Abstract
Congenital anomalies of the kidney and urinary tract (CAKUT) are the most common cause of end-stage kidney disease in children. While the genetic aberrations underlying CAKUT pathogenesis are increasingly being elucidated, their consequences on a cellular and molecular level commonly remain unclear. Recently, we reported rare heterozygous deleterious LIFR variants in 3.3% of CAKUT patients, including a novel de novo frameshift variant, identified by whole-exome sequencing, in a patient with severe bilateral CAKUT. We also demonstrated CAKUT phenotypes in Lifr−/− and Lifr+/− mice, including a narrowed ureteric lumen due to muscular hypertrophy and a thickened urothelium. Here, we show that both in the ureter and bladder of Lifr−/− and Lifr+/− embryos, differentiation of the three urothelial cell types (basal, intermediate and superficial cells) occurs normally but that the turnover of superficial cells is elevated due to increased proliferation, enhanced differentiation from their progenitor cells (intermediate cells) and, importantly, shedding into the ureteric lumen. Microarray-based analysis of genome-wide transcriptional changes in Lifr−/− versus Lifr+/+ ureters identified gene networks associated with an antimicrobial inflammatory response. Finally, in a reverse phenotyping effort, significantly more superficial cells were detected in the urine of CAKUT patients with versus without LIFR variants indicating conserved LIFR-dependent urinary tract changes in the murine and human context. Our data suggest that LIFR signaling is required in the epithelium of the urinary tract to suppress an antimicrobial response under homeostatic conditions and that genetically induced inflammation-like changes underlie CAKUT pathogenesis in Lifr deficiency and LIFR haploinsufficiency.
Introduction
Congenital anomalies of the kidney and urinary tract (CAKUT) comprise various structural malformations that result from defects in the morphogenesis of the kidney and urinary tract, including renal agenesis (absence of kidney) or duplex (double) kidneys, renal dysplasia (kidney with abnormal structures) or hypoplasia (small kidneys with reduced nephron number), as well as fused (grown together) and ectopic (aberrantly located) kidneys, hydronephrosis (distended renal pelvis and calices), megaureter (dilated ureter) and vesicoureteral reflux (VUR) (1). All CAKUT phenotypes sum up to a prevalence of 3–9/1000 live births (2–4) and represent a significant health burden because they cause around 40% of cases of end-stage kidney disease in children (5). Although over 50 genes are known to cause CAKUT when mutated (6–8), and chromosomal abnormalities (4), microdeletions or microduplications can also be causative (9,10), the majority of CAKUT patients (around 60%) remain genetically unsolved (8).
In the last few years, novel CAKUT-causing genes have been identified using next-generation sequencing (NGS) approaches and different strategies of data analysis (6,11–15). Using whole-exome sequencing (WES) and a trio-based de novo strategy, we recently identified a novel heterozygous de novo frameshift variant in the leukemia inhibitory factor receptor (LIFR) gene in a patient with severe bilateral CAKUT (16). LIFR encodes a transmembrane receptor that can form multimeric complexes with certain other receptors, such as interleukin 6 signal transducer also called glycoprotein 130 (GP130). Receptor complexes containing LIFR are utilized by several interleukin 6 (IL6) cytokine family members including the LIF interleukin 6 family cytokine (LIF) to stimulate JAK/STAT, MAPK and PI3K signaling (17). Subsequent LIFR sequence analysis demonstrated rare heterozygous deleterious LIFR variants in 3.3% of CAKUT patients who presented with hydronephrosis, obstructive megaureter and renal hypodysplasia (16). Moreover, Lifr−/− mice known to die perinatally due to neurologic and skeletal anomalies (18,19) were found to display similar phenotypes including hydronephrosis, megaureter, blind-ending ureter, renal hypoplasia and microcystically dilated renal tubules (16). Together with a previous report on a patient with combined CAKUT and Stüve–Wiedemann syndrome (20), which is caused by biallelic LIFR mutations (21), and a recent study demonstrating that LIF signaling via LIFR is required for renal development in Xenopus laevis (22), these data strongly support the notion that LIFR is critical for the development of the renal system across various species and causes CAKUT when mutated.
In our recent study, we also observed a thickening of the epithelial compartment in the ureter of Lifr−/− and Lifr+/− mouse embryos (16). Together with high expression of Lifr in the epithelium of Lifr+/+ mouse ureters at embryonic day (E)18.5 and in normal human renal pelvis (16), this suggested a specific function of LIFR in the development and/or homeostasis of the ureter epithelium. The epithelial lining of the ureter and bladder, the urothelium, is organized in three cell layers of variable thickness and function: a single layer of abundant cuboidal basal cells (B-cells) that anchor the tissue to the lamina propria via the basement membrane, one or more layers of intermediate cells (I-cells) presumed to serve as precursors and an apical cell layer consisting of large hexagonal cells known as superficial cells (S-cells) or umbrella cells (23,24). S-cells feature highly impermeable tight junctions and a specialized unique asymmetric unit membrane composed of uroplakin (UPK) proteins to mount a barrier against the luminal content (24–26). During embryogenesis, S- and B-cells differentiate gradually from the underlying I-cells (27,28). Under conditions of homeostasis, urothelial cell turnover is very slow but rapidly increases after epithelial injury or bacterial infection (26,28).
