Uromodulin in Mineral Metabolism

Abstract

Purpose of review:

Uromodulin (UMOD), also known as Tamm-Horsfall protein, is the most abundant protein in human urine. UMOD has multiple functions such as protection against urinary tract infections (UTIs) and nephrolithiasis. This review outlines recent progress made in UMOD’s role in renal physiology, tubular transport, and mineral metabolism.

Recent findings:

UMOD is mostly secreted in the thick ascending limb (TAL) and to a lesser degree in the distal convoluted tubule (DCT). UMOD secretion is regulated by the calcium sensing receptor. UMOD upregulates ion channels (e.g. ROMK, TRPV5, and TRPM6) and cotransporters (e.g. NKCC2 and NCC) in the TAL and DCT. Higher serum UMOD concentrations have been associated with higher renal function and preserved renal reserve. Higher serum UMOD has also been linked to a lower risk of cardiovascular disease and diabetes mellitus.

Summary:

With better serum UMOD detection assays the extent of different functions for UMOD is still expanding. Urinary UMOD regulates different tubular ion channels and cotransporters. Variations of urinary UMOD secretion can so contribute to common disorders such as hypertension or nephrolithiasis.

Introduction

Uromodulin, also known as Tamm Horsfall protein, is the most abundant protein in human urine. The protein was first described by Tamm and Horsfall as a urinary mucoprotein which prevents viral hemagglutination (1). Over thirty years later a protein was purified from the urine of pregnant women which due to its immunosuppressive characteristics was named Uromodulin (UMOD) (2). Subsequently, UMOD and Tamm Horsfall protein were found to be identical (3). In this review I will refer to this protein as UMOD. Different functions have been described for UMOD: Protection against urinary tract infections (UTIs) (4–6), and calcium (Ca²⁺)-containing kidney stones (7–9), contribution to urinary concentration (10, 11), and modification of tubular ion transport for potassium (K⁺), sodium (Na⁺), Ca²⁺, and magnesium (Mg²⁺) (12–15). Moreover, UMOD modifies innate immunity (16–20). Heterozygous mutations in the UMOD gene cause a tubulointerstital nephropathy called autosomal dominant tubulointerstital kidney disease (ADTKD-UMOD) (21). This review describes the latest findings regarding the physiological significance of UMOD in the regulation of minerals in the human body, also known as mineral metabolism.

UMOD trafficking and secretion

The UMOD gene is localzied on chromosome 16p12.3 and encodes a 640 amino acid protein (22). The UMOD protein consists of an N-terminal signal peptide, followed by three epidermal growth factor-like (EGF-like) domains, which are important for protein-protein interaction, a D8C domain of unknown significance, a fourth EGF-like domain, a long zona pellucida domain, which is crucial for UMOD polymerization (23–26), and a C-terminal GPI-anchor (27) (Fig. 1A). UMOD has a high number of 48 cysteine residues forming 24 intramolecular disulfide bridges. About 60% of all UMOD mutations affect one of these 48 conserved cysteine residues and cause protein misfolding and accumulation of the mutant UMOD protein in the endoplasmic reticulum (ER) (28). Recent publications have linked UMOD mutations to increased inflammation, fibrosis, altered unfolded protein response and mitochondrial function (29–32). Regarding further information on ADTKD-UMOD I refer the interested reader to a comprehensive review given the space limitations (33).

Figure 1:

UMOD synthesis and secretion. A. UMOD contains EGF-like domains (green) which are important for protein-protein interaction, a cysteine-rich domain (D8C) (blue) of unknown significance, and a zona pellucida domain (red) which is crucial for UMOD multimerization. B. During UMOD synthesis the molecule is translocated to the ER where the signal peptide is cleaved, N-glycosylation and glycosylphosphatidylinositol (GPI)-anchoring occur, and disulfide bridges are formed. In the Golgi apparatus N-glycans are further modified. UMOD is then sorted to the apical membrane where it is linked to the cell membrane by it’s GPI- anchor. Hepsin, a type II transmembrane serine protease, cleaves UMOD and releases it into urine.

