Exploring the value of using metabolomic analysis of urine to diagnose urological pathologies

Abstract

Urological cancers are in the top ten most common causes of cancer across the globe. Although some cancers are seen in both sexes, the prevalence of urological cancers increases by up to three-quarters in men. Currently, the only route to reliable diagnosis is through an invasive procedure, such as a needle biopsy or cystoscopy. These methods can cause the patient concern and discomfort and carry the risk of complications such as infection and bleeding. The need for reliable non-invasive testing methods is clear and pressing. This project sought to aid in the development of new diagnostic methods in easily obtained, minimally invasive samples such as urine, using metabolomics approaches. Advances in mass spectrometry are allowing the detection of large numbers of metabolites, even at low levels. The derived mass spectrometry spectra can be assessed with a multivariate statistical approach to identify metabolites which could be of diagnostic importance.

The potential of the urine metabolome to define clinically relevant changes linked to disease and health is being increasingly appreciated. Urine can be easily obtained with minimal patient discomfort. However, the stability of the urine metabolomes after being obtained from the patient over time and under different temperatures needs further characterization. To assess the feasibility of using urine for metabolomic assessment from a variety of sources, the study examines the impact of storage, temperature, time, and centrifugation speeds in comparison to raw urine, using samples from healthy volunteers, lung cancer and lung disease symptom control patients (which were the only clinical urine samples available at the time). Initial findings observed that although many metabolites remained stable some underwent significant changes even at -20 oC. Following pre-treatment at different temperature and centrifugation speeds however, the observed biochemical pathway analysis suggested that these differences could potentially be overcome when used for biomarker discovery. The only caveat being comparing raw to previously frozen urine should be avoided. The robustness of urine over time and temperature was again tested in the comparison of unprocessed lung cancer and symptom control patients. In this case, the groups were consistently metabolomically distinct. These findings together indicate collection with no preprocessing in a primary care or domestic environment does appear to be a viable option.

In the analysis of prostate cancer (PCa) in comparison to benign prostatic hyperplasia (BPH) and symptom control patients the study investigated a total of 104 urine samples (46 PCa, 29 BPH, 29 controls) and 156 plasma/serum samples (79 PCa, 11 BPH, 66 agematched symptom controls) obtained following informed consent, using flow infusion electrospray mass spectrometry (FIE-MS). Multivariate assessments of the derived data showed that the experimental classes were distinctive. The major sources of variation were identified and receiver operating characteristic (ROC) area under the curve (AUC) assessments showed that a combination of five metabolites had AUC values (accuracies) of between 89% and 93%. In comparison, plasma and serum indicated poor separation between the experimental classes with lower overall AUC values of between 55% and 89%. Assessment of the biological function of the key metabolites in urine suggested bioenergetic shifts in metabolism of PCa cells, in particular highlighting increased mitochondrial metabolism of fatty acids. There was evidence of inflammatory mediators derived through arachidonic acid (AA) metabolism across all experimental classes. In the BPH cohort there appeared to be changes in primary bile acid synthesis, alongside higher levels in several metabolites indicating increased mitogenesis. This suggests that there is the potential to determine differential metabolic markers between the experimental groups.

A pilot study comparing other urological diseases was also conducted comparing PCa to symptom controls, bladder cancer (BC) and kidney cancer (KC). This consisted of 15 urine samples (7 PCa, 3 BC, 5 Male controls) and 30 serum and plasma samples (12 PCa, 5 BC, 8 KC, 5 male controls). Few substantial differences were observed when comparing all the significant groups. However, some differences were seen when comparing between malignant diseases. In urine, lipid and bile acid derivatives were observed when comparing metabolites found in the PCa and BC cohorts, ANOVA targeted significant (P < 0.001) changes. However, ANOVA suggested no statistically significant differences between the BC samples and the symptom control group. The plasma samples suggested the same significant differences between the BC and PCa group, where reciprocally, lipid metabolism appeared to be distinct. The KC cohort showed unique
increases in lipid metabolites, but these also showed considerable variability. Upon analysis of the serum samples there were no statistically significant differences. Taken together, the findings of this study suggest the urine metabolome can be used to further our understanding of the pathogenesis of urological cancers, although the identification of accurate diagnostic markers could be more challenging. However due to the relatively small sample size further experimentation is needed. This study provided the opportunity to demonstrate the robustness of urine as a diagnostic tool. There appears to be great potential for more widespread use of this biofluid, which is both easy to obtain and non-invasive. Urine analysis does not come without caveats, however, and attention needs to be paid to collection and storage protocols. The diagnostic capability of urine in respect of urological cancers shows great promise, particularly in PCa. Early findings have identified metabolic changes in bioenergetics and inflammatory pathways. This has resulted in the identification of key metabolic fingerprints which have the potential to be developed into diagnostic tools for both malignant and benign prostate disease. This has both enhanced and furthered the knowledge in the use of urine as a diagnostic tool in urological disease.

Date of Award	2023
Original language	English
Awarding Institution	Aberystwyth University
Supervisor	Luis Mur (Supervisor) & Manfred Beckmann (Supervisor)