Here, we set out to characterize the urothelial changes associated with reduced or lost Lifr function both in Lifr+/− and Lifr−/− mice and in a CAKUT patient with LIFR haploinsufficiency (16) on a cellular and a molecular level. We provide evidence that LIFR is essential for urothelial integrity by preventing an inflammatory response under homeostatic conditions. Our data suggest a link between inflammatory processes genetically induced by Lifr deficiency or LIFR haploinsufficiency and CAKUT pathogenesis.
Results
Desquamation of S-cells in the ureters and bladder of Lifr−/− and Lifr+/− embryos
To further characterize the urothelial changes previously described in Lifr−/− embryos and to elucidate their role in the CAKUT phenotypes observed (16), we performed histological and molecular stainings on proximal ureter sections at different developmental stages. Urothelial differentiation was determined by analyzing the expression of KRT5, ∆NP63 and UPK1B that combinatorially mark B-cells (KRT5+∆NP63+UPK1B−), I-cells (KRT5−∆NP63+UPK1B+low) and S-cells (KRT5−∆NP63−UPK1B+high) (27) using immunofluorescence. In wild-type embryos at E14.5, a mono-layered urothelium expressed ∆NP63 only, indicating the presence of differentiated I-cells (Fig. 1A). At E16.5, a two-layered architecture was established containing apically located UPK1B+ S-cells as well as ∆NP63+ B-cells, some of which also expressed KRT5 indicating onset of B-cell differentiation. At E18.5, stratification into three layers with full differentiation of B-, I- and S-cells was present. In Lifr−/− ureters, the temporal development of epithelial stratification and differentiation was largely unchanged with a possible slight delay in B-cell differentiation at E16.5. However, an alteration observed at all analyzed stages was a narrowed ureteral lumen. Moreover, we detected cells located in the lumen at E18.5 (Fig. 1A). Additional in situ hybridization and immunofluorescence analysis revealed that these luminal cells were strongly positive for the S-cell markers Upk1b/UPK1B and Upk3b, thus representing S-cells, while the adluminal cell layer showed some discontinuity in the expression of these markers (Fig. 1B).
Narrowed ureteric lumen, a thickened urothelium, desquamated S-cells and discontinuity of uroplakin (UPK) expression in Lifr−/− mouse embryo ureter development. (A) Hematoxylin and eosin (HE, left) and immunofluorescence (center and right) staining using antibodies against the B-cell marker KRT5 (green, center), the B- and I-cell marker ∆NP63 (red, center), the I- and S-cell marker UPK1B (red, right) and DAPI counterstain (blue) on transverse sections of the proximal ureter of Lifr−/− and Lifr+/+ embryos at the indicated stages revealed reduced ureteral lumen, a thickened urothelium and desquamated cells in Lifr−/− compared to wild-type embryos, while the temporal development of epithelial stratification and differentiation was largely unchanged. (B) RNA in situ hybridization using specific probes for Upk1b and Upk3b expressed in S-cells as well as immunofluorescence staining using an UPK1B antibody (red) and DAPI counterstain (blue) on transverse sections of the proximal ureter of Lifr−/− and Lifr+/+ embryos at E18.5 showed that the desquamated cells corresponded to S-cells and that the expression of these S-cell markers was discontinuous in the adluminal cell layer of Lifr−/− embryos. At least three mice were analyzed per genotype and staining; scale bars represent 50 μm (A and B).
Next, we explored whether these urothelial changes also exist in the bladder of Lifr−/− embryos as well as in the ureter and bladder of Lifr+/− embryos at E18.5. Histological and molecular analyses of frontal bladder sections revealed a thickened and ruffled epithelium with reduced and patchy Upk1b/Upk3b expression and luminal shedding of S-cells in Lifr−/− compared to wild-type embryos (Fig. 2). These histological changes were also seen in both ureter and bladder of approximately 50% of Lifr+/− embryos (Supplementary Material, Fig. S1). We conclude that, in a dose-dependent manner, Lifr is required to maintain the integrity of the S-cell layer in the urothelium of the ureter and the bladder shortly after S-cell differentiation has occurred.
Desquamation of S-cells, thickened and ruffled urothelium and discontinuity of uroplakin (UPK) expression in the bladder of Lifr−/− mouse embryos at E18.5. Hematoxylin and eosin (HE) staining, immunofluorescence staining using an UPK1B antibody (red) and DAPI counterstain (blue) and RNA in situ hybridization using Upk1b- and Upk3b-specific probes on frontal bladder sections revealed desquamated Upk1b/UPK1B and Upk3b expressing cells corresponding to S-cells, a thickened and ruffled urothelium and discontinuous UPK expression in Lifr−/− embryos. At least three mice were analyzed per genotype and staining; scale bars represent 500 μm (images of whole bladder sections) or 50 μm (higher magnification images).
