Translational medicine in fish-derived peptides: from fish endocrinology to human physiology and diseases

Melanin-concentrating hormone interferes with the hypothalamic-pituitary-gonad axis in the Mozambique tilapia

This study was conducted to elucidate the influence of melanin-concentrating hormone (MCH) along the reproductive-axis in the female tilapia Oreochromis mossambicus. Administration of MCH (4 μg / 0.1 ml saline) for 22 days resulted in significantly lower gonadosomatic index compared to controls. Significant reduction in the mean numbers of follicles at different stages of development such as previtellogenic (stages I-III), vitellogenic (stage IV) and preovulatory (stage V) follicles was observed in MCH-treated fish compared with controls. On the other hand, the rate of atresia was significantly higher in follicles at stages II, III and IV in MCH-treated fish. In addition, in the pituitary gland, sparsely labelled gonadotropin releasing hormone (GnRH)-immunoreactive fibres were observed in MCH-treated fish in contrast to their intense labelling in controls. The serum level of luteinizing hormone (LH) showed significant decrease, but the serum cortisol level rose significantly following MCH treatment compared to those of controls. Collectively, these results indicate for the first time, that MCH treatment blocks follicular development during the ovarian cycle, possibly through the suppression of GnRH-LH axis in fish. The results also indicate that MCH may activate the stress-axis pathway in fish.

Variation in the Evolution and Sequences of Proglucagon and the Receptors for Proglucagon-Derived Peptides in Mammals

The mammalian proglucagon gene (Gcg) encodes three glucagon like sequences, glucagon, glucagon-like peptide-1 (GLP-1), and glucagon-like peptide-2 that are of similar length and share sequence similarity, with these hormones having cell surface receptors, glucagon receptor (Gcgr), GLP-1 receptor (Glp1r), and GLP-2 receptor (Glp2r), respectively. Gcgr, Glp1r, and Glp2r are all class B1 G protein-coupled receptors (GPCRs). Despite their sequence and structural similarity, analyses of sequences from rodents have found differences in patterns of sequence conservation and evolution. To determine whether these were rodent-specific traits or general features of these genes in mammals I analyzed coding and protein sequences for proglucagon and the receptors for proglucagon-derived peptides from the genomes of 168 mammalian species. Single copy genes for each gene were found in almost all genomes. In addition to glucagon sequences within Hystricognath rodents (e.g., guinea pig), glucagon sequences from a few other groups (e.g., pangolins and some bats) as well as changes in the proteolytic processing of GLP-1 in some bats are suggested to have functional effects. GLP-2 sequences display increased variability but accepted few substitutions that are predicted to have functional consequences. In parallel, Glp2r sequences display the most rapid protein sequence evolution, and show greater variability in amino acids at sites involved in ligand interaction, however most were not predicted to have a functional consequence. These observations suggest that a greater diversity in biological functions for proglucagon-derived peptides might exist in mammals.

Low-Molecular-Weight Peptides from Salmon Protein Prevent Obesity-Linked Glucose Intolerance, Inflammation, and Dyslipidemia in LDLR-/-/ApoB100/100 Mice

BACKGROUND

We previously reported that fish proteins can alleviate metabolic syndrome (MetS) in obese animals and human subjects.

OBJECTIVES

We tested whether a salmon peptide fraction (SPF) could improve MetS in mice and explored potential mechanisms of action.

METHODS

ApoB(100) only, LDL receptor knockout male mice (LDLR(-/-)/ApoB(100/100)) were fed a high-fat and -sucrose (HFS) diet (25 g/kg sucrose). Two groups were fed 10 g/kg casein hydrolysate (HFS), and 1 group was additionally fed 4.35 g/kg fish oil (FO; HFS+FO). Two other groups were fed 10 g SPF/kg (HFS+SPF), and 1 group was additionally fed 4.35 g FO/kg (HFS+SPF+FO). A fifth (reference) group was fed a standard feed pellet diet. We assessed the impact of dietary treatments on glucose tolerance, adipose tissue inflammation, lipid homeostasis, and hepatic insulin signaling. The effects of SPF on glucose uptake, hepatic glucose production, and inducible nitric oxide synthase activity were further studied in vitro with the use of L6 myocytes, FAO hepatocytes, and J774 macrophages.

