methodological aspects on measurement of clara cell protein in urine as a biomarker for airway...

60 L. ANDERSSON ET AL.

Copyright © 2006 John Wiley & Sons, Ltd. J. Appl. Toxicol. 2007; 27: 60–66

DOI: 10.1002/jat

JOURNAL OF APPLIED TOXICOLOGYJ. Appl. Toxicol. 2007; 27: 60–66Published online in Wiley InterScience(www.interscience.wiley.com) DOI: 10.1002/jat.1184

Methodological aspects on measurement of Clara cellprotein in urine as a biomarker for airway toxicity,compared with serum levels

Lena Andersson,1,* Per-Arne Lundberg2 and Lars Barregard1

1 Department of Occupational and Environmental Medicine, Sahlgrenska University Hospital and Academy, GöteborgUniversity, Göteborg, Sweden

2 Clinical Chemistry, Sahlgrenska University Hospital, Göteborg, Sweden

Received 7 July 2006; Revised 22 September 2006; Accepted 25 September 2006

ABSTRACT: The Clara cell protein CC16, secreted from Clara cells in the lung, is discussed as a potential biomarker

for toxic effects on the airways. An increased concentration of CC16 in serum may be caused by increased permeability

of the lungs, caused by high levels of air pollution. Since CC16 is eliminated by renal excretion, it may be possible to use

urine instead of serum samples. Few studies have been conducted on urinary CC16 (U-CC16), however.

The aim was to investigate the optimal way of sampling and quantifying CC16 in urine samples and compare CC16

in human serum and urinary samples. Repeated sampling was performed in two groups of healthy subjects. First morning

urine, 24 h urine, and matched blood and urine samples were collected.

The excretion of U-CC16 increased over the day, e.g. from 0.08 µµµµµg h−−−−−1 in the morning to 0.28 µµµµµg h−−−−−1 in daytime and

0.16 µµµµµg h−−−−−1 in the evening (medians in males). Morning samples (µµµµµg h−−−−−1) from males properly reflected the 24 h excretion

(r ===== 0.91). The best correlation with 24 h excretion was obtained with creatinine-corrected first morning urine samples

(r >>>>> 0.9). Generally, females had lower excretion of CC16 than males (medians 2.5 µµµµµg 24 h−−−−−1 in females and 16 µµµµµg 24 h−−−−−1

in males). There was significant intraindividual variation, but the interindividual variation was larger in both groups.

There was an association between serum CC16 (S-CC16) and U-CC16 in morning samples. With optimal methods for

sampling U-CC16, urine samples may be used in experimental studies of air pollution. Copyright © 2006 John Wiley &

Sons, Ltd.

KEY WORDS: Clara cell protein; CC16; methodological study; urine sampling

* Correspondence to: L. Andersson, Department of Occupational and

Environmental Medicine, Sahlgrenska University Hospital and Academy,

P.O. Box 414, S-405 30 Göteborg, Sweden.

E-mail: [email protected]

excretion before U-CC16 can be used as a biomarker. In

the literature (e.g. Bernard et al., 1993), several factors

have been shown to affect the levels of CC16 in urine.

The most important are CC16 originating from the pros-

tate (which also secretes CC16), renal function and preg-

nancy. The first portion of urine from men contains CC16

from the prostate (Bernard et al., 1992). The size of the

portion affected by postrenal excretion has not been

quantified. The excretion from the prostate must be elimi-

nated from the urine sample before it can be used as a

surrogate for serum.

