Renal cell carcinoma (RCC) accounts for 11,000 deaths per year in the United States. When detected early, generally serendipitously by imaging conducted for other reasons, long term survival is generally excellent. When detected with symptoms, prognosis is poor. Under these circumstances, a screening biomarker has the potential for substantial public health benefit. The purpose of this study was to evaluate the utility of urine metabolomics analysis for metabolomic profiling, identification of biomarkers, and ultimately for devising a urine screening test for RCC. Fifty urine samples were obtained from RCC and control patients from two institutions, and in a separate study, urine samples were taken from 13 normal individuals. Hydrophilic interaction chromatography-mass spectrometry was performed to identify small molecule metabolites present in each sample. Cluster analysis, principal components analysis, linear discriminant analysis, differential analysis, and variance component analysis were used to analyze the data. Previous work is extended to confirm the effectiveness of urine metabolomics analysis using a larger and more diverse patient cohort. It is now shown that the utility of this technique is dependent on the site of urine collection and that there exist substantial sources of variation of the urinary metabolomic profile, although group variation is sufficient to yield viable biomarkers. Surprisingly there is a small degree of variation in the urinary metabolomic profile in normal patients due to time since the last meal, and there is little difference in the urinary metabolomic profile in a cohort of pre- and postnephrectomy (partial or radical) renal cell carcinoma patients, suggesting that metabolic changes associated with RCC persist after removal of the primary tumor. After further investigations relating to the discovery and identity of individual biomarkers and attenuation of residual sources of variation, our work shows that urine metabolomics analysis has potential to lead to a diagnostic assay for RCC. Renal cell carcinoma (RCC) accounts for 11,000 deaths per year in the United States. When detected early, generally serendipitously by imaging conducted for other reasons, long term survival is generally excellent. When detected with symptoms, prognosis is poor. Under these circumstances, a screening biomarker has the potential for substantial public health benefit. The purpose of this study was to evaluate the utility of urine metabolomics analysis for metabolomic profiling, identification of biomarkers, and ultimately for devising a urine screening test for RCC. Fifty urine samples were obtained from RCC and control patients from two institutions, and in a separate study, urine samples were taken from 13 normal individuals. Hydrophilic interaction chromatography-mass spectrometry was performed to identify small molecule metabolites present in each sample. Cluster analysis, principal components analysis, linear discriminant analysis, differential analysis, and variance component analysis were used to analyze the data. Previous work is extended to confirm the effectiveness of urine metabolomics analysis using a larger and more diverse patient cohort. It is now shown that the utility of this technique is dependent on the site of urine collection and that there exist substantial sources of variation of the urinary metabolomic profile, although group variation is sufficient to yield viable biomarkers. Surprisingly there is a small degree of variation in the urinary metabolomic profile in normal patients due to time since the last meal, and there is little difference in the urinary metabolomic profile in a cohort of pre- and postnephrectomy (partial or radical) renal cell carcinoma patients, suggesting that metabolic changes associated with RCC persist after removal of the primary tumor. After further investigations relating to the discovery and identity of individual biomarkers and attenuation of residual sources of variation, our work shows that urine metabolomics analysis has potential to lead to a diagnostic assay for RCC. The study of all endogenously produced metabolites, known as metabolomics (or metabonomics), is the youngest of the omics sciences. It is becoming increasingly clear that, of all of the omics techniques, metabolomics has the greatest potential for biomarker discovery because this technique defines the signature of the actual processes that are occurring within the body rather than examining compounds (such as untranscribed DNA or pre- or post-translationally modified proteins) that may be superfluous to these processes (1Dettmer K. Hammock B.D. Metabolomics—a new exciting field within the “omics” sciences.Environ. Health Perspect. 