The UMOD precursor enters the cellular secretory pathway resulting in a 85 kDa glycoprotein (34) (Fig. 1B). Apical UMOD is then cleaved by Hepsin, a type II transmembrane serine protease, and released UMOD into urine (Fig. 1B) (23). UMOD homopolymers contribute to the protective effect of UMOD against UTIs and kidney stones, and modify innate immunity (5, 9, 27, 35). Interaction between the internal hydrophobic patch (IHP) within the UMOD ZP domain and the external hydrophobic patch (EHP), which is between the cleavage site and the transmembrane domain, prevents intracellular UMOD polymerization (25). Crystallography further narrowed down the region within UMOD that mediates self-polymerization to an interdomain linker conneting the ZP-N and ZP-C domains (36). Additionally, UMOD is also sorted to the basolateral membrane to a lesser degree, where it is released in the interstitium and enters the circulation as monomeric UMOD (37–39). It was thought that UMOD is exclusively secreted in the thick ascending limb (TAL) (40). Recent findings have shown that UMOD is also secreted in DCT1 (up to approximately 10% of the TAL UMOD secretion) (41).

It has remained unclear how tubular UMOD secretion is modified. The Ca²⁺ sensing receptor (CaSR) reduces UMOD secretion. In mouse models Casr activating mutations lowered UMOD secretion, whereas inactivating mutations increased UMOD secretion (42). Humans treated with cinacalcet also had reduced urinary UMOD secretion (42).

Physiological variations of UMOD secretion

An association between urinary UMOD secretion and eGFR has been suggested (37, 43–45). Twenty-four hour urinary UMOD secretion has been linked to urine volume and kidney size, while age and diabetes mellitus have been associated inversely with UMOD secretion, suggesting that urinary UMOD reflects tubular activity (46). Common variants near the UMOD gene have also been associated with increased renal UMOD expression, higher urinary UMOD secretion, hypertension, eGFR, and decline of renal function (46–53).

Urinary UMOD may estimate nephron mass because a significant reduction of urinary UMOD has been demonstrated in living kidney donors (54). Urinary UMOD may also serve as a predictor for acute kidney injury (AKI) after cardiac surgery. A preoperative lower spot urine UMOD/creatinine ratio implicated a higher odds of more severe AKI postoperatively (55, 56). Lower urinary and serum UMOD levels have also been described in patients with earlier kidney allograft failure (57–59). How urinary UMOD protects against AKI and progressive kidney disease is still unclear. In addition, there is still conflicting data whether high urinary UMOD is protective or aggravating towards CKD. While one hypothesis suggests that urinary UMOD enhances tubular salt absorption and worsens hypertension and subsequent CKD, another line of thought suggests that UMOD modifies the immune system (39, 53, 60–63). There is evidence that UMOD deficiency results in a state of inflammation (61). UMOD modifies the abundance and phagocytic activity of mononuclear phagocytic cells and granulopoesis which affects the response to AKI (39). A cross-talk between the TAL and the proximal tubule has been suggested with UMOD inhibiting chemokine signaling and thereby reducing the risk of adjacent proximal tubule injury (62, 64). Regarding the impact of UMOD on the immune system I refer the interested reader due to space limitation to comprehensive reviews about this topic (60, 65).

