9 to 0.2% for dilution ratio from
1:2 to 1:8 (SI Table S-2). The 1:16 dilution (in the third sample)
showed a poor %bias, thus >8 dilution ratio is not recommended.
These results suggest that the reportable concentrations for
urinary total DES/IDS may go up to approximately 480 ng/mL
of DES/IDS following an 8-fold dilution when necessary.
Analytical Performance: Precision and Accuracy/Recov-
ery. The intra- and interassay precision and accuracy/recovery
of three validation samples (low, mid, and high QC samples) were
determined in triplicate per day, during three different days. The
intra-assay imprecision (%CV) and accuracy (%bias) for total DES/
IDS ranged from 5.0 to 16.4 and -11.1 to 0.3, respectively (SI
Table S-3). The interassay imprecision and accuracy percentages
ranged from 1.6 to 5.1 and from -1.1 to -0.2, respectively. The
total recovery of the assay was also analyzed through the analysis
of mid- and high-validation samples. Intra- and interassay recovery
for total DES/IDS ranged from 88.9 to 99.0% and 98.9 to 99.8%,
respectively. SPE recovery was also assessed using urine samples
spiked with mid and high levels of DES prior to or after SPE.
The percentage of total DES/IDS recovery ranged from 109.6 to
111.6% (SI Table S-4).
Stability. Short-term stability (up to 20 h at RT and at 4 °C),
as well as stability following 3 F/T cycles (using freshly collected
urine stored at 4 °C for no more than 4 h) was assessed using
three urine samples (run in triplicate) from apparently healthy
Figure 3. Urinary total DES/IDS as a biomarker for COPD (A) The
level of urinary total DES/IDS in COPD patients were about 3 times
higher than those in healthy volunteers (p ) 1.2 × 10-5;
t test). The boundary of the box closest to zero indicates the 25th
percentile, a line within the box marks the median, and the boundary
of the box farthest from zero indicates the 75th percentile. Whiskers
(error bars) above and below the box indicate the 90th and 10th
percentiles. (B) Except one patient with COPD, the urinary total DES/
IDS levels were well correlated to the urinary free DES/IDS levels
(R2 ) 0.85, Pearson’s correlation).
donors. No significant reduction of DES/IDS was observed
following storage at room temperature (20 h, 98.9% recovery), or
4 °C (20 h, 100.7% recovery) or after three freeze/thaw cycles
(106.5% recovery) (SI Table S-4).
Creatinine Normalized Urinary Total DES/IDS Levels in
Healthy Individuals and COPD Patients. To assess the utility
of this method for measuring total urinary DES/IDS as a
biomarker for COPD, we analyzed a total of 36 urine samples
(random urine samples) including 18 COPD patients and 18 age
and gender matched apparently healthy donors (Table 2, first
cohort). The levels of urinary total DES/IDS in all healthy
volunteers and COPD patients were above the LLOQ ranging
between 1.1 and 49.4 ng/mL. The level of total DES/IDS was
normalized to the levels of urinary creatinine. For healthy
volunteers, the level of urinary total DES/IDS were 6.8 ± 3.2 ng/
mg creatinine (mean ± SD). Urinary total DES/IDS in COPD
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010 3749
patents were about 3 times higher than those in healthy volunteers
(COPD: 20.4 ± 14.9 ng/mg creatinine; p ) 1.2 × 10-5; t test)
(Figure 3A). Despite a small number of healthy smokers in the
study, it was noticed that four healthy smokers had higher urinary
total DES levels (10.8 ± 3.1 ng/mg creatinine) creatinine compared
to nonsmokers and ex-smokers among the healthy group (5.6 ±
2.1 ng/mg creatinine). The smoking related elevation of urinary
total DES was in line with previous reports.7,8 In comparison to
the results from other studies using LC-MS-based methods, the
results for healthy volunteers reported in this study was lower
compared to the results from Boutin and colleagues (13.2 ± 1.9
ng/mg creatinine for healthy smokers and 14.9 ± 2.9 ng/mg for
healthy nonsmokers22), in which a stable isotopic desmosine
standard was also included. In contrast, we also found that the
levels of urinary total DES/IDS in COPD patients were twice as
high as those reported in the same study.22 The reason for the
discrepancy is not clear, but it is likely due to the small size of
cohorts, the nature of spot urine samples used, methods used to
measure creatinine, and/or different disease severity of patients
recruited in the studies. A larger cohort with whole day urine
samples (or multiple visits if spot urine samples are preferred) is
urgently required to establish the validity of this marker for
disease progression and pharmacodynamic monitoring for com-
pounds inhibiting proteases that cause lung degradation.
We recognized that one potential limitation of applying this
method in clinical chemistry laboratories is the requirement of
acid hydrolysis, where it requires 24 h for sample preparation and
the use of 12 M hydrochloric acid adds another degree of health
and safety requirements. As free DES and IDS has been shown
to be detectable in urine samples,6 we also investigated if free
DES/IDS levels can be used to estimate elastin degradation. To
do so, a second cohort consisting of a total of 40 samples (20
COPD patients and 20 healthy volunteers, see Table 2, second
cohort, for the demographics) were processed and analyzed in
the same method procedure except that the acid hydrolysis step
was omitted. Two free desmosine levels were below LLOQ and
were excluded. We found that except one patient with COPD, the
urinary total DES/IDS levels were well correlated to the urinary
free DES/IDS levels (R2 ) 0.85, Pearson’s correlation, Figure
3750 Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
3B). However, the differences between the two groups in terms
of free DES/IDS were not as significant as the total DES/IDS
(free DES/IDS: p ) 0.001 vs total DES/IDS: p ) 1.2 × 10-5).
There was a substantial overlap between the two groups in
urinary free DES/IDS levels. These results suggests that
despite urinary free and total DES/IDS levels being highly
correlated, it is preferred to measure urinary total DES/IDS
level as a biomarker for elastin degradation for COPD.
Although the scope of the current study focuses on urinary
DES/IDS, it is perceivable that total DES/IDS measurements in
other body fluids such as sputum (to a lesser degree, plasma)
may reflect more directly to the lung pathologies in respiratory
diseases. Some initial studies by Turino and associates have shown
some promising results.5,6 Further effort for analytical validation
of these methodologies and clinical validation of sputum/plasma
total DES/IDS as a biomarker for respiratory diseases are urgently
required.
CONCLUSION
In summary, we have developed and validated a fast and
reliable LC-MS method using stable isotopic DES as an internal
standard for measuring total urinary total DES/IDS. Our results
demonstrate that the LC-MS/MS method provides sensitive,
reproducible, and accurate quantification of total DES/IDS in
human urine samples as a biomarker for monitoring elastin
degradation in diseases such as COPD.
ACKNOWLEDGMENT
We thank Professor Simon Robins from University of Aberdeen
and Dr. Ian Kay from Thermo Scientific for their valuable
suggestions and technical support. The first two authors contrib-
uted equally to this work.
SUPPORTING INFORMATION AVAILABLE
Figures S-1-S-4, Tables S-1-S-5. This material is available free
of charge via the Internet at http://pubs.acs.org.
Received for review January 19, 2010. Accepted March 4,
2010.
AC100152F