2004; 112: A396-A397Crossref PubMed Scopus (141) Google Scholar). In addition, there is a relatively small number of metabolites to examine (with the notable exception of plants, which produce a plethora of secondary metabolites) as compared with genes, transcripts, and proteins in their respective omics fields, and therefore the data germane to metabolomics are more easily handled and analyzed. Proponents of metabolomics provide convincing justification that this technique offers more immediate translational benefit than the other omics fields (2Schmidt C. Metabolomics takes its place as latest up-and-coming “omic” science.J. Natl. Cancer Inst. 2004; 96: 732-734Crossref PubMed Scopus (62) Google Scholar, 2Schmidt C. Metabolomics takes its place as latest up-and-coming “omic” science.J. Natl. Cancer Inst. 2004; 96: 732-734Crossref PubMed Scopus (62) Google Scholar). The use of metabolomics through examination of patient urine is in theory an ideal means to study diseases of the urinary tract given that low molecular weight compounds (such as small molecule metabolites) are freely filtered into the urine. In addition, obtaining this biofluid can be done quickly, easily, and in a non-invasive manner in the clinic. Thus, urine metabolomics has potential utility in metabolic profiling as well as for biomarker discovery for cancers of the urinary tract (3Kind T. Tolstikov V. Fiehn O. Weiss R.H. A comprehensive urinary metabolomic approach for identifying kidney cancer.Anal. Biochem. 2007; 363: 185-195Crossref PubMed Scopus (388) Google Scholar). Once urinary biomarkers are discovered and validated, they could conceivably be used for prognosis as well as to predict response to targeted therapies as obtaining urine is always more feasible than gaining access to tumor tissue. There have been several studies looking at single compounds in the urine as markers of non-malignant renal disease. These compounds include N-acetyl-β-d-glucosaminidase, neutrophil gelatinase-associated lipocalin, human kidney injury molecule-1, and interleukin-18 for kidney injury (4Han W.K. Waikar S.S. Johnson A. Betensky R.A. Dent C.L. Devarajan P. Bonventre J.V. Urinary biomarkers in the early diagnosis of acute kidney injury.Kidney Int. 2008; 73: 863-869Abstract Full Text Full Text PDF PubMed Scopus (473) Google Scholar, 5Waikar S.S. Bonventre J.V. Biomarkers for the diagnosis of acute kidney injury.Curr. Opin. Nephrol. Hypertens. 2007; 16: 557-564Crossref PubMed Scopus (75) Google Scholar); one of the same molecules, human kidney injury molecule-1, has also been proposed as a marker for RCC 1The abbreviations used are: RCC, renal cell carcinoma; HILIC, hydrophilic interaction chromatography; PCA, principal component analysis; LDA, linear discriminant analysis; FDR, false discovery rate; PC, principal component; TX, University of Texas Health Science Center at San Antonio; CA, University of California, Davis; AU, approximately unbiased; BMI, body mass index. (6Han W.K. Alinani A. Wu C.L. Michaelson D. Loda M. McGovern F.J. Thadhani R. Bonventre J.V. Human kidney injury molecule-1 is a tissue and urinary tumor marker of renal cell carcinoma.J. Am. Soc. Nephrol. 2005; 16: 1126-1134Crossref PubMed Scopus (147) Google Scholar). The metabolite glucose, when present in has been for as a biomarker of in our have on examining the to a of metabolites in the urine known as well as can in the diagnosis of RCC. study, using relatively that these two can be for the time that this is to a technique (3Kind T. Tolstikov V. Fiehn O. Weiss R.H. A comprehensive urinary metabolomic approach for identifying kidney cancer.Anal. Biochem. 2007; 363: 185-195Crossref PubMed Scopus (388) Google Scholar). on this and our urine metabolomics to be a technique for of RCC, there exist associated with this technique that to be to its In the study, of these and our study on the utility of urine metabolomics in RCC. of urine samples from patients with clear cell RCC and control patients were obtained from two separate A separate of was obtained from a normal group of to the of on changes of metabolomic analysis of mass data that and control patients can be using a larger cohort of the urinary metabolic profile is dependent on site of when sources of variation in the data were was that a small of urinary metabolites to the and control these are the metabolites to for identification and biomarker Surprisingly urine samples from patients several postnephrectomy were from urine samples of patients, suggesting that metabolic of the were in the term by removal of the tumor. In addition, there was little variation in the urinary due to the time since the last meal, suggesting that samples are for this of These data confirm the utility of urine metabolomics analysis for RCC and a of that of the data and that to be the technique is After by the the of or or the of urine samples from clear cell RCC patients of and and at and were obtained from the at the University of California, Center or the University of Texas Health Science Center at San patients from the University of California, data were patients in the have known kidney or renal TX, two samples were obtained from RCC samples to as and after in the same patients as urine samples were obtained to or the of all University of California, and In a separate study, were taken from a cohort of 13 normal at of the and at after In all collection urine was in a urine and at within of were on analysis was performed as (3Kind T. Tolstikov V. Fiehn O. Weiss R.H. A comprehensive urinary metabolomic approach for identifying kidney cancer.Anal. Biochem. 