RESULTS

Mice fed HFS+SPF or HFS+SPF+FO diets had lower body weight (protein effect, P = 0.024), feed efficiency (protein effect, P = 0.018), and liver weight (protein effect, P = 0.003) as well as lower concentrations of adipose tissue cytokines and chemokines (protein effect, P ≤ 0.003) compared with HFS and HFS+FO groups. They also had greater glucose tolerance (protein effect, P < 0.001), lower activation of the mammalian target of rapamycin complex 1/S6 kinase 1/insulin receptor substrate 1 (mTORC1/S6K1/IRS1) pathway, and increased insulin signaling in liver compared with the HFS and HFS+FO groups. The HFS+FO, HFS+SPF, and HFS+SPF+FO groups had lower plasma triglycerides (protein effect, P = 0.003; lipid effect, P = 0.002) than did the HFS group. SPF increased glucose uptake and decreased HGP and iNOS activation in vitro.

CONCLUSIONS

SPF reduces obesity-linked MetS features in LDLR(-/-)/ApoB(100/100) mice. The anti-inflammatory and glucoregulatory properties of SPF were confirmed in L6 myocytes, FAO hepatocytes, and J774 macrophages.

The brain–adipose axis: A review of involvement of molecules

Many molecules are involved in the regulation of feeding behavior, and they and their receptors are located in the brain hypothalamus and adipocytes. On the basis of evidence suggesting an association between the brain and adipose tissue, we propose the concept of the brain-adipose axis. This model consists of (l) the expression of endogenous molecules and/or their receptors in the hypothalamus and peripheral adipose tissue, (2) the function of these molecules as appetite regulators in the brain, (3) their existence in the general circulation as secreted proteins and (4) the physiological affects of these molecules on fat cell size and number. These molecules can be divided into two anorexigenic and orexigenic classes. In adipose tissue, all orexigenic molecules possess adipogenic activity, and almost all anorexigenic molecules suppress fat cell proliferation. Although the manner, in which they present in the circulating blood connect the brain and peripheral adipocytes, remains to be well-organized, these observations suggest the positive feedback axis affecting molecules in the hypothalamus and adipose tissue. Analysis of the disturbance and dysregulation of this axis might promote the development of new anti-obesity drugs useful in treating the metabolic syndrome.

Comparative pharmacokinetics and pharmacodynamics of urocortins 1, 2 and 3 in healthy sheep

BACKGROUND AND PURPOSE The urocortin (Ucn) peptides are emerging as potential therapeutic targets for heart disease. However, pharmacokinetic (PK) and pharmacodynamic (PD) data are lacking. Therefore, we investigated the PK/PD for all three Ucns. EXPERIMENTAL APPROACH Seven sheep received 1 µg·kg(-1) boluses of Ucn1, Ucn2 and Ucn3. Population PK/PD models were developed to describe the time course of the haemodynamic effects. RESULTS The population estimate for Ucn1 clearance (0.486 L·h(-1)) was lower than that for Ucn2 (21.7 L·h(-1)) and Ucn3 (220 L·h(-1)), while steady-state volumes of distribution were similar for Ucn1 and Ucn2 (∼8 L) but substantially larger for Ucn3 (23.5 L). Ucn1 disposition was adequately described by a two-compartment model, with a one-compartment model required for Ucn2 and Ucn3. The half-life for Ucn1 was 2.9 h (α phase) and 8.3 h (β phase), and 15.7 and 4.4 min for Ucn2 and Ucn3 respectively. All Ucns produced significant increases in heart rate, cardiac output and left ventricular systolic and mean arterial pressures, and decreases in left atrial pressure and peripheral resistance. Delayed-effect pharmacodynamic models best described the time course of haemodynamic responses, with effects more rapid and less prolonged for Ucn2 and Ucn3 than Ucn1. Similar and physiologically plausible estimated baseline (E(0)) effects were exhibited by all Ucns, whereas EC(50) values were generally greater for Ucn1. CONCLUSIONS AND IMPLICATIONS Relative to Ucn1, both the PK and haemodynamic responses to Ucn2 and Ucn3 occurred more rapidly. Our data provide important comparative information, useful to the rational design of future clinical studies.