Uromodulin in tubular ion transport

Uromodulin modifies salt absorption in the TAL and contributes to hypertension

Mice lacking UMOD (Umod^−/−) display altered expression of renal transporters in the TAL and in the distal nephron (11). Specifically, UMOD stimulates NKCC2 by phosphorylation, reducing intracellular Cl^- concentration (13). Umod^−/− mice respond to furosemide with a blunted response regarding Na⁺, K⁺, and Cl^- excretion but not regarding urine volume (13). These findings are consistent with an intracellular role of UMOD regarding upregulation of NKCC2 activity (Fig. 2). The significance of NKCC2 stimulation by UMOD in mice and humans can be exemplified by higher urinary UMOD secretion in individuals with the UMOD single nucleotide polymorphism SNP4293393 (53). Mice overexpressing wild-type UMOD (Tg^Umodwt/wt had a 80% higher Umod secretion than WT mice) show higher systolic blood pressures, left ventricular hypertrophy, and had higher phosphorylated NKCC2 in total kidney lysate. In Tg^Umodwt/wt animals a low salt diet improves blood pressures and furosemide treatment leads to a higher natriuretic effect compared to WT animals. NKCC2 activation by UMOD seems to be at least partially mediated through the SPAK/OSR1 pathway. Graham et al confirmed the role of UMOD in modulating blood pressure in Umod^−/− mice by demonstrating lower blood pressures and higher urinary Na⁺ excretion compared to control mice (66).

Figure 2:

UMOD modifies ion channels and cotransporters along the TAL and DCT. In the TAL UMOD stimulates from the intracellular space NKCC2 and ROMK cell surface abundance. In the DCT UMOD stimulates NCC possibly via intracellular SPAK/OSR1 signaling. Luminal UMOD enhances TRPM6 cell surface abundance by forming a urinary multi-protein complex. Urinary galectin-1 stabilizes lattice formation between TRPM6 N-glycan and UMOD, thereby impairing dynamin-2 dependent TRPM6 endocytosis. In low Mg²⁺ conditions more urinary UMOD is secreted, possibly to compensate for the low Mg²⁺ state by increasing tubular Mg²⁺ absorption via TRPM6. UMOD increases apical TRPV5 abundance in a similar fashion as TRPM6 by lattice formation and impairing endocytosis.

ROMK forms a functional unit with NKCC2 in the TAL and is crucial for Na⁺ absorption by recycling K⁺ into the tubular lumen which has been absorbed via NKCC2 (Fig. 2). Physical interaction between ROMK and UMOD increases ROMK current and surface expression (12). These findings are consistent with Umod^−/− mice displaying urinary K⁺ losses (66).

In the TAL of WT and Umod^−/− mice RNA-seq data were analyzed when provided with a high salt diet (67). In WT mice a significant upregulation of heat-shock proteins with Hspa1b and a (aka. Hsp70), and other members of the heat-shock family was found but not in Umod^−/− mice. Hsp70 has been implicated in hypertension (68). Additional evidence for a role of UMOD polymers in hypertension was described in the stroke-prone spontaneously hypertensive rat (SHRSP) which serves as a model for maternal chronic hypertension indicating a possible role of UMOD in pre-eclampsia (69). Elevated urinary UMOD levels have been described in pre-eclampsia patients (70). The signifcance of UMOD for hypertension and its relationship to renal dysfunction has been exemplified in a human cohort by the SNP rs12917707 (G/G allele) which was associated with lower renal function and left atrial remodeling (71).

An additional role for UMOD was recently shown for the sodium-chloride cotransporter (NCC) (Fig. 2). The authors showed in in vivo and in vitro experiments that UMOD upregulates the sodium-chloride cotransporter (NCC) via phosphorylation (41). The authors speculated about possible involvement of the SPAK-OSR1 pathway in this mechanism.

Serum UMOD concentrations are associated with hypertension, cardiovascular events, CKD, and diabetes mellitus

More reliable assays measuring serum UMOD levels in humans have complemented our understanding of UMOD in the development of hypertension beyond genome-wide association studies (GWAS). In a cohort of 529 patients with coronary artery disease a protective effect of higher serum UMOD concentrations on mortality was identified (72). A larger study of 3316 individuals referred for angiography confirmed this protective effect by showing that patients who have a higher serum UMOD have a more a favorable metabolic profile, lower rates of hypertension, diabetes mellitus, heart failure, and a lower risk for 10-year mortality (38).