2007; 363: 185-195Crossref PubMed Scopus (388) Google Scholar). urine was with an of at samples were for at to for was performed using T. and 13 for HILIC, by as the at of at and T. were from was used by a A were performed with a or and an After a at a to was at and a to was at The was to were to the of the linear mass was to at a of was to The of the was at mass were from to and mass in and The in individual and all was performed using to data the the data were from the into using the to with were as time two and time were as time mass and number of were in than of the samples of all this was using a of was performed using data. data was The data that and of detected and were from into the detected in a control data and in a control data and were detected using a that and mass within a time and a mass The were and a mass to an or was in a from to in the data and from to in the normal patient data The data were for analysis was with the and for and that the of the metabolomic was in the RCC samples than the control was to the two and analysis that to in this study, as and differential analysis, are on the that of on the as for other the variance in our data to with the to the metabolite to analysis in all to the analysis was performed to or individual samples a of their The with each in a separate and the the number of at each there is a single each were by Scholar). an analysis of variance approach and the of of two that can be at each the in analysis, was performed on and were using the R. an for the in PubMed Scopus Google Scholar). The of for each were using normal and The were for or relating to the two of RCC and control or pre- and and variation to site and urine component analysis was for The of the was to to a small number of principal components that of the variation in the data Scholar). linear discriminant analysis R. The use of in Google was performed on the principal components to a linear using for each of The was to be for all The group was on the of a to a of was by the of that were The was used to evaluate the of the metabolic profile that the group of a on the of the The is on one patient at a and the is used for the The the of principal components on all metabolites to the by for The is by the of samples when for each using the is each patient at a time all patients have been and the all the analysis was performed to identify metabolites the RCC from normal using a test or were by the false discovery the of false the the false discovery a and approach to R. Soc. Scholar). was as on the of in a approach was used to the that a metabolite is a by the of a M. C. R. A approach for the analysis of Google Scholar). a that for each of metabolites a of difference in difference in two for the is for all metabolites with a test Under the that metabolite are for all the are to a from to the metabolites for which the that at one metabolite has is the are to a that to than A of and the of The in the were using with a The was used to the that a metabolite is the two The is the that a metabolite with a given or is the two In to the the were used to metabolites in of A metabolomics has sources of variation that can be to and other from urine and other performed variance component analysis to examine the of in a metabolomics K. T. of variation in 2005; PubMed Scopus Google Scholar). The of sources of variation were using a linear in the of using the The of each were as is the of group variation is the of variation is the of variation is the of tumor variation within and is the residual variation by other than the in the group could be by variation by other than the and tumor each variance components were The variance was to be the of The of each of variation was as a of the variance to the of all variance the of group variation, and the of variation due to variation in as to mass spectrometry the of compounds that in urine (3Kind T. Tolstikov V. Fiehn O. Weiss R.H. A comprehensive urinary metabolomic approach for identifying kidney cancer.Anal. Biochem. 2007; 363: 185-195Crossref PubMed Scopus (388) Google Scholar, Fiehn O. of compounds of of hydrophilic interaction and mass Biochem. PubMed Scopus Google Scholar). were by the technique because of the and T. Tolstikov V. Fiehn O. Weiss R.H. A comprehensive urinary metabolomic approach for identifying kidney cancer.Anal. Biochem. 2007; 363: 185-195Crossref PubMed Scopus (388) Google the in this in the study, to identify all of the detected rather on of the use of the mass and for the of diagnostic for RCC. The was that a group of potential biomarkers was more to for than single discovered as a potential metabolites can be in the sources of in metabolomics analysis in normal to the as to body mass or the urinary 13 with The were was and individual from to the 13 were in the study, were were early of and were were more to in the after the urine was taken in the and taken in the several urine samples from each to after a in the same and a to after a on a separate When by and the were as metabolites were shows the of of of all metabolites in each from each analyze the of all urinary metabolites within and these normal analysis of metabolite of metabolites was It can be from this analysis that all samples