Are Circulating Phospholipid Fatty Acids a Reasonable Surrogate for Levels in The Prostate?

Abstract

BackgroundEpidemiologic studies of circulating fatty acids and prostate cancer risk have been inconsistent, which has complicated nutritional recommendations for prostate cancer prevention. Inconsistencies may partly be due to use of blood concentrations as surrogate biomarkers for the fatty acid profile of target tissue. Since the target (prostate) tissue is of clinical interest, we evaluated the associations between blood and prostate fatty acid concentrations.MethodsPhospholipid fatty acid (PLFA) profiles were measured in plasma and benign prostate tissue from 20 patients who underwent prostatectomy. For each participant, three prostate tissue specimens varying in size (10 – 50 mg) and location were collected. PLFA profiles were measured in a single 0.2ml aliquot of plasma, and in each tissue sample separately. To evaluate correlations of tissue PLFAs between specimens ≤20 mg and >20 mg, Pearson correlation coefficients were calculated. Intraclass Correlation Coefficients (ICC) and coefficients of variation (CV) were used to evaluate the reliability of prostate tissue PLFA from three distinct specimens of each participant. Pearson correlations were also calculated between plasma and mean tissue PLFA concentrations.ResultsMean PLFA concentrations from smaller (≤20 mg) and larger (>20 mg) tissue were nearly identical. Correlations between tissue specimens of varying size were strongest for omega‐3 (ω‐3) fatty acids, ranging from 0.67 for alpha‐linoleic acid to 0.99 for eicosapentaenoic acid. For selected trans‐fatty acid concentrations, correlations between specimens ≤20 mg and >20 mg varied substantially, ranging from 0.10 for trans‐18:2, to 0.83 for trans‐18:1. Correlations for individual omega‐6 fatty acid concentrations were moderate (0.51 to 0.72), although still statistically significant. Intraclass Correlation Coefficients (ICCs) comparing tissue concentrations from three distinct specimens per participant were moderate to strong for most PLFAs, ranging from 0.66 for linoleic acid to 0.96 for eicosapentaenoic acid, with only one ICC below 0.50 (TFA 18:2: ICC=0.28). Mean coefficients of variation (CV) between the three distinct tissue specimens were all < 0.16. For most fatty acids, mean tissue concentrations were lower than plasma concentrations. Correlations between plasma and tissue concentrations were strongest for EPA (0.93) and total ω‐3 fatty acids (0.90); however, for ALA, the correlation between plasma and tissue was very poor (0.03).ConclusionsPLFA profiles can be reliably measured in small quantities of prostate tissue, and concentrations are largely homogeneous within the prostate. Correlations between plasma and tissue concentrations are moderate to strong for most PLFAs. The overall strong correlations between plasma and tissue concentrations suggest that for most individual PLFAs, plasma measures are adequate surrogate markers of target (prostate) tissue; these findings are important for interpreting nutritional associations of PLFAs and prostate cancer riskSupport or Funding InformationResearch Support: Pacific Northwest Prostate Cancer SPORE (P50CA97186); Fred Hutchinson Cancer Research Center.