Distribution of urocortins and corticotropin-releasing factor receptors in the cardiovascular system

Urocortins are human homologues of urotensin I, a fish corticotropin-releasing-factor- (CRF-) like peptide secreted from the urophysis. There are three urocortins: urocortin 1, urocortin 2, and urocortin 3 in mammals. We have shown that urocortin 1 and urocortin 3 are endogenously synthesized in the myocardial cells of human heart and may act on CRF type 2 receptor (CRFR2) expressed in the heart. Expression levels of urocortin 1 in the heart and plasma urocortin 1 levels are elevated in patients with heart failure. Recent studies have shown that urocortins have various biological actions in the cardiovascular system, such as a vasodilator action, a positive inotropic action, a cardioprotective action against ischemia/reperfusion injury, and suppressive actions against the renin angiotensin system and the sympathetic nervous system. Urocortins and CRFR2 may therefore be a potential therapeutic target for cardiovascular diseases, such as congestive heart failure, hypertension, and myocardial infarction.

Cellular models for the study of the pharmacology and signaling of melanin-concentrating hormone receptors

Cellular models for the study of the neuropeptide melanin-concentrating hormone (MCH) have become indispensable tools for pharmacological profiling and signaling analysis of MCH and its synthetic analogues. Although expression of MCH receptors is most abundant in the brain, MCH-R(1) is also found in different peripheral tissues. Therefore, not only cell lines derived from nervous tissue but also from peripheral tissues that naturally express MCH receptors have been used to study receptor signaling and regulation. For screening of novel compounds, however, heterologous expression of MCH-R(1) or MCH-R(2) genes in HEK293, Chinese hamster ovary, COS-7, or 3T3-L1 cells, or amplified MCH-R(1) expression/signaling in IRM23 cells transfected with the G(q) protein gene are the preferred tools because of more distinct pharmacological effects induced by MCH, which include inhibition of cAMP formation, stimulation of inositol triphosphate production, increase in intracellular free Ca(2+) and/or activation of mitogen-activated protein kinases. Most of the published data originate from this type of model system, whereas data based on studies with cell lines endogenously expressing MCH receptors are more limited. This review presents an update on the different cellular models currently used for the analysis of MCH receptor interaction and signaling.

Increased expression of urotensin II-related peptide and its receptor in kidney with hypertension or renal failure

Urotensin II-related peptide (URP) is a novel vasoactive peptide that shares urotensin II receptor (UT) with urotensin II. In order to clarify possible changes of URP expression in hypertension and chronic renal failure (CRF), the expressions of URP and UT were studied by quantitative RT-PCR and immunohistochemistry in kidneys obtained from spontaneous hypertensive rats (SHR), Wistar-Kyoto rats (WKY), and WKY with CRF due to 5/6 nephrectomy. Expression levels of URP mRNA and UT mRNA were significantly higher in the kidneys obtained from SHR compared with age-matched WKY (at 5-16 and 16 weeks old, respectively). A dissection study of the kidney into three portions (inner medulla, outer medulla and cortex) showed that the expression levels of URP mRNA and UT mRNA were highest in the inner medulla and the outer medulla, respectively, in both SHR and WKY. The expression levels of URP and UT mRNAs were greatly elevated in the remnant kidneys of CRF rats at day 56 after nephrectomy, compared with sham-operated rats (about 6.5- and 11.9-fold, respectively). Immunohistochemistry showed that URP immunostaining was found mainly in the renal tubules, vascular smooth muscle cells and vascular endothelial cells. UT immunoreactivity was localized in the renal tubules and vascular endothelial cells. These findings suggest that the expressions of URP and UT mRNAs in the kidney are enhanced in hypertension and CRF, and that URP and its receptor have important pathophysiological roles in these diseases.

Characterization of the insulinostatic effect of urotensin II: a study in the perfused rat pancreas

We have investigated the effect of urotensin II (UII) on insulin secretion at normal and high glucose concentrations as well as induced by secretagogues acting on the B cell via different mechanisms. The study was performed in the perfused rat pancreas. UII, at 1 nM, blocked the insulin response to an increase in perfusate glucose concentration from 5.5 to 9 mM while failed to significantly modify insulin secretion at higher glucose levels (from 9 to 13 mM). The insulinotropic effect of this glucose challenge was reduced by 10 nM UII. UII, at 1 nM, inhibited tolbutamide-induced insulin secretion, whereas, it did not affect KCl-induced insulin release. UII inhibited exendin-4-induced insulin secretion, an effect not observed in pertussis toxin-treated rats.