When analysing urinary proteins, it is an advantage if

the excretion rate (µg h−1) can be quantified from prefer-

ably a 24 h collection. However, an entire 24 h period of

urine collection can be problematic in field studies. Nor-

mally, spot urine is used and the concentration is often

corrected for creatinine or specific gravity (SG) to com-

pensate for the dilution (Carrieri et al., 2001). No data

were found on the association between creatinine-

corrected levels of CC16 and excretion rates, the variabil-

ity in the excretion of CC16 in urine within and between

individuals, and storage stability. The aim of this study

was to evaluate optimal methods of determining CC16 in

urine by exploring the impact of the postrenal excretion,

Introduction

The Clara cell protein CC16 is secreted from the Clara

cells to the epithelial lining fluid of the lung and is

transported up along the bronchiolar tree. Hermans and

Bernard (1999) report that CC16 can leak to serum

through the lung epithelial barrier. It has been shown that

high levels of air pollution may increase the permeabil-

ity of the lung epithelium (Broeckaert et al., 2000). Con-

sequently, serum CC16 (S-CC16) has been used as a

non-invasive marker in different studies of effects of air

pollution on the airways. Recently, the use of urinary

CC16 (U-CC16) was investigated as well (Timonen

et al., 2004).

In large-scale studies, urine sampling offers several

advantages compared with serum sampling. The non-

invasive urine samples are more readily accepted and the

costs are lower. It is, however, necessary to clarify the

renal handling of the protein and factors affecting its

METHODOLOGICAL ASPECTS FOR MEASUREMENT OF CC16 61


DOI: 10.1002/jat

first 100 ml of the sample before collecting the urine.

This volume was taken into account when the total vol-

ume was recorded. Replicate samples (5 ml) were taken

directly from all fresh portions and transferred to

polypropylene centrifugation tubes (Sarstedt Australia

Pty. Ltd., Ingle Farm, Australia) and stored cold (4 °C)

until analysis (max 7 days). Samples used for stability

control were immediately frozen (−20 °C) and analysed

after 2 weeks and after 2 months. Samples from day 1

were used for the study of postrenal excretion. Samples

from day 2 were used for examination of the variability

within days. First morning samples from day 1 and 2

were used for calculation of variability between days. All

samples from day 1 and 2 were used for a study of the

association between excretion rates and concentrations in

samples corrected for creatinine and SG, respectively.

Group B. On day 1, one sample was taken from the

first morning urine. The second sample was taken after

6 h, and another, from the next morning’s first urine

(Table 1). The samples were used for studies of stability,

variability within subjects between days, and associations

between S-CC16 and U-CC16. Samples used for stabil-

ity control were frozen (−20 °C) and analysed after

6 months.

Blood Sampling

Group B. The first blood sample was taken 2 h after

the first morning urine sample. The second serum sample

was taken in the afternoon, and the third, on the next

morning 2 h after the urine sample (see Table 1). The

whole procedure was repeated 1 week later. For 13 of

the 24 subjects (six males and seven females), blood

samples was also taken at noon (Fig. 4). All blood

samples were collected by venipuncture in vacuum tubes

(BD Vacutainer™, Franklin Lakes, NJ, USA). The serum

samples were used to study the association between

matched blood and urine samples, and the variability of

S-CC16 within and between days.

Determination of Serum and Urinary CC16

Analyses of CC16 were performed using the Human

Clara Cell Protein ELISA kit from BioVendor

(BioVendor Laboratory Medicine, Inc., Brno, Czech

Republic). Standard and sample solutions were incubated

for 1 h in microtiter wells coated with rabbit polyclonal

antibodies against human CC16. After washing, biotin-

labeled antibodies against CC16 were added, followed

by thorough washing, and addition of Streptavidin-

horseradish-peroxidase. After 1 h incubation and a final

washing, the conjugate reacted with H2O2-tetramethyl-

benzidine. The reaction was stopped by adding an acidic

solution before absorbance was measured at 450 nm. The

Table 1. Urine (U) and serum (S) sampling for deter-mination of Clara cell protein in Group A and Group B

Group A Group B

Day 1 Day 2 Day 1 Day 2

Morning U U U, S U, S

Noon U U U, Sa U, Sa

Afternoon U S S

Evening U

Next morning U, S U, S

a Serum samples taken from 13 of the 24 subjects for the study of diurnal

S-CC16 variation. Morning samples were taken about 8 am, noon samples

about 2 pm and afternoon samples at about 5 pm.

and establishing the optimal time for sampling. Also

examined were the storage stability and the correlation

between matched samples of urine and serum.