Regarding CKD development higher serum UMOD levels associated with higher kidney function supporting the notion that plasma UMOD is a biomarker of renal function and represents residual renal mass (73–75). Serum UMOD concentrations have been shown to reliably correlate with renal function and have even allowed to distinguish between CKD0 and CKD1 stages (76). One wonders if serum UMOD concentration may be sensitive enough to replace serum creatinine as a more sensitive marker for early CKD. An additional study found an association between UMOD SNPs and CKD in diabetic patients and showed that higher serum UMOD concentrations are associated with a lower CKD risk (77). For a more detailed review regarding the relationship between UMOD and CKD risk in larger populations, I refer the interested reader to a comprehensive review (78).

Evidence shows an association between UMOD and the risk for diabetes mellitus (79, 80). Lower serum UMOD levels in humans correlate inversely with HgbA1c and serum glucose levels (38, 79, 81, 82). Reduced urinary and serum UMOD levels were described in patients with T1DM, T2DM, and diabetic nephropathy (83–89). How UMOD is linked to diabetes mellitus is not clear. One wonders how these findings can be reconciled with the results outlining higher CKD risk and mortality due to higher urinary UMOD levels. Is the UMOD function in the blood and the urine different? Could the beneficial effect of a higher UMOD plasma level be a systemic immunomodulatory effect improving systemic inflammation which is common with metabolic syndrome?

Uromodulin functions as a macromolecule and enhances Ca²⁺ absorption in the DCT to reduce the risk for hypercalciuria and kidney stones

Umod^−/− mice develop spontaneoulsy intrarenal Ca²⁺-containing crystals resulting in renal calcinosis, ureteral obstruction, and hydronephrosis (7–9). We also found large bladder stones in Umod^−/− mice (Fig. 3, unpublished data). In these animals crystals were detected in deep medulla and renal papillae and compared to the Randall’s plaques in humans (90). Renal papillary calcification was found as early as 2 months (9). The crystals contained mostly Ca²⁺-phosphate in the form of hydroxyapatite. Urinary supersaturation was elevated for Ca²⁺ and Ca²⁺-oxalate. Umod^−/− mice showed a compensatory increase of the urinary protein osteopontin, another inhibitor of Ca²⁺ crystal formation (7). Urinary osteopontin and UMOD work synergistically in preventing renal Ca²⁺ crystals. Previously, UMOD had been shown to inhibit Ca²⁺-oxalate stone aggregation as a polyanionic macromolecule (91). UMOD and osteopontin are thought to coat the crystal surface and thereby prevent crystal nucleation and excert an anti-crystal effect (92–94).

Figure 3:

Umod^−/− mice form bladder stones. We confirmed large bladder stones in Umod^−/− mice.

An additional mechanism to prevent kidney stones is the upregulation of the apical Ca²⁺ channel TRPV5 by UMOD (14). Extracellular UMOD, secreted in the TAL stimulated TRPV5 current density by increasing the number TRPV5 channels in the apical membrane. UMOD required dynamin-2 dependent endocytosis for TRPV5 upregulation. The significance of endocytosis for TRPV5 stimulation by UMOD was confirmed in fibroblasts lacking caveolin-1 (Cav1^−/−), another key protein for endocytosis, whereas co-transfection of WT Cav1 plasmid in Cav1^−/− fibroblasts rescued the stimulation of TRPV5 by UMOD. In immunofluorescent (IF) studies of the kidney, Umod^−/− mice displayed a lower TRPV5 abundance compared to WT mice (14). The literature on UMOD in human stone-formers is conflicting due to heterogenous and not well characterized cohorts. Some of the conflicting data may also stem from different preparations or storage conditions of the urine (95).

Calcium and uric acid stone formers excrete a lower urinary UMOD concentration (96). Whereas in control individuals a linear correlation between urinary Ca²⁺ and UMOD was found, this correlation was not present in Ca²⁺-stone formers suggesting that these individuals may be unable to secret more UMOD with a higher urinary Ca²⁺ concentration (96). In a recent study of different kidney stone-formers, stone-formers had 32% lower urinary UMOD concentration and a lower sialic acid content compared to controls, supporting the significance of correct UMOD glycosylation in the prevention of nephrolithiasis (97). Moreover, a GWAS identified a SNP within UMOD that protected against formation of Ca²⁺-containing kidney stones (98).