Altered ratios of the three urothelial cell types in Lifr−/− ureters indicate elevated cellular turnover of S-cells due to increased proliferation and enhanced differentiation from I-cells
To determine whether the adhesive properties of urothelial cells are affected by loss of Lifr, we performed immunofluorescence stainings of proteins that mark tight and adherens junctions on transverse ureter sections of wild-type and Lifr−/− embryos at E18.5. We found that the tight junction protein ZO-1 was expressed on the apical side of the majority of S-cells in wild-type and Lifr−/− ureters. Similarly, the adherens junction proteins CDH1 and CTNNB1 as well as the peripheral membrane protein EZRIN were expressed in all layers of the urothelium and appeared unchanged upon loss of Lifr. In S-cells shed into the ureteric lumen of Lifr−/− ureters, all analyzed marker proteins were dramatically downregulated (Supplementary Material, Fig. S2), indicating that S-cells in Lifr−/− ureters lower their expression of junction proteins and loose epithelial identity only during or after shedding.
To explore whether S-cell shedding in ureters of Lifr−/− embryos is associated with altered turnover of urothelial cells, we quantified urothelial cell types by serial section analysis of KRT5, ∆NP63 and UPK1B expression at E18.5. We detected a significant decrease of I-cells and a significant increase of B-cells at both proximal and distal levels of the ureter in Lifr−/− versus wild-type embryos. Additionally, there was a trend toward a higher number of S-cells in Lifr−/− embryos that did not, however, reach statistical significance, but detachment of these cells may lower their actual number (Fig. 3A). Since I-cells are precursors of B- and S-cells (27,28), our finding indicates a higher differentiation rate of I-cell progenitors.
Altered ratios of urothelial I-, B- and S-cells indicating elevated cellular turnover and enhanced proliferation of urothelial cells in ureters of Lifr−/− mouse embryos at E18.5. (A) Quantification of urothelial cell types identified by immunofluorescence staining of B-cell marker KRT5 (green), B- and I-cell marker ∆NP63 (red, upper panel), I- and S-cell marker UPK1B (red, lower panel) and nuclei (blue) revealed a significant decrease of I-cells representing precursors of B- and S-cells, a significant increase of B-cells and a slight increase of S-cells in proximal and distal ureters of Lifr−/− versus Lifr+/+ embryos. (B) Analysis of cell death using a TUNEL assay on transverse sections of proximal and distal ureters at E18.5 does not detect apoptosis in either Lifr+/+ or Lifr−/− urothelium (adluminal side of the dotted white line). The nuclei were stained using DAPI (blue). Inset shows apoptotic bodies (green) in the kidney as a positive control. (C) Immunofluorescence staining of the venous marker EMCN (red) and the pan-endothelial marker IB4 (green) at E18.5 revealed no differences in Lifr+/+ or Lifr−/− urothelium. DAPI (blue) was used to stain the nuclei. (D) Cellular proliferation determined by KI-67 (red) immunofluorescence staining and counterstaining with the B-cell marker KRT5 (green, left) or the B- and I-cell marker ∆NP63 (green, right) and DAPI (blue) at E18.5 in proximal Lifr+/+ and Lifr−/− ureters. Bar graphs show proliferation in percentage determined by the quantification of KI-67-positive cells relative to the total number of KRT5-positive B- versus DAPI-positive I-/S-cells (left), relative to ∆NP63-positive B-/I- versus DAPI-positive S-cells (center) or relative to the total urothelial cell number as detected by DAPI counterstaining (B-/I-/S-cells, right). While proliferation of B-cells or B-/I-cells was scarcely altered, proliferation of I-/S-cells or S-cells was (significantly) elevated in Lifr−/− proximal ureters. At least three mice were analyzed per genotype and staining; scale bars represent 50 μm (A-D). *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ns = not significant.