a of metabolite metabolite for were more than for the of metabolite for each urine of to be for were more the from the urine with the from the urine approximately as well as with the from the urine as by the the the and urine samples was than that the and urine The urine from to the urine from to suggesting that is than the of and on changes of urinary metabolite compared the for individual metabolites the samples and samples and samples using the metabolites, metabolites the and samples There were metabolites was the samples and samples after for FDR, of metabolites were as at It that more variation of metabolites was in the as compared with the taken of taken on a suggesting that changes in urine metabolite analysis are more in after a as compared with and the of the metabolites, the of on metabolomic variance were these reasons, control for or time after in In the of urine samples from patients with clear cell RCC and and control patients were obtained from the at CA, and clear cell RCC and postnephrectomy and radical) were obtained from at and were from the same patients obtained at from to after and was due to the of on in using normal analysis was performed to the urine samples from RCC and control patients to study The and and samples obtained from the and collection There was of the control urine from the RCC and the RCC urine and of the RCC and RCC urine to be and a although the RCC from the RCC Thus, urine metabolomic in due to of and samples to of the samples from the same collection identify metabolites that the group and the compared the for individual metabolites the control group and RCC group in the data metabolites, metabolites the two after for FDR, metabolites were the two at the data there were metabolites after these metabolites, was as at the metabolites as metabolomic profile biomarkers, a approach to the of a metabolite a by the of a on the of test The that metabolites than that a metabolite is the control and RCC in the data These metabolites be in the of in performed of all obtained from each sample. the analysis that the urinary metabolite profile is dependent on site of samples separate from samples for the of the A of all the urine samples from is in the by the principal components of the variance in the data was by the The of shows RCC and control samples as on the analysis The of and that in samples to It is shown that and of of the the and RCC is with analysis shown and using samples from our study (3Kind T. Tolstikov V. Fiehn O. Weiss R.H. A comprehensive urinary metabolomic approach for identifying kidney cancer.Anal. Biochem. 2007; 363: 185-195Crossref PubMed Scopus (388) Google and the utility of urine metabolomics analysis in RCC from control The of the the and samples with the analysis shown the and samples a single rather than two Thus, urine metabolite can separate RCC from control patients and are in urine samples taken and after removal of the primary tumor. which of the are the two the of and and that the two test and to have suggesting that were the for the group in the samples Thus, the with a in of RCC The of the the the two to components of to yield sufficient for the present the the components of were for the pre- and the and the that can separate urine samples into a number of the discriminant The was to a linear using the principal component as new that the for each of the number of to The by the of The of the discriminant and the of samples to the group are shown to and site of collection of of group using for each of the number of of of of of is the of RCC samples as RCC of of is the of as RCC of of of is the of RCC samples as RCC of is the of as RCC in a new The of the linear discriminant was by the shows the obtained from and linear discriminant for the and The that for the the control and RCC samples can be in of RCC for the data The of was within the In the of a was as due to the of the pre- and samples as shown in as well as shown The the and samples are in of of the the of samples to their group of using discriminant for each of of of of of is the of RCC samples as RCC of of is the of as RCC of is the of RCC samples as RCC of is the of as RCC in a new in metabolomics data can from several sources that are to and from samples due to and other that are and of to changes in data which can the of data and to variance component analysis was performed to the of in our metabolomics the of sources of variation using the samples at the University of California, and the samples at the University of Texas at San because the two data sources of group The group variation in the data from the of the group variation in the data was due to the removal of the primary tumor. each metabolite within the data were as is the of group variation is the of variation is the of variation is the of tumor variation within the and is the residual which is the of variance that by the in the The was on the of each of the were as and variance components were The of of sources of variation that of the metabolites small little of the A small of the metabolites variation the two metabolites at the in of group variation than suggesting that for metabolites the group was the to variation in the these metabolites are to in viable biomarkers for RCC. our on these metabolites, which a in performed and for metabolites and the as The metabolites a of compared with that of using the metabolites, suggesting that the metabolites with variation are in the the of the metabolites, the greatest of variation was residual that these metabolites be as biomarkers. the of each variance of sources of data in a new the data for each metabolite were as is the of variation is the of variation is the of variation is the of tumor variation within the and is the residual The was on the of each of the were as and variance components were The of of sources of variation for the data were to that for the data The that of the metabolites small of is that than of variance was by the of the metabolites, the greatest of variation was residual of each variance component for the data of sources of data in a new kidney is the of deaths and of is the of in patients per year in the United States. is the is that of are at When detected with symptoms, prognosis of RCC is T. renal cell carcinoma in the in PubMed Scopus Google Scholar); RCC has a survival R.H. identification of for Int. Full Text Full Text PDF PubMed Scopus Google Scholar). these reasons, and non-invasive are for identifying RCC at an to The new of in which the of metabolites by an is for the of diagnostic on metabolite profiling as well as for the discovery of individual biomarkers and the of its urine is a biofluid to examine in this in of the that RCC is a this was for an of this In a study using a number of have shown that urine metabolomics analysis has the of RCC from control patients (3Kind T. Tolstikov V. Fiehn O. Weiss R.H. A comprehensive urinary metabolomic approach for identifying kidney cancer.Anal. Biochem. 2007; 363: 185-195Crossref PubMed Scopus (388) Google Scholar). The present study that a means of potential urinary biomarkers, and that to be in data from In to our study and that patients with RCC can be from control patients by urine metabolomics analysis, is shown that site of collection substantial in metabolomics to into the collection and the that all samples were in the same by the same of samples at the University of California, with at the University of Texas at San with in other cancers using D. D. A. M. D. profiling with 2008; PubMed Scopus Google Scholar). Thus, investigations to be at the of by the use of more than one a for of a screening this is urine metabolomics studies be to a single collection have been with other omics of a further analysis of control 2008; Scholar). have also now the urine in normal have shown in normal and control looking at metabolites D. T. B.D. of metabolites in normal human 2007; Scopus Google have at in that is relatively small C. D. of a urine in for 2007; Scopus Google Scholar). A of our study is that there was variation in urine metabolomics analysis when to the time of the last the of the metabolites, the of on their metabolomic was and can be on this study that is to control for normal in urinary metabolomics have the of on the urine an of for have the of on the urine in In in our study of RCC patients and a single urine was obtained from each within the same collection site from to in these metabolomic to be variation, as well as and to be in have that, although there is a degree of group variation in the urine metabolomics RCC and is that there is substantial residual variation as metabolites that to the group variation be in studies as potential biomarkers, the sources of residual variation to be to the potential of urine of residual variation may include of time of urine and In addition, our of of other of and molecular analysis or T. Tolstikov V. Fiehn O. Weiss R.H. A comprehensive urinary metabolomic approach for identifying kidney cancer.Anal. Biochem. 2007; 363: 185-195Crossref PubMed Scopus (388) Google may yield for variance component analysis and analysis due to in was that samples as for samples from the same patients, a and was that urine metabolomic from samples from the same that metabolic changes persist in patients at several after removal of the primary a that may be by the of known that RCC R.H. identification of for Int. Full Text Full Text PDF PubMed Scopus Google Scholar). In addition, our of of pre- and samples from the University of Texas at San as a control for of collection as well as the of using metabolomic to identify metabolomic and biomarkers of the study has sources of and their by variance component The analysis that to the of metabolomics When a metabolomics study, to the of study the of and to the and a with to all that can the of these lead to and can a in A is to a to of with of metabolites to the number of samples at a single given the of study is as in the study, on several may be In this the of metabolomics studies on of and to metabolomics variation be in the of by the of the urine samples and by and data collection have relating to urine collection and have the utility of urine metabolomics analysis for kidney from control have using variance analysis, that potential biomarkers are there potential sources of and that to be when studies and in for biomarkers in the