Uromodulin increases Mg²⁺ absorption in the DCT via TRPM6.

In order to better understand what genes are upregulated in hypomagnesemia the transcriptome of the DCT of mice fed with a low or high Mg²⁺ diet was studied (99). UMOD was described as of one of the highest upregulated mRNAs in the low Mg²⁺ diet cohort but it remained unclear how UMOD compensates for a low Mg²⁺ state. One study showed that Umod^−/− mice have hypermagnesuria and increased expression of genes involved in Mg²⁺ homeostasis, suggesting that UMOD contributes to maintenance of Mg²⁺ balance in a low Mg²⁺ state (15). The TAL is responsible for the majority of tubular Mg²⁺ absorption. One would hypothesize a blunted response to furosemide in the Umod^−/− mice if the Mg²⁺ wasting would be due to a dysfunctional TAL. Previously, in Umod^−/− mice a blunted response of the TAL after furosemide was published (13, 53, 100). While we found hypercalciuria and hypermagnesuria in Umod^−/− mice at baseline, we found no significant difference between Umod^−/− and WT mice after furosemide regarding urine volume, natriuresis, kaliuresis, Ca²⁺, or Mg²⁺ excretion, thus making the TAL as the source for the hypermagnesuria unlikely (15). The different response to furosemide compared to the other groups may be due to differences in age, furosemide dose, genetic background, mouse models, urine collection method, and corrections for urine volume or body weight. Umod^−/− mice have less apical TRPM6 abundance in IF studies of the DCT compared to WT mice. UMOD enhances the cell surface abundance of the Mg²⁺ channel TRPM6 and thereby increases thereby TRPM6 current density (Fig. 2) (15). UMOD excerts its effect on TRPM6 from the extracellular space but requires the ubiquitously secreted urinary protein galectin-1 (15). TRPM6 and UMOD physically interact, allowing for lattice formation between the apical TRPM6 channels and UMOD polymers thereby impairing TRPM6 endocytosis (15). We identified a novel TRPM6 N-glycan site which is required for the UMOD upregulation of TRPM6 current density. Finally, urinary UMOD secretion was upregulated in animals fed a low Mg²⁺ diet, suggesting a feedback mechanism with higher urinary UMOD secretion as a compensatory mechanism to upregulate TRPM6 in a low Mg²⁺ state and to support tubular Mg²⁺ absorption (Fig. 2).

Conclusion and Future Directions:

UMOD has been known as a urinary protein for over a century. With the help of improved detection tools we still learn more about the multi-faceted roles of this protein. UMOD modifies multiple apical tubular ion channels from the intracellular and extracellular space. New insight has shown that not only pathological mutations but also physiological variations of normal urinary UMOD secretion contribute to common diseases. Measurement of urinary UMOD secretion could serve as an inexpensive and noninvasive test to assess nephron mass, predict AKI risk, and kidney transplant longevity. In addition, measurement of serum UMOD may serve as a screening tool for evaluating CKD risk, cardiovascular complications, diabetes mellitus, diabetic nephropathy, and possibly overall mortality. Further studies will have to confirm UMOD’s role as a serum marker for early CKD. Future human studies investigating the effect of UMOD on nephrolithiasis should focus on well characterized subpopulations of stone formers and incorporate detailed dietary information.

Key points:

• UMOD has multi-faceted roles in hypertension, the formation of Ca²⁺-containing kidney stones, and renal K⁺ and Mg²⁺ homeostasis.

• These effects are due to direct effects of UMOD on ion channels (ROMK, TRPV5, TRPM6) and cotransporters (NKCC2 and NCC) along the TAL and DCT.

• Pathological mutations in UMOD result in autosomal-dominant tubulo-interstitial kidney disease (ADTKD).

• Variable secretion of wild-type (WT) UMOD also contributes to different common diseases by exposing individuals with higher urinary UMOD to a higher risk of hypertension and individuals with a lower urinary UMOD to a higher risk of kidney stones, acute kidney injury, chronic kidney disease, and diabetes mellitus.