We evaluated all possible causes of altered cell ratios in a tissue compartment, i.e. presence of programmed cell death, compromised nutrient supply and altered proliferation rates, in proximal and distal Lifr−/− or wild-type ureters at E18.5. When testing for signs of apoptosis by using the dUTP nick end-labeling (TUNEL) assay, no signals were detected in either Lifr−/− or wild-type ureters (Fig. 3B). Additionally, no differences were found in the expression of the vasculature markers endomucin (EMCN) and GSL I isolectin B4 (IB4) using immunofluorescence stainings (Fig. 3C). Thus, neither cell death nor a reduced vascularization resulting in a diminished nutrient supply of the urothelium is causative for the reduction of I-cells and shedding of S-cells in Lifr−/− ureters. Next, we evaluated proliferation rates in the different urothelial compartments of the proximal ureter by staining for the KI-67 antigen that labels proliferating cells and determining KRT5 and ∆NP63 expression to identify B-, I- and S-cells using immunofluorescence. Quantification of the KI-67 labeled cells uncovered significant differences in proliferation rates between Lifr−/− and wild-type ureters (Fig. 3D). While proliferation of B-cells or B-/I-cells was hardly altered, proliferation of I-/S-cells or S-cells was (significantly) elevated in Lifr−/− versus wild-type ureters. When considering all three cell types together, proliferation was significantly higher in the urothelium of Lifr−/− ureters (Fig. 3D). Hence, an increased cellular turnover due to an enhancement of I-cell differentiation and S-cell proliferation accompanies desquamation of S-cells in the urothelium of E18.5 Lifr−/− embryos.
Gene expression microarray analysis detects transcriptional changes in Lifr−/− ureters and identifies affected networks associated with renal and urological disease, cellular development, organismal injury and inflammatory response
To determine the molecular changes in Lifr−/− ureters that may be associated with the identified histological alterations, e.g. S-cell desquamation, we performed microarray analyses on RNA of E18.5 Lifr−/− and wild-type ureters using two independent pools each. We found that the expression of 258 genes was consistently differentially regulated by ≥1.5-fold in Lifr−/− versus wild-type ureters, whereby 49 genes were downregulated, while 209 genes showed an upregulated expression (Supplementary Material, Tables S1 and S2). As expected, Lifr was among the downregulated (−4.8x) transcripts confirming the specificity of the microarray analysis and the genotype of the mice. According to Ingenuity Pathway Analysis (IPA) software, the transcriptional changes in Lifr−/− versus wild-type ureters affected networks associated with (i) organismal injury and abnormalities, renal and urological disease (score, 65); (ii) cellular development, tissue morphology (score, 23); (iii) inflammatory disease, inflammatory response (score, 21); and (iv) immune cell trafficking (score, 15) with a score ≥15, among others (Fig. 4A).
Differentially expressed genes in ureters of E18.5 Lifr−/− versus Lifr+/+ mouse embryos by network analysis of transcriptome data (A) and RNA section in situ hybridization (B) (A) Using a list of genes that were consistently differentially regulated by ≥1.5-fold in microarray analysis on RNA of pooled Lifr−/− versus Lifr+/+ ureters (Supplementary Material, Tables S1 and S2), Ingenuity Pathway Analysis (IPA) allocated a score of ≥15 to gene networks associated with organismal injury and abnormalities, renal and urological disease, cellular development, tissue morphology, inflammatory disease, inflammatory response and immune cell trafficking, among others. (B) RNA in situ hybridization analysis to evaluate microarray results and determine spatial expression of selected candidate genes on proximal ureter sections of E18.5 Lifr−/− and Lifr+/+ ureters. Probes and fold changes as detected by microarray analysis are as indicated. Insets display positive signals in different compartments of the kidney as internal controls. At least three mice were used per genotype and probe; scale bar represents 50 μm. Further genes analyzed by RNA in situ hybridization are shown in Supplementary Material, Fig. S3.
To validate our microarray results and to determine the spatial expression of the candidate genes, we performed RNA in situ hybridization analysis on proximal ureter sections of E18.5 Lifr−/− and wild-type ureters (Fig. 4B, Supplementary Material, Fig. S3). We limited this effort to genes with an intensity average of above 400 and an average fold change of ≤−1.5 or ≥3.0 and additionally analyzed the lymphocyte antigen 6 family member D (Ly6d) gene and the sclerostin domain containing 1 (Sostdc1) gene. We confirmed reduced expression of Lifr (average fold change, −4.8) and Col25a1 (−2.2) in all ureter layers (Fig. 4B). Genes showing an upregulated expression in microarray analysis of the Lifr−/− ureter, Spink1 (+6.2), Lrp2 (+5.9), Slpi (+5.1), Gcnt1 (+4.7), Slc34a1 (+4.3), Serpinf2 (+4.2), Spp2 (+3.9), Ly6d (+2.2) and Sostdc1 (+2.2), were weakly expressed in the adluminal S-cell layer in Lifr+/+ ureters. Expression appeared robustly (Slpi, Gcnt1, Ly6d) and weakly (all others) increased in this domain in the Lifr−/− ureter (Fig. 4B). Notably, some of the top upregulated transcripts, i.e. Spink1, Lrp2, Slpi and Ly6d, have been reported to play an important role in inflammatory response or injury (29–32), and Lrp2, Slc34a1 and Sostdc1 have been associated to the kidney or urinary tract (33–35), potentially linking their ureteral overexpression during development to CAKUT pathogenesis.