Methods

Subjects

Group A. For a detailed study of U-CC16, 20 healthy (no

diabetes, hypertension, or airways or kidney diseases),

non-smoking subjects (ten female and ten male, mean

age 35 years, range 22–57 years) were examined.

Group B. To study CC16 in matched samples of serum

and urine, 24 healthy, nonsmoking subjects (13 female

and 11 male, mean age 37 years, range 21–57 years)

were recruited.

All subjects were recruited from the staff of the

Department, and they filled in a questionnaire about

age, smoking habits, and possible diseases or medication.

The study was approved by the Ethical Committee of

Göteborg.

Postrenal Excretion

All male subjects in group A separated their first urine

sample in three portions of 50 ml each and a fourth por-

tion with the remaining urine. This allowed quantification

of the postrenal excretion in every portion.

Urine Sampling

Group A. On day 1, one sample was taken from the first

morning urine. Another sample was taken before noon,

according to the scheme in Table 1. During day 2, sam-

ples from every urine portion during 24 h were collected

and classified as first morning, noon, afternoon, and

evening samples. All urine samples were timed, and the

total volumes recorded. After the assessment of postrenal

excretion (see Results), the male subjects discarded the



DOI: 10.1002/jat

absorbance is proportional to the con-centration of

CC16. Calibrators with target values in the ranges 5.0–

9.4 µg l−1 and 15.8–23.7 µg l−1 were provided in each kit

and used at the start and at the end of the assay in every

run. Our results were always within the accepted range.

Precision, as assessed in duplicate samples within runs,

was considered acceptable with a coefficient of variation

of 11%.

For both serum and urine samples, 4 µl of undiluted

sample was mixed with 96 µl dilution buffer directly in

the wells. To evaluate samples with low U-CC16, new

analyses with less diluted samples were performed. The

limit of detection was <0.04 µg l−1, defined as such a con-

centration of Clara cell protein giving absorbance higher

than mean absorbance of blank: Ablank + 3 × SDblank. This

was tested in our own laboratory as well.

Determination of Creatinine and Specific Gravity

Analyses of creatinine were performed in fresh urine

using the Jaffé method (Roche Diagnostics, limit of

detection 0.01 mmol l−1). Specific gravity was measured

in fresh urine with a Ceti, Digit 012 refractometer

(Medline, Oxfordshire, UK).

Data Analysis

Excretion rates of CC16 were calculated from the

concentration, volume and sampling time for each urine

portion. Moreover, all urine samples were corrected for

creatinine and SG. Specific gravity calculations were

performed with SGstandard = 1.020. Precision (repeatability)

was assessed from analyses of duplicates from the same

urine or serum samples in the same run.

In the descriptive analyses performed, S-CC16 levels

were found to be approximately normally distributed,

while the distribution of U-CC16 was skewed. The vari-

ability in CC16 within individuals was examined and the

coefficient of variation (CV) was calculated on first

morning urine portions (excretion rates and creatinine-

corrected concentrations) and serum morning samples on

different days in each subject. The total variability in

S-CC16 and log-transformed U-CC16 was partitioned

within and between subjects using PROC NESTED (SAS

Institute, Inc., Cary, NC, USA). The CC16 excretion

rates as well as concentrations corrected for creatinine or

SG in first morning or daytime urine samples were com-

pared with the 24 h CC16 excretion using linear regres-

sion and Pearson’s correlation coefficient. The same

technique was used for comparison of matched serum

and urine samples. Statistical analyses were performed

using the PROC NESTED and PROC VARCOMP pro-

cedures in the SAS System, version 9.1 (SAS Institute,

Inc., Cary, NC, USA).

Results

Postrenal Excretion

In the majority of male subjects, CC16 concentrations

were higher in the first two urine portions (50 ml each).

Stabilization was reached in the third and fourth portion.

The CC16 level in the first portion was three to 200

times higher than the concentration in the third portion.

