Urinary vesicles: in splendid isolation.

Urine provides an attractive, non-invasive alternative to tissue, blood or other biofluid as a potential source of biomarkers of systemic and renal disease. The urinary protein content can be divided in two parts: soluble and solid-phase components; the solid-phase components can be subdivided into sediment precipitated with low-speed centrifugation and low-density nanometer-sized particles (vesicles <100 nm in diameter) precipitated by ultracentrifugation. In normal human adult urine, 48% of the total urinary protein excreted is contained in the sediment, 49% is soluble and the remaining 3% is in urinary vesicles [1]. Pisitkun et al. [2] were the first to describe these vesicles in human urine and called them ‘exosomes’; they went on to demonstrate the potential for exosomes as a starting material for biomarker discovery in urine. They chose the term exosomes because the vesicles originate from the membranes of internal multivesicular bodies that are released into the urine when the outer membrane of these bodies fuses with the apical plasma membrane of polarized epithelial cells [2]. Although the discovery of exosomes was intriguing, the question was why we should pay them any attention? Several recent studies have shown why exosomes are attractive for biomarker discovery. First, proteomics has shown that urinary exosomes are derived from all the epithelial cells lining the urinary tract, including glomerular podocytes and renal tubular cells from proximal and distal nephron segments [2]. The protein family commonly associated with exosomes is the tetraspanins, including CD9, CD63 and CD81. Second, other exosomal proteins directly represent the proteome of the source cells, for example, the presence of the sodium-potassium-chloride cotransporter type 2 (NKCC2), the sodium-chloride cotransporter (NCC) or the water channel aquaporin-2 (AQP2) reflects exosomes from the thick ascending limb of the loop of Henle, the distal convoluted and the collecting duct, respectively [2]. This property of exosomes is clinically useful as a biomarker not only of the nature of a renal disease, but also the site of renal injury [3]. For example, the abundance of the aldosterone-sensitive NCC was found to be higher in urinary exosomes of animal models of hyperaldosteronism and in patients with primary hyperaldosteronism [4]. Third, exosomes also contain transcription factors, messenger RNA (mRNA) and microRNA (miRNA), which have been shown to change in a disease [5–7]. Finally, the distal traffic of protein or RNA occurring after exosomes are released into the glomerular filtrate could potentially affect downstream cellular functions and be a novel mechanism of intra-renal signalling. Indeed, a recent study has demonstrated the ability of exosomes to mediate cell-to-cell communication along the nephron with functional AQP2 being transferred between cells [8]. The studies cited above provide a biological rationale for focussing more attention on urinary exosomes. However, as with many technical discoveries of this nature, there have been unresolved methodological problems. Since their initial discovery, the method of harvesting exosomes from various biological fluids, including urine, has relied on a two-step differential centrifugation process [2, 9, 10]. Although widely used, it is still uncertain whether differential centrifugation can isolate all urinary vesicles. This uncertainty was the basis of a recent and important study by Musante et al. [11] reported in the current issue of the journal. In their study, these investigators tried to optimize the process of vesicle isolation and yield, and, specifically, to define the properties of previously uncharacterized vesicles retained in the ultracentrifugation supernatant [11]. Using pooled second morning urines from six healthy volunteers, Musante et al. [11] set out to analyze whether urinary vesicles are still present in the original ultracentrifugation supernatant (‘the crude fraction’). To do so, they subjected the ultracentrifugation supernatant to overnight ammonium sulfate precipitation. A 10 000 g centrifugation step was then used to generate supernatant and pellet. This second supernatant was subjected to two further centrifugation steps and the resulting pellet was analyzed (‘the aqueous fraction’). The pellet obtained IN F O C U S