Reverse phenotyping: S-cell desquamation also occurs in a CAKUT patient with LIFR haploinsufficiency
Previously, in patient A004 with bilateral renal anomalies, we identified a novel heterozygous de novo LIFR frameshift variant resulting in LIFR haploinsufficiency due to nonsense-mediated decay of the mutant mRNA (16). In the present study, we explored whether the ureteral findings identified here in Lifr−/− and Lifr+/− mice were also detectable in patient A004 with LIFR haploinsufficiency. Microscopic re-evaluation of a hematoxylin- and eosin (HE)-stained section from the distal ureter obtained from patient A004 at the age of 1 year revealed a thickened urothelium and a ruffled surface possibly shedding cells compared to a thinner urothelium and a smooth apical surface in a section from the normal ureter of a 15-year-old male control individual (Fig. 5A). Furthermore, the average number of S-cells per 100 ml urine was significantly enhanced in patient A004 at age 12 compared to age- and sex-matched patient A018B with bilateral renal anomalies not carrying a LIFR variant, as determined for three independent urine samples (Fig. 5B and C). Notably, both patients had undergone nephrectomy and preemptive kidney transplantation due to end-stage kidney disease at 1 year of age, whereby they retained their own bladders. We conclude that S-cell desquamation occurs in the bladder and probably in the ureter of patient A004 with LIFR haploinsufficiency, indicating conserved LIFR-dependent phenotypic changes in the human and murine context.
Urothelial changes in the ureter and increased S-cell desquamation in the urine of CAKUT patient A004 with a heterozygous de novo LIFR frameshift variant resulting in LIFR haploinsufficiency (16). (A) Hematoxylin and eosin (HE) staining of the distal ureter obtained at 1 year of age from CAKUT patient A004 with LIFR haploinsufficiency (16) revealed thickened and ruffled urothelium compared to the ureter of a 15-year-old male control individual. Scale bars represent 50 μm (main image) or 1 mm (inset). (B, C) Quantification of S-cell number per ml of second morning urine obtained from CAKUT patient A004 with LIFR haploinsufficiency at 12 years of age compared to age- and sex-matched CAKUT patient A018B carrying no LIFR variants, who both had nephrectomy at one year of age retaining their own bladders. HE and immunofluorescence staining of UPK1B-positive (red) S-cells (B) revealed significantly more desquamated S-cells per 100 ml urine in patient A004 versus patient A018B (C) demonstrating that S-cells desquamate from the bladder urothelium of patient A004 and suggesting translational relevance of our murine findings. At least three independent urine samples were collected per patient over a period of one year; scale bar represents 50 μm. *P ≤ 0.05.
Discussion
In an effort to identify novel genetic aberrations underlying CAKUT pathogenesis and to characterize their consequences on a cellular and molecular level, we recently reported a spectrum of morphological changes in the kidney and the urinary tract of patients with rare heterozygous deleterious LIFR variants and of Lifr−/− as well as Lifr+/− prenatal mouse embryos, including reduced ureteral lumen, hydronephrosis and renal hypodysplasia (16). Similarly, reduced nephrogenesis was reported in GP130-deficient mice (36), and overexpression of a dominant negative Lifr impaired normal kidney development in Xenopus laevis (22) implicating LIF signaling via the GP130/LIFR heterodimeric receptor complex in renal development of vertebrates and lower vertebrates. Here, we demonstrated an additional phenotype associated with reduced or lacking LIFR function in the murine and human urinary tract, namely, increased luminal shedding of S-cells from the urothelium of both the ureter and bladder, in association with a proliferative state and an increased differentiation of I-cells. Our molecular analysis suggests that this phenotype may arise from a failure to suppress an antimicrobial gene program in the S-cell layer.
LIFR suppresses an antimicrobial gene program in the S-cell layer
Previous studies have shown that during embryogenesis, S-cells arising from I-cells undergo endoreplication and withdraw from the cell cycle (26,27,37). Endowed with tight junctions and specialized semicrystalline arrays of apical UPK proteins, S-cells are critical for urothelial barrier function toward urine osmolarity and toxins under conditions of homeostasis (23,25,28,38). During infection, pathogens encounter S-cells as the first line of defense and invade them. Consequently, S-cell shedding and subsequent regeneration present an efficient and rapid response to remove trapped pathogens and prevent them from reaching deeper ureteral layers and the blood stream (39).