There was a statistically significant difference between

the second and third portion but no statistically significant

difference was detected between the third and fourth

portion. Therefore, the first 100 ml of each male urine

collection was discarded in all subsequent samples. Male

subjects were supplied with a plastic cup with the 100 ml

level marked. Thus the effect of CC16 originating from

the prostate was considered to be eliminated.

Excretion Rate

The CC16 excretion rate was significantly higher during

day and evening than during the night (Fig. 1). It

was more common for values to be below the limit of

detection in first morning urine samples than in day and

evening samples. The total 24 h excretion was much

higher in men (median 16 µg) than in women (median

2.5 µg) (not shown in the Figure).

Spot Urine Samples versus 24 h Collections

For male subjects, the highest correlation between the

excretion rate at different times of the day and the total

24 h excretion was seen in first morning samples (r =0.91, Fig. 2). The same correlation was not seen in

noon and evening samples (r = 0.41 and r = 0.63, respec-

tively). In female subjects, the correlation coefficients

were >0.98 for all times of the day. For creatinine-

corrected levels (µg g−1), the highest correlation with the

total 24 h excretion of U-CC16 (µg 24 h−1) was again

found in first morning samples (females r = 0.85, males

r = 0.98) compared with samples taken at noon (females

r = 0.64, males r = 0.67) (Fig. 3a and 3b, respectively).

The correlation was high also for the male subjects’ first

morning samples corrected for SG (females r = 0.16,

males r = 0.98).

Variability Within and Between Individuals

To study variability, first morning samples from two

different days were used in subjects (groups A and B)

with detectable values in both first morning samples

(n = 32). About 90% of the total variability was variabil-

ity between individuals and the remaining 10% was



DOI: 10.1002/jat

Figure 1. Median values of CC16 excretion (µg h−1), with 10, 25, 75– and 90 percentiles, in ten female andtenmale subjects at different times of the day

variability within individuals. Within-subject variability

can also be expressed as relative standard deviation

(RSD) in untransformed U-CC16, which was 63%

between days (59% and 42%, respectively, for male and

female subjects).

Figure 2. Association between excretion rates in firstmorning urine samples and the 24 h excretion in males(Pearson r = 0.91) and females (r = 0.98)

Precision and Storage

In 46 fresh urinary samples with a median concentration

of 0.9 µg l−1 (range 0.04–33.8 µg l−1), the CV for determi-

nation of CC16 in duplicates was 11%. To assess storage

stability, 40 samples were used. After both 2 weeks

and 2 months, no significant decrease in concentrations

was seen (point estimates 114% and 109% of the initial

concentrations), but the precision (reproducibility)

decreased (CV 22% and 26%, respectively) when paired

results from fresh and stored samples were compared. For

13 of the subjects in group B, urine samples were also

analysed after 6 months. Again, here was no significant

decrease in concentrations (point estimate 95%) but the

reproducibility was not satisfactory (CV 57%).

Variability of CC16 Levels in Serum

These calculations are based on the 13 subjects in group

B where a total of four serum samples was taken;

morning, noon, afternoon and the next morning, to study

diurnal variation. The level of serum CC16 decreased

over the day despite sex, i.e. in the opposite direction

compared with urine (Fig. 4). In the morning the con-

centrations ranged from 2.0 to 14.5 µg l−1 (mean 7.5, SD

2.7). The median S-CC16 was lowest in the noon



DOI: 10.1002/jat

Figure 3. Associations between creatinine-correctedvalues and the 24 h urinary CC16 (U-CC16) excretion(µg 24 h−1). Pearson correlation coefficients were 0.98(first morning) and 0.67 (noon) in males (a) and 0.85(first morning) and 0.64 (noon) in females (b)

samples. However, when calculating the individual

changes, the mean CC16 decrease was 12% in the second

serum sample (6 h after the morning sample) and 17% in

the third sample (9 h after morning sample). A separate

analysis of all 24 subjects showed a mean decrease

of 25% in the afternoon sample. The within-subject vari-

ability for S-CC16 in morning samples was 22% of the

total variability, 78% being variability between subjects.