CONCLUSION

1) B cells are less sensitive to UII at a high glucose level than at a low glucose. 2) The inhibitory effect of UII on both glucose and tolbutamide-induced insulin release, suggests the implication of ATP-dependent K(+) channels. The insulinostatic effect of UII was not observed during KCl stimulation, a condition in which these channels are overridden. 3) The insulinostatic effect of UII can also be mediated by its inhibitory action on the adenylate cyclase/cAMP system via a pertussis toxin-sensitive G(i) protein.

Evidence for endogenous urotensin-II as an inhibitor of insulin secretion. Study in the perfused rat pancreas

In the perfused rat pancreas, infusion of urotensin-II (UII), a somatostatin-like peptide, inhibits glucose-induced insulin secretion. We have resorted to specific antagonists of the UII receptor (UT), palosuran and urantide, to investigate whether endogenous UII also behaves as an inhibitor of beta-cell secretion. The insulinostatic effect of UII was counteracted by palosuran and by urantide but not by a somatostatin-receptor antagonist (cyclo-somatostatin). Furthermore, the insulinostatic effect of somatostatin was not reversed by palosuran. These results suggest that UII and somatostatin blocked beta-cell secretion via distinct receptors. Finally, in the absence of exogenous UII, both palosuran and urantide potentiated glucose-induced insulin release, thus supporting the concept that endogenous UII is an insulinostatic peptide. By virtue of their insulinotropic effect, UT antagonists may be considered potential drugs for treating the impaired insulin secretion characteristic of type 2 diabetic patients.

Seasonal changes in peptide, receptor and ion channel mRNA expression in the caudal neurosecretory system of the European flounder (Platichthys flesus)

The caudal neurosecretory system (CNSS) of the euryhaline flounder Platichthys flesus has suggested roles in osmoregulatory, reproductive and nutritional adaptation, as fish migrate between seawater (winter) and brackish/freshwater (summer) environments. This study examined seasonal changes in mRNA expression profile of functionally important genes in the CNSS. cDNAs encoding neuropeptides, receptors and ion channels were cloned by reverse transcriptase polymerase chain reaction (RT-PCR) and screening of a flounder CNSS cDNA library. The expression profile of cloned genes was determined by real-time RT-PCR at 2-month intervals throughout the year in CNSS from seawater-adapted fish. Plasma cortisol (measured by radioimmunoassay) showed a peak in April, the time of spawning. Expression levels of mRNA for peptides urotensins I and II (UI, UII) and corticotropin releasing factor (CRF) all showed a seasonal cycle, with lowest expression in April and highest in August-October. The expression of CRF2(UI), UT(UII) and CRF1 receptors was not correlated with corresponding peptide expression. Receptors for potential neuromodulators of CNSS activity also displayed a seasonal mRNA expression profile. Glucocorticoid, 5-hydroxytryptamine, kappa-opioid and glutamate receptor expression peaked around April, suggesting that modulation of electrical activity of the neurosecretory Dahlgren cells is of particular importance at this time. Expression of mRNA for L-type Ca(2+) and Ca-activated K(+) channels was lower during the summer months. These channels underlie electrical bursting activity in Dahlgren cells. Ion channel mRNA expression was also lower in CNSS from flounder fully adapted to freshwater as opposed to seawater, consistent with previously reported observations of reduced bursting activity in Dahlgren cells from freshwater-adapted CNSS. These findings support the hypothesis that the CNSS is functionally reprogrammed to cope with changes in physiological challenge as fish migrate between sea and estuaries in winter and spring.

Localization of corticotropin-releasing factor, urotensin I, and CRF-binding protein gene expression in the brain of the zebrafish,Danio rerio