Here, when analyzing the urothelial changes in mice and patient A004 with LIFR deficiency or haploinsufficiency, we observed desquamated S-cells in combination with increased cell proliferation and enhanced differentiation of I-cells mimicking the cellular changes associated with bacterial infection or injury. This came as a surprise because the mice were kept under specific pathogen-free conditions and no recurrent urinary tract infections were diagnosed in the patient in the last few years. Moreover, using IPA network analysis on transcriptome data from Lifr−/− and Lifr+/+ ureters, we identified networks associated with organismal injury as well as inflammatory disease, inflammatory response or immune cell trafficking to be affected by Lifr deficiency. Consistently, some of the genes showing a strongly upregulated ureteral expression in Lifr−/− embryos according to microarray analysis, Spink1, Lrp2, Slpi and Ly6d, have been reported to be involved in inflammatory processes as a response to infection or injury (29–32). Of these, expression in the adluminal S-cell layer of Lifr−/− embryos was robustly increased in Ly6d encoding a protein involved in infection that is an inflammation marker (30,40) and Slpi coding for secretory leukocyte peptidase inhibitor, a suppressor of inflammatory response (31). SLPI is mainly found as a secretory product of epithelial cells in the respiratory, digestive and reproductive tracts as well as in the kidney. It inhibits bacterial growth, controls the processing of inflammatory mediators and, thus, counteracts excessive inflammatory responses to initiate healing processes (31,41). Together, these findings indicate that Lifr deficiency provokes an overshooting microbial defense program, resembling a ‘chronic’ condition, in which the natural course of inflammation is lost, resulting in disease progression instead of recovery.
An important role in the regulation of inflammation is not without precedence for IL6 (39,42–44) or other cytokines of the IL6 family, such as LIF (45). Il6-deficient mice were unable to establish a normal immune response to localized tissue damage (43,44), and a patient with a biallelic mutation in the IL6ST gene encoding the transmembrane receptor GP130 utilized by IL6 was reported to present with recurrent infections and an impaired acute-phase response (46). For LIF, one of two ligands that signal through a GP130/LIFR dimeric receptor (17), pro-inflammatory and anti-inflammatory actions have been described (45). An anti-inflammatory role for LIFR was detected in skeletal muscle regeneration (47), corroborating our data in the urothelium that LIFR may act in an anti-inflammatory manner.
LIFR function is mediated by STAT signaling in urothelial S-cells
Different effector pathways including JAK/STAT, MAPK and PI3K are known to be stimulated by LIFR signaling (17). In metanephric mesenchymes, i.e. embryonic structures giving rise to the kidney, dissected from rat embryos, exposure to LIF induced STAT3 phosphorylation (36). In a cellular model expressing LIFR mutant proteins, which had been identified in CAKUT patients, compared to wild-type LIFR, we previously detected reduced STAT3 phosphorylation (16). Together, these data suggest that STAT3 represents a crucial effector pathway of LIFR signaling in the urinary tract. Moreover, STAT3 was shown to have a critical role in protection from inflammation-induced heart damage (48). Additionally, conditional deletion of STAT3 in murine bone marrow cells during hematopoiesis causes Crohn’s disease-like inflammatory cell infiltration in the small and large intestine, increased proliferation of cells of the myeloid lineage and pseudo-activated innate immune response during hematopoiesis (49). Here, using pharmacological inhibition experiments in ureter explant cultures of wild-type mouse embryos, we tested a possible involvement of STAT3 in urothelial maintenance. Our preliminary ex vivo experiments show that STAT3 inhibition leads to increased S-cell shedding (Supplementary Material, Fig. S4) indicating that, in fact, JAK/STAT3 acts downstream of LIFR to suppress inflammation-like changes in the urothelium during homeostasis. In turn, these data suggest that mounting an antibacterial response occurs at least partly by inhibiting LIFR signaling and/or STAT3 function.
Ureteral changes induced by Lifr deficiency or LIFR haploinsufficiency and CAKUT
Most processes that induce cell removal in the urothelium, such as mechanical or chemical damage or infection, result in an inflammatory response and prolonged cell desquamation and cause hyperplasia of the urothelium (26). Here, we propose that such pathology, i.e. S-cell desquamation and elevated cell proliferation in the urothelium, can be caused genetically by LIFR deficiency and likely contributes to the renal and urinary tract malformations observed in Lifr-deficient mice and in CAKUT patients carrying pathogenic LIFR variants (16). The consequences of a thickened urothelium and increased S-cell shedding are a narrow lumen, especially in the distal ureter, which creates a physical obstruction resulting in hydro- or megaureter, hydronephrosis (1) and possibly also renal dysplasia. Although the loss of S-cells may be compensated by increased I-cell proliferation and differentiation, it is likely that under the ‘chronic’ conditions induced by LIFR deficiency, the S-cell barrier function is weakened, and underlying cells and tissues are at least temporarily exposed to urinary toxins increasing the risk for tissue damage and inflammation (28) and enhancing the ureteral stenosis. Moreover, reduced Upk expression due to paucity of S-cells might play an additional role in CAKUT pathogenesis since dilatation of the renal pelvis as well as clear-cut hydronephrosis was detected in Upk3-deficient mice (38). Similar to our findings in Lifr-deficient mice, we, here, observed thickened and ruffled urothelium in the ureter and significantly more desquamated S-cells in the urine of patient A004 with LIFR haploinsufficiency presenting with obstructive megaureter, massive hydronephrosis and renal dysplasia as well as renal agenesis (16) compared to an age- and sex-matched control CAKUT patient without pathogenic LIFR variants. These findings suggest conserved Lifr-dependent urinary tract anomalies in mice and humans and, consequently, relevance of our data from Lifr-deficient mice for CAKUT patients carrying pathogenic LIFR variants.