Consequently, matched serum morning samples on differ-

ent days showed high correlations, CV 17%.

Urinary CC16 versus Serum CC16

Only 19 of 24 the subjects had values over the limit of

detection in both first morning urine samples. Also, one

Figure 4. Diurnal variation in serum levels of CC16,with median values and 10, 25, 75– and 90 percentiles,in 13 subjects

male subject had indications of renal tubular damage

(high excretion of low molecular weight proteins), his

data were not included. Therefore the calculations are

based on 18 subjects. There was a strong linear associa-

tion (r = 0.76, P = 0.0003) between the mean (n = 2)

CC16 excretion rates in the first morning urine samples

and the mean (n = 2) CC16 concentrations in serum

as shown in Fig. 5. There was no similar association

between the CC16 levels in daytime serum samples (9 h

after morning samples) compared with U-CC16 samples

(6 h after the morning urine sample), (r = 0.34, P = 0.12).

Note, however, that daytime serum and urine samples

were not perfectly matched in time.

Figure 5. Association (Pearson r = 0.76) between themean urinary CC16 (U-CC16) from two morning urinesamples and the mean serum CC16 (S-CC16) from twomatched morning serum samples



DOI: 10.1002/jat

with 24 h excretion was obtained with creatinine-

corrected first morning urine samples. In summary,

if spot urine samples are used, we recommend first

morning samples, either timed with calculation of excre-

tion rates per hour, or creatinine-corrected (or both). For

males, the first 100 ml should be discarded.

Serum CC16

In agreement with a previous study (Helleday et al.,

2006), a decrease in S-CC16 concentration during

daytime was found. By contrast, it was found that the

excretion of U-CC16 increased over the day. Possibly

this could be the reason for the decrease in S-CC16.

Morning serum samples from different days within

subjects showed relatively stable CC16 levels, which is

an advantage when S-CC16 is to be used as a marker in

epidemiological or experimental studies.

Urinary CC16 versus Serum CC16 as a Marker forEffects on the Airways

Although some female U-CC16 levels were below the

detection limit, there was a correlation between serum

and urinary levels of CC16 (Fig. 5). Consequently,

U-CC16 could be used as a surrogate for S-CC16. The

urinary levels were, however, less stable than serum

levels on different days within individuals and there are

still question marks surrounding postrenal excretion,

sampling time and diluted samples. The variation

between individuals was also large, probably larger than

it would be after exposure to air pollution. Therefore, to

compare a group of individuals, large populations will be

needed. On the other hand, since there is an association

between S-CC16 and U-CC16 levels, this opens up an

interest in further investigating the use of urine samples.

In experimental studies the variability within individuals

is more important. Power calculations were performed

with a paired t-test with a normal approximation for α =0.05 based on our geometric group mean in U-CC16

of 0.04 in 32 subjects (the t-test was applied to log-

transformed U-CC16 levels). There will be 80% power

to detect a 50% increase in geometric mean U-CC16 in

a group of individuals (n = 32) before and after exposure

to some factor.

Acknowledgement—Eva M. Andersson is acknowledged for statisticaladvice.

References

Arsalane K, Broeckaert F, Knoops B, Clippe A, Buchet JP, Bernard A.1999. Increased serum and urinary concentrations of lung Clara cellproteins in rats acutely exposed to ozone. Toxicol. Appl. Pharmacol.

159: 169–174.

Discussion

The present study shows both possibilities and limitations

in the use of U-CC16. It is an easy sampling method

and CC16 seems to be stable in frozen urine after

2 months of storage. The repeatability was good (CV

11%) for fresh samples, but the precision deteriorated

somewhat after storage. In thawed samples, the precipi-

tate was solved carefully and considered homogenous.

Although, since the amount used in the ELISA was

only 4 µl, it might be less homogenous and therefore

CV increased. The samples from group B, stored for

6 months, showed decreased levels of CC16. The excre-

tion should be quantified using either timed samples or

creatinine-corrected values. There was a major influence

of postrenal excretion in males. According to our results,

the first 100 ml should be eliminated from every male

urine portion. No previous data were found on the size of

the urine sample affected by postrenal excretion. In a

recent study on U-CC16, midstream samples were used

(Timonen et al., 2004).

