ARTICLE pubs.acs.org/ac
Potential Effect of Diaper and Cotton Ball Contamination on NMR- and LC/MS-Based Metabonomics Studies of Urine from Newborn Babies Aaron M. Goodpaster, Eshwar H. Ramadas, and Michael A. Kennedy* Department of Chemistry and Biochemistry, Miami University, Oxford, Ohio 45056, United States
bS Supporting Information ABSTRACT: Nuclear magnetic resonance (NMR) and liquid chromatography/mass spectrometry (LC/MS) based metabonomics screening of urine has great potential for discovery of biomarkers for diseases that afflict newborn and preterm infants. However, urine collection from newborn infants presents a potential confounding problem due to the possibility that contaminants might leach from materials used for urine collection and influence statistical analysis of metabonomics data. In this manuscript, we have analyzed diaper and cotton ball contamination using synthetic urine to assess its potential to influence the outcome of NMR- and LC/MS-based metabonomics studies of human infant urine. Eight diaper brands were examined using the “diaper plus cotton ball” technique. Data were analyzed using conventional principal components analysis, as well as a statistical significance algorithm developed for, and applied to, NMR data. Results showed most diaper brands had distinct contaminant profiles that could potentially influence NMR- and LC/MS-based metabonomics studies. On the basis of this study, it is recommended that diaper and cotton ball brands be characterized using metabonomics methodologies prior to initiating a metabonomics study to ensure that contaminant profiles are minimal or manageable and that the same diaper and cotton ball brands be used throughout a study to minimize variation.
N
uclear magnetic resonance (NMR) and liquid chromatography/mass spectrometry (LC/MS) based metabonomics research aimed at human disease biomarker discovery has increased significantly over the past decade.1-4 The increase in popularity is due to the promise that metabonomics research could contribute to personalized health care, noninvasive diagnosis of diseases, earlier diagnosis of diseases that have high fatality rates, and potential biomarker discovery for diseases that currently are difficult to diagnose.5-7 Typically, either NMR or LC/ MS is performed, but since these two techniques are complementary, both techniques are performed when possible to achieve the most comprehensive screening of the entire metabolome. Wilson's group has demonstrated the utility of using both techniques together and illustrated that the two techniques applied in the same study allowed different aspects of the metabolome to be investigated.8-11 Metabonomics has shown great promise in identifying potential biomarkers for human diseases, specifically in human cancers,12,13 and metabonomics in oncology has recently been reviewed.14,15 Outside of oncology, metabonomics has been used less often to study other human diseases. The few reported studies include investigations of the effect of age, gender, and diurnal variation in human samples.16,17 Metabonomics has also been used to study mice metabolic profiles to detect changes that occurred based on age,8 gender,11 obesity,9 diet,18 and body weight loss.19 Diagnosis and treatment of children's diseases could benefit greatly from metabonomics research, specifically in the area of newborn children's diseases, potentially reducing the stress that r 2011 American Chemical Society
children must endure if biomarkers can be discovered and validated for predicting individual susceptibilities, early diagnosis, monitoring disease progression, monitoring treatment efficacy, or revealing new aspects of disease etiology. A significant issue for metabonomics investigations of newborn diseases is sample collection, especially urine sample collection. Urine is a potentially ideal biofluid for metabonomics studies because sample collection is minimally invasive. There have been two metabonomics studies on pediatric patients. The first investigated urine samples from children with nephrouropathies where the investigators were trying to identify distinct metabolic profiles associated with different nephrourological disorders in children.20 The children in this study were old enough to provide urine samples using a standard procedure of collecting the urine into a cup. The second study investigated bronchoalveolar lavage fluid from cystic fibrosis patients to determine if a distinct metabolic profile could be determined for varying levels of inflammation.21 Neither of these studies involved urine collection from babies. Several methods have been adopted for collection of urine samples including midstream specimens, clean catch specimens, urine collection bags, suprapubic catheter, and from disposable diapers.22 All these methods have considerable drawbacks, are difficult, or are impractical when trying to collect urine from Received: September 29, 2010 Accepted: December 6, 2010 Published: January 6, 2011 896
dx.doi.org/10.1021/ac102572b | Anal. Chem. 2011, 83, 896–902
Analytical Chemistry
ARTICLE
newborn babies. Collection of urine from disposable diapers seems to be the preferred way of collecting urine from a baby. However, since manufactures started producing ultraabsorbent diapers, it is now impossible to routinely collect urine samples directly from the diaper itself. A simple way to overcome this problem is to insert either a cotton ball or a cotton pad into the disposable diaper, which allows urine to absorb in the cotton. Roberts and Lucas were the first to use this “diaper plus cotton ball” method of urine collection.23 They concluded that the collection procedure did not influence urinary concentrations of the constituents considered in their study. Since this paper, there have been several other studies conducted using this procedure or a slight modification involving use of a cotton pad.24-29 Despite the widespread use of the “diaper plus cotton ball” technique for urine sample collection in babies, there has been no reported investigations of the potential effect that diaper or cotton ball contamination could have on NMR- or LC/MSbased metabonomics studies, which is the goal of this paper. Here, eight diaper brands were examined using synthetic urine together with an experimental design intended to mimic urine collection in the human infant. Results indicated that most diaper brands exhibited distinct contaminant profiles that could potentially influence NMR- and LC/MS-based metabonomics studies. On the basis of this study, it is recommended that diapers and cotton ball brands be carefully characterized using metabonomics methodologies prior to initiating a study to ensure contaminant profiles are minimal or manageable and that the same diaper and cotton ball brands be used throughout a study to minimize unwanted variation.
needed, 250 mL of the concentrated stock solution was diluted to 2 L to give the final concentrations above. Sample Collection. For “cotton ball alone” samples, the cotton ball was placed into a sterile urine collection cup and 50 mL of SU was poured over the cotton ball. The lid was placed on the collection cup and the sample was incubated at 37 °C for 3 h. The “diaper alone” and “diaper plus cotton ball ” samples were collected by shaping the diaper into the collection cup, trimming excess diaper, and placing a cotton ball in the middle of the diaper for “diaper plus cotton ball” samples. An amount of 150 mL of SU was poured onto the center of the diaper or onto the cotton ball if present. After 1.5 h of incubation, an additional 150 mL of SU was added to the cups because the diaper absorbed all of the initially added SU. Samples were then allowed to incubate for another 1.5 h before collection. For “diaper plus cotton ball” samples, SU was collected by placing the cotton ball in the barrel of the syringe, inserting the plunger, and squeezing out the absorbed SU. About 4-7 mL of SU could be obtained using this technique. To collect “diaper alone” samples, a pipet was inserted into the center of the diaper and liquid sitting at the bottom was withdrawn. Diapers had to be slightly squeezed in order to obtain enough SU. Samples were put into a 15 mL sterile conical tube and stored in a -20 °C freezer until data collection. Sample Preparation. SU from the three groups were thawed on ice prior to preparation for NMR and LC/MS analysis. Details regarding sample preparation can be found in the Supporting Information. Nuclear Magnetic Spectroscopy. 1H NMR spectra were acquired at 298 K on a Bruker US2 Avance III 850 MHz spectrometer operating at 850.10 MHz equipped with a 5 mm TXI tripleresonance probe with inverse detection and controlled by TopSpin 2.1.4 (Bruker, Germany). Detailed methods used for NMR data collection are summarized in the Supporting Information. Liquid Chromatography/Mass Spectrometry. For LC/MS analysis, an Agilent Technologies 1200 rapid resolution system (Agilent Technologies, Palo Alto, CA, U.S.A.) consisting of a binary pump, temperature-controlled autosampler, column oven, and a variable wavelength UV detector was coupled to a Bruker micrOTOF mass spectrometer. Detailed methods used for LC/MS data collection are summarized in the Supporting Information. Principal Components Analyses. AMIX software (Bruker Biospin, Billerica, MA, U.S.A.) was used to perform PCA on NMR data. NMR spectra were binned into 0.01 ppm wide buckets over the region δ 10.0-0.15 ppm. The region around the water (δ 4.60-5.00) plus regions corresponding to resonances from known SU compounds were also excluded, as follows: δ 4.60-5.00, δ 4.05-4.07, δ 3.93-3.94, δ 3.851-3.855, δ 3.033.06, δ 3.97-3.98, δ 4.014-4.019, δ 4.14-4.15, and δ 5.606.00. Buckets with variances