In ureters of E18.5 mice with Lifr deficiency, our microarray expression data revealed strong upregulation of three genes associated with the kidney or urinary tract, i.e. Lrp2, Slc34a1 and Sostdc1, potentially linking their ureteral overexpression during development to CAKUT pathogenesis. Sostdc1, a gene expressed in the developing murine kidney from E13.5 and most abundantly at E17.5 (35), encodes a secreted protein that interacts with WNT co-receptor LRP6 (50) and also functions as a bone morphogenic protein (BMP) antagonist, particularly of BMP7 (35,51), thus impacting two signaling pathways involved in kidney and nephron induction (1). Increased SOSTDC1 expression, commonly detected in transcriptomes from ureter samples of CAKUT patients (52), lowered the contraction frequency of wild-type ureters and potentially contributed to the phenotypical changes of Gata2-deficient ureters (53), and BMP7 mutations have been reported in human CAKUT (54) suggesting that overexpression of SOSTDC1 contributes to murine and human CAKUT pathogenesis. Lrp2 encoding LDL-receptor-related protein 2/megalin is an endocytic receptor expressed in S-shaped bodies and widely accepted to be important during the development of the proximal renal tubule (33). Internalizing a variety of ligands, LRP2/megalin is thought to play a role in cholesterol uptake during embryogenesis (33). Loss of this receptor in congenital or acquired disease results in renal reabsorption defects such as the Donnai–Barrow syndrome (OMIM #222448) (55). It is conceivable that Lifr deficiency-induced overexpression of Lrp2/megalin may impact cholesterol metabolism, as shown for hepatocyte nuclear factor-1β encoded by a gene known to be associated with CAKUT when mutated (56). Slc34a1 coding for solute carrier family 34 (type II sodium/phosphate cotransporter), member 1 (NaPi-IIa), is also expressed in the proximal renal tubule where it reabsorbs phosphate to ensure the maintenance of phosphate homeostasis (34). The disorders associated with mutations in Slc34a1 are hypophosphatemic nephrolithiasis/osteoporosis (OMIM #612286) (57), Fanconi renotubular syndrome 2 (OMIM #613388) (58) and idiopathic infantile hypercalcemia (OMIM #616963) (59), all characterized by an impairment of renal phosphate handling and increased urinary phosphate and calcium excretion (60). Therefore, detecting upregulated Slc34a1 expression in Lifr deficiency and idiopathic hypercalcemia in 15% of infants with renal dysplasia (61) may point toward a link between calcium and phosphate homeostasis and CAKUT.
In summary, our study identifies an upregulated expression of the kidney/urinary tract-associated genes Lrp2, Slc34a1 and Sostdc1 in Lifr-deficient ureters. Furthermore, our results link Lifr deficiency in mice and LIFR haploinsufficiency in humans to inflammatory processes in the urothelium of the urinary tract that may contribute to CAKUT pathogenesis, possibly mediated by a reduction of STAT signaling.
Materials and Methods
Patients and patient samples
The study was approved by the Ethics Board of Hannover Medical School, Hannover, Germany; each family provided informed consent for participation in the study. Ureter sections and urine samples from CAKUT patient A004 carrying a novel heterozygous de novo frameshift variant in the LIFR gene (NM_002310.5), LIFR:c.1273_1276delGTTA, and p.(Val425Ilefs*2) (chromosomal position according to GRCh37/hg19 chr5:38506022_38506025) (16) were analyzed and compared with ureter sections from a control individual and urine samples from CAKUT patient A018B not harboring any LIFR variants according to whole-exome sequencing. Male patient A004 presented with right-sided renal agenesis and left-sided dysplastic kidney with hydronephrosis and obstructive megaureter at birth and received preemptive kidney transplantation and left-sided nephrectomy due to end-stage kidney disease at the age of 1 year (16). The ureter from patient A004 obtained during nephrectomy was compared to a normal ureter from a male individual obtained at 15 years of age. Urine sampling was done when patient A004 was 12 years of age and compared to urine samples from 14-year-old male patient A018B who presented with bilateral renal dysplasia and vesicoureteral reflux at birth and received preemptive kidney transplantation due to end-stage kidney disease at 1 year of age.