Factors affecting Urinary Excretion of CC16

The majority of the female subjects had low U-CC16,

which has been observed previously (Bernard et al.,

1993; Timonen et al., 2004). The high CC16 levels

among some men may be explained by some CC16

originating from the prostate even after discarding the

first 100 ml. The excretion of U-CC16 in male subjects

increases with age, probably due to an increase in

postrenal excretion (Bernard et al., 1993), but after

discarding the first 100 ml, no similar association between

age and U-CC16 excretion was found. As mentioned

above, the excretion rate of CC16 has a diurnal variation.

The excretion increases during daytime before it

decreases in the night. One explanation may be an in-

crease in the glomerular filtration rate over the day

(Arsalane et al., 1999, Hansen et al., 2002; Kabanda

et al., 1995). Another is that the increased urinary flow

rate during daytime will decrease the reabsorption and

thereby increase the excretion.

How to Sample and Quantify Urinary CC16

The best correlation between the excretion rate in first

morning samples and the total 24 h excretion was seen in

male subjects (Fig. 2). In female subjects, the evaluation

was more difficult due to low values, some of which

were even below the detection limit. A comparison

between the total 24 h excretion and concentrations

corrected for SG and creatinine showed creatinine-

corrected concentrations to be superior. This was the case

for both female and male subjects. The best correlation



DOI: 10.1002/jat

Bernard A, Roels H, Lauwerys R, Witters R, Gielens C, Soumillion A,Van Damme J, De Ley M. 1992. Human urinary protein 1: evidencefor identity with the Clara cell protein and occurrence in respiratorytract and urogenital secretions. Clin. Chim. Acta 207: 239–249.

Bernard A, Thielemans N, Lauwerys R, Vandeleene B, Lambert AE.1993. The renal handling of protein 1 (Clara cell protein): effect ofage, sex and renal dysfunction. Contrib. Nephrol. 101: 66–70.

Broeckaert F, Clippe A, Knoops B, Hermans C, Bernard A, 2000. Claracell secretory protein (CC16): features as a peripheral lungbiomarker. Ann. NY Acad. Sci. 923: 68–77.

Carrieri M, Trevisan A, Battista Bartolucci G. 2001. Adjustment toconcentration-dilution of spot urine samples: correlation betweenspecific gravity and creatinine. Int. Arch. Occup. Environ. Health 74:63–67.

Hansen H, Hovind P, Jensen B, Parving H. 2002. Diurnal variations ofglomerular filtration rate and albuminuria in diabetic nephropathy.Kidney Int. 61: 163–168.

Helleday R, Segerstedt B, Forsberg B, Mudway I, Nordberg G, BernardA, Blomberg A. 2006. Exploring the time-dependence of serum claracell protein as a biomarker of pulmonary injury in humans. Chest

130: 672– 675.Hermans C, Bernard A. 1999. State of the art. Secretory proteins of

pulmonary epithelial cells: characteristics and potential applicationsas peripheral lung markers. Am. J. Respir. Crit. Care Med. 159: 646–678.

Kabanda A, Goffin E, Bernard A. 1995. Factors influencing serumlevels and peritoneal clearances of low molecular weight proteinsin continuous ambulatory peritoneal dialysis. Kidney Int. 48: 1946–1952.

Timonen KL, Hoek G, Heinrich J, Bernard A, Brunekreef B, De HartogJ, Hämeri K, Ibald-Mulli A, Mirme A, Peters A, Tittanen P, KreylingWG, Pekkanen J. 2004. Daily variation in fine and ultrafineparticulate air pollution and urinary concentrations of lung Clara cellprotein CC16. Occup. Environ. Med. 61: 908–914.

methodological aspects on measurement of clara cell protein in urine as a biomarker for airway...

Documents