Our current understanding of the corticotropin-releasing factor (CRF) system distribution in the teleost brain is restricted by limited immunohistochemical studies and a lack of complete transcriptional distribution maps. The present study used in situ hybridization to localize and compare CRF, urotensin I (UI), and CRF-binding protein (CRF-BP) expression in the brain of adult zebrafish (Danio rerio). All three peptides were localized in the preoptic area, periventricular hypothalamic and tectal regions, and dorsal part of the trigeminal motor nucleus. CRF and UI were both expressed in several nuclei of the dorsal telencephalon, whereas CRF and CRF-BP were both expressed in the ventral nucleus of the ventral telencephalon. Sole expression of CRF and CRF-BP was apparent in the olfactory bulbs and superior raphe nucleus, respectively, whereas only UI was observed in the corpus mamillare, nucleus of the medial longitudinal fascicle, dorsal tegmental nucleus, nucleus lateralis valvulae, and nucleus interpeduncularis. A major finding of this study was the general regional overlapping of CRF-BP with its ligands and a tendency to be expressed in tandem with CRF rather than UI. Overall, the mRNA expression patterns outlined in this study support the stress-related neuroendocrine, autonomic, and behavioral functions generally ascribed to the vertebrate CRF system and suggest some unique functional roles for CRF and UI in the teleost brain.

Human urotensin II as a link between hypertension and coronary artery disease

Hypertension is a well-known risk factor for atherosclerosis, but the molecular mechanisms that link elevated blood pressure to the progression of atherosclerosis remain unclear. Human urotensin II (U-II), the most potent endogenous vasoconstrictor peptide identified to date, and its receptor (UT receptor) are involved in the etiology of essential hypertension. In patients with essential hypertension, U-II infused into the forearm brachial artery has been shown to induce vasoconstriction. Recent studies have demonstrated elevated plasma U-II concentrations in patients with essential hypertension, diabetes mellitus, atherosclerosis, and coronary artery disease. U-II is expressed in endothelial cells, macrophages, macrophage-derived foam cells, and myointimal and medial vascular smooth muscle cells (VSMCs) of atherosclerotic human coronary arteries. UT receptors are present in VSMCs of human coronary arteries, the thoracic aorta and cardiac myocytes. Lymphocytes are the most active producers of U-II, whereas monocytes and macrophages are the major cell types expressing UT receptors, with relatively little receptor expression in foam cells, lymphocytes, and platelets. U-II accelerates foam cell formation by up-regulation of acyl-coenzyme A:cholesterol acyltransferase-1 in human monocyte-derived macrophages. In human endothelial cells, U-II promotes cell proliferation and up-regulates type 1 collagen expression. U-II also activates nicotinamide adenine dinucleotide phosphate (NADPH) oxidase and plasminogen activator inhibitor-1 in human VSMCs, and stimulates VSMC proliferation with synergistic effects observed when combined with oxidized low-density lipoprotein, lysophosphatidylcholine, reactive oxygen species or serotonin. These findings suggest that U-II plays key roles in accelerating the development of atherosclerosis, thereby leading to coronary artery disease.

Further insights into the neurobiology of melanin-concentrating hormone in energy and mood balances

Melanin-concentrating hormone (MCH) is a critical hypothalamic anabolic neuropeptide, with key central and peripheral actions on energy balance regulation. The actions of MCH are, so far, known to be transduced through two seven-transmembrane-like receptor paralogues, named MCH1R and MCH2R. MCH2R is not functional in rodents. MCH1R is an important receptor involved in mediating feeding behaviour modulation by MCH in rodents. Pharmacological antagonism at MCH1R in rodents diminishes food intake and results in significant and sustained weight loss in fat tissues, particularly in obese animals. Additionally, MCH1R antagonists have been shown to have anxiolytic and antidepressant properties. The purpose of this review is to highlight the recent numerous pieces of evidence showing that pharmacological blockade at MCH1R could be a potential treatment for obesity and its related metabolic syndrome, as well as for various psychiatric disorders.

Expression and characterization of melanin-concentrating hormone receptors on mammalian cell lines

The neuropeptide melanin-concentrating hormone (MCH) is expressed in central and peripheral tissues where it participates in the complex network regulating energy homeostasis as well as in other physiologically important functions. Two MCH receptor subtypes, MCH-R1 and MCH-R2, have been cloned which signal through activation of Gi/o/q proteins and hence regulate different intracellular signals, such as inhibition of cAMP formation, stimulation of IP3 production, increase in intracellular free Ca2+ and/or activation of MAP kinases. Most of the data were obtained with cell systems heterologously expressing either of the MCH receptors. Fewer reports exist on studies with cell lines which endogenously express MCH receptors. Here, we describe human and other mammalian cell lines with which MCH receptor activation can be studied under "natural" conditions and we summarize the characteristics and signaling pathways of the MCH receptors in the different cell systems.

Physiological roles of urocortins, human homologues of fish urotensin I, and their receptors

Urocortin 1, a human homologue of fish urotensin I, together with its related-compounds (urocortins 2 and 3), comprises a distinct family of stress peptides. Urocortin 1 has a high affinity for both corticotropin-releasing factor (CRF) type 1 receptor (CRF1) and CRF type 2 receptor (CRF2), and urocortins 2 and 3 have a high affinity for CRF2, while CRF has a low affinity for CRF2 and a high affinity for CRF1. These differences of the binding affinity with receptors make the biological actions of these peptides. Besides the binding affinity with receptors, the limited overlap of the distribution of CRF and urocortins may also contribute to the differences of physiological roles of each peptide. Urocortins show 'stress-coping' responses such as anxiolysis and dearousal in the brain. In the periphery, recent studies show the potent effects of urocortins on the cardiovascular and immune systems. In this review article, we take a look over the series of peptides included in this family, especially in terms of the versatility of biological actions, along with the various characters of the receptors.

Urocortins as cardiovascular peptides

Urocortins (Ucn) 1, 2 and 3, human homologues of fish urotensin I, form the corticotropin-releasing factor (CRF) family, together with CRF, urotensin I and sauvagine. Ucn 3 is a novel member of this family and is a specific ligand for CRF type 2 receptor. CRF type 2 receptor is thought to mediate the stress-coping responses, such as anxiolysis, anorexia, vasodilatation, a positive inotropic action on myocardium and dearousal. Endogenous ligands for the CRF type 2 receptor expressed in the cardiovascular tissues, such as the myocardium, have long been unknown. We have shown expression of Ucn 3 as well as Ucn 1 in the human heart. Ucn 3 is also expressed in the kidney, particularly distal tubules. Studies in various rat tissues showed that high concentrations of immunoreactive Ucn 3 were found in the pituitary gland, adrenal gland, gastrointestinal tract, ovary and spleen in addition to the brain, heart and kidney. These observations suggest that Ucn 3 is expressed in various tissues including heart and kidney, and may regulate the circulation in certain aspects of stress and diseases, such as inflammation. Ucn 1 and 3 appear to have important pathophysiological roles in some cardiovascular diseases.

Elevated plasma levels of immunoreactive urotensin II and its increased urinary excretion in patients with Type 2 diabetes mellitus: association with progress of diabetic nephropathy

Urotensin II (UII) is the most potent vasoconstrictor peptide ever identified. In order to clarify the pathophysiological role of UII in diabetes mellitus, we examined plasma immunoreactive UII levels and urinary excretion of immunoreactive UII in 10 control subjects and 48 patients with Type 2 diabetes mellitus. The patients were divided into three groups according to the renal function: Group I with Ccr > or = 70 ml/min, group II with 30 < or = Ccr <70 ml/min and group III with Ccr <30 ml/min. Plasma immunoreactive UII levels were elevated in the three diabetic groups compared with normal controls (P <0.05). Group III patients had significantly higher plasma immunoreactive UII levels (15.9 +/- 2.2 fmol/ml, mean +/- S.E.M., n=6) by approximately 1.6-fold than did group I (10.9 +/- 0.9 fmol/ml, n=17) and group II (10.8 +/- 0.8 fmol/ml, n=25) (P <0.05). Urinary excretion of immunoreactive UII was significantly increased in group III patients (52.4 +/- 14.8 pmol/day) by more than 1.8-fold compared with control subjects, groups I and II (P <0.005). Fractional excretion of immunoreactive UII significantly increased as renal function decreased. Presence of diabetic retinopathy or neuropathy had negligible effects on plasma immunoreactive UII levels and urinary immunoreactive UII excretion. Reverse phase HPLC analyses showed three immunoreactive peaks in normal plasma extracts and multiple immunoreactive peaks in normal urine extracts. Thus, Type 2 diabetes mellitus itself is a factor to elevate plasma immunoreactive UII levels, and accompanying renal failure is another independent factor for the increased plasma immunoreactive UII levels in Type 2 diabetic patients. Increased urinary immunoreactive UII excretion in Type 2 diabetic patients with advanced diabetic nephropathy may be due not only to the elevated plasma immunoreactive UII levels but also to increased UII production and/or decreased UII degradation in the diseased kidney.