Animals
Animals were kept in accordance with the National Institute of Health guidelines for the care and use of laboratory animals. All experiments were approved by the Ethics Board of the Lower Saxony State Office for Consumer Protection and Food Safety. The Lifrtm1lmx (synonym Lifr+/−) mouse line (19) was purchased from The Jackson Laboratory (#002402, Bar Harbor, ME, USA) and maintained on an NMRI outbred background. Lifr−/− embryos were obtained from heterozygous matings. Wild-type littermates served as controls. For timed pregnancies, vaginal plugs were checked in the morning after mating, and noon was considered as E0.5. Since Lifr−/− mice die perinatally (18,19), embryos were analyzed at E18.5 and earlier, i.e. at E16.5 and E14.5. Embryos and urogenital systems were dissected in phosphate-buffered saline (PBS). Specimens were fixed in 4% paraformaldehyde (PFA) in PBS followed by dehydration in methanol and stored in 100% methanol at −20°C until further processing. For genotyping and sex determination, genomic DNA was extracted from ear tissue of adult mice or embryonic tissue biopsies, and PCR using specific primers was performed as described before (16).
Ureter ex vivo cultures
The ureters of E18.5 wild-type embryos from NMRI matings were dissected in PBS, explanted onto 0.4-μm polyester membrane Transwell supports (#3450, Corning Inc., Lowell, MA, USA) and cultured in an air–liquid interface as previously described (27). Culture medium was changed every second day. To analyze the effect of STAT3 inhibition, the ureters were cultured for 6 days in the presence of dimethyl sulfoxide (DMSO) or pSTAT3-inhibitor S3I-201 (Axon 2313/Batch 1, Axon Medchem BV, Groningen, the Netherlands) dissolved in DMSO and added to the medium at a final concentration of 200 μM. The ureters were fixed on the Transwell membrane with ice-cold methanol, postfixed in 4% PFA in PBS, dehydrated in methanol and stored in 100% methanol at −20°C until further processing.
Histological analysis
For histological analysis, urogenital systems of at least three mice for each genotype were paraffin-embedded, and transverse sections of 5-μm thickness were generated. Histological analysis was additionally performed on 5-μm transverse sections of formalin-fixed paraffin-embedded human ureter tissue from CAKUT patient A004 and an individual with a normal ureter. HE staining was performed according to standard protocols.
Immunofluorescent detection of antigens
Immunofluorescence analysis on 5-μm transverse paraffin-embedded ureter sections was done as previously described (27). The following antibodies were used: mouse anti-UPK1B (dilution 1:250, #WH0007348M, Sigma-Aldrich, St. Louis, MO, USA), rabbit anti-ΔNP63 (1:250, #619001, BioLegend, San Diego, CA, USA), rabbit anti-KRT5 (1:250, #PRB-160P, Covance, Princeton, NJ, USA), anti-CDH1 (1:200, kind gift of Rolf Kemler, Max Planck Institute of Immunobiology and Epigenetics, Freiburg, Germany), rabbit anti-CTNNB1 (1:200, #sc-7199, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), mouse anti-Ezrin (1:200, #MS-661, Thermo Fisher Scientific, Waltham, MA, USA), rabbit anti-ZO-1 (1:100, #61-7300, Thermo Fisher Scientific), rat anti-mouse KI-67 (1:50, #M7249, Agilent/Dako, Santa Clara, CA, USA), rat anti-mouse EMCN (1:10, V.7C7, kind gift from D. Vestweber, Max Planck Institute of Molecular Biomedicine, Münster, Germany) and fluorescein-labeled isolectin B4 (IB4, 1:100, #FL-1101, Vector Laboratories, Burlingame, CA, USA). Fluorescence staining was performed using Alexa 488/555-conjugated secondary antibodies (1:250, #A-11034 or #A-21422, Thermo Fisher Scientific) or biotin-conjugated secondary antibodies (1:250, #111-065-033, Dianova, Hamburg, Germany) and the TSA Tetramethylrhodamine Amplification Kit (1:100, #NEL701001KT, PerkinElmer, Waltham, MA, USA). Labeling with primary antibodies was performed at 4°C overnight after citrate antigen unmasking (#H-3300, Vector Laboratories). Secondary antibodies were incubated at room temperature for 2 h in Tris-buffered blocking solution. Counterstaining of the nuclei was performed using 4′,6-diamidino-2-phenylindole (DAPI). For each developmental stage, genotype and experiment, at least three independent specimens were analyzed.
RNA in situ hybridization analysis
Non-radioactive RNA in situ hybridization analysis on transverse ureter sections followed a published protocol (62). For each genotype and experiment, at least three independent specimens were analyzed.
Apoptosis and cell proliferation assays
Apoptosis was analyzed on 5-μm transverse paraffin-embedded sections using a TUNEL assay and the ApopTag Plus Fluorescein In Situ Apoptosis Detection Kit (#S7111, Merck, Darmstadt, Germany) following the manufacturer’s instructions. Ureter sections of three independent specimens were analyzed per genotype at E18.5. Detection of apoptosis in the kidney served as a positive control. Cell proliferation rates were analyzed by the detection of KI-67-positive cells on 5-μm transverse ureter sections. A minimum of four consecutive sections each of both proximal ureters from three independent specimens were analyzed per genotype. