Neuromedin U (NMU), a neuropeptide isolated from porcine spinal cord and named because of its activity as a rat uterus smooth muscle contraction inducer, is emerging as a new player in the tumorigenesis and/or metastasis of many types of cancers. Expressed in a variety of tissues, NMU has been shown to possess many important activities in the central nervous system as well as on the periphery. Along with the main structural and functional features of NMU and its currently known receptors, we summarized a growing number of recently published data from different tissues and cells that associate NMU activity with cancer development and progression. We ask if, based on current reports, NMU can be included as a marker of these processes and/or considered as a therapeutic target.
Since it was identified in 1985, neuromedin U (NMU) expression has been found in many species in various isoforms. The longer forms NMU-17 (toad), NMU-21 (goldfish), NMU-23 (rat, tree frog), NMU-25 (human, pig, rabbit, dog, chicken, frog, goldfish), and the shorter forms NMU-8 (pig, dog) and NMU-9 (guinea pig, chicken) have been identified (reviewed in Reference). Although there is still discussion about whether the longer forms of NMU intermediate precursors of the truncated forms are, there is agreement that the peptide is widely conserved between organisms, with emphasis on the amidated C-terminal pentapeptide (-Phe-Arg-Pro-Arg-Asn-NH2). The NMU sequence conservation indicates that its biological functions are tightly correlated with the peptide structure (reviewed in Reference).
In addition to different NMU isoforms, Mori K et al. detected NMU-precursor-related peptides, NURP33 and NURP36, produced from the same NMU precursor in rat pituitary, small intestine, and brain tissues. The NURPs are involved in the pituitary release of prolactin, but they are unable to activate known NMU receptors. Thus, the molecular mechanisms that controls various peptides’ synthesis from NMU precursors (alternative splicing, protease availability during peptide processing) seem to be vital for the regulation of their biological activity.
The picture is even more complicated, as a 36-amino-acid-long peptide, named neuromedin S (NMS), was also found in rat brain. However, NMS is not a splicing variant of NMU. Although NMU and NMS have similar structures and receptor affinities and both primarily function as neuropeptides, NMU appears to be the focus of attention in the field of cancer development and progression.
Human neuromedin U is encoded by the NMU gene, located on chromosome 4 (4q12), and is synthesized as a 174-amino-acid precursor. The first 34 aa at the N-terminus are the signal peptide, which is typical for the precursors of many other regulatory peptides. The mature NMU-25 sequence is located within the C-terminus of the pre–pro-peptide and is flanked by pairs of basic residues forming cleavage sites. The bioactivity of NMU from different species depends on two main features: a highly conserved pentapeptide at the C-terminus and post-translational amidation of the C-terminal amino acid, which is typical of many gastrointestinal hormones and determines receptor binding capacity. Human pre–pro-NMU cleavage mainly generates NMU-25, but the presence of other putative proteolytic sites in the precursor suggests the possibility of releasing a series of other peptides, as shown in rats.
The distribution of NMU-25 in humans showed its expression in the central nervous system, gastrointestinal tract, oesophagus to rectum, genitourinary tract, thyroid gland, spleen, lymphocytes, adipose tissue, mast cells, endothelial cells, keratinocytes, and placenta.
Neuromedin U plays its function through interaction with two main receptors: neuromedin U receptor 1 (NMUR1, previously FM-3, GPR-66) and neuromedin U receptor 2 (NMUR2, previously TGR-1), encoded by separate genes located on chromosomes 2 and 5, respectively. Both receptors share relatively high sequence and amino acid homology (~50%) and demonstrate comparable sub-nanomolar affinity to NMU. Both NMUR1 and NMUR2 were discovered as growth hormone secretagogue receptor (GHSR) and neurotensin receptor (NTSR) homologues, but testing of ghrelin, neurotensin, different types of neuromedins, and other similarly structured factors showed NMU and NMS as the only ligands for NMUR1 and NMUR2. Human receptor activation is, to some extent, NMU origin and isoform independent, as rat NMU-23, canine NMU-8, and porcine NMU-25 or NMU-8 induce signalling just as does human NMU-25.
First reports identifying neuromedin U as a cognate ligand of orphan G-protein coupled receptors were published almost simultaneously by American and Japanese groups, and all distribution data were based on NMUR mRNA detection. The initial studies showed diversification in NMURs tissue distribution. NMUR1 was found to be prominently expressed in the periphery (e.g., gastrointestinal tract, male genitourinary system, lungs, kidneys, cardiovascular and immune system) and NMUR2 expression was mainly detected in the central nervous system. Nonetheless, further studies complicated the picture and showed NMUR1’s presence in the cerebellum, hippocampus, and hypothalamus, while NMUR2 mRNA has been identified in peripheral tissues of genitourinary and gastrointestinal tracts and in many other organs. This controversy emerged from the development of advanced detection techniques over recent years, and it can also be the effect of relatively high amino acid homology between NMUR1 and NMUR2, which implicates the shortage of effective experimental tools, such as highly specific antibodies. As antibody staining appeared to be non-specific and mainly inconsistent with RNA expression data (especially in the case of NMUR1, as seen in the Human Protein Atlas, among other sources), mRNA level is still used as the first measure when particular receptor presence is determined.
In addition to classical NMURs, Lin et. al. (2015) revealed a lack of the third exon in an NMUR2 splice variant, NMUR2S, in human ovarian cancer. The expression of NMUR2S was also confirmed in various other human cancer cell lines.
Interestingly, the NMU receptor–ligand binding signal was also detected in cell lines without NMUR1 and NMUR2 expression. In non-small-cell lung cancer (NSCLC), Takahashi et al. (2006) found that the heterodimer formed by growth hormone secretagogue receptor 1b (GHSR1b) and neurotensin receptor 1 (NTSR1) was involved in NMU-related signalling. Thus, it cannot be excluded that other NMU receptors could be found in the future.
All currently known NMU receptors NMUR1, NMUR2, NMUR2S and GHSR1b/NTSR1 belong to the large family of G-protein-coupled receptors (GPCRs). They have membrane localization and, except NMUR2S, classical GPCR structures with an extracellular N-terminus responsible for ligand binding, seven transmembrane domains, and an intracellular C-terminus. Comparing two classical NMURs, NMUR2 has a shorter third intracellular domain as well as an N-terminus and a longer C-terminus than NMUR1. The truncated variant of NMUR2 (NMUR2S), due to the lack of the sixth transmembrane domain and the third extracellular loop, forms only six transmembrane domains with both the N- and C-termini, localized extracellularly, which is thought to be the major reason for NMUR2S’s negative modulation of NMU signalling.
As previously mentioned, the approximately 50% amino acid homology between NMUR1 and NMUR2 implicates the shortage of effective tools to clearly define the functions of individual receptors. Many studies on the signal transduction triggered by NMUR activation have been performed with the use of cell lines with ectopic overexpression of particular NMUR and have been based on extensive knowledge of the GPCR receptor family.
It is known that GPCRs propagate signals in the cell through heterotrimeric G-proteins. Upon ligand binding, GPCRs and G-proteins change their conformation and transduce signals inside the cell (Figure 1).
Both activated NMUR1 and NMUR2 regulate signalling pathways involving inositol phosphates and calcium as secondary messengers. Studies based on the various sensitivities of different G-protein subunits to the pertussis toxin (PTX; Gα q insensitive, Gα i sensitive) have established which subunit of G-proteins are involved in NMURs signal transduction. PTX pre-treatment of HEK-293 or COS-7 or CHO cells overexpressing either NMUR1 or NMUR2 did not interfere with calcium mobilization after NMU application, indicating that Gα q is the main player in the process.
Subsequently, phospholipase C was identified as an enzyme underlying phosphoinositide metabolism and the calcium mobilization response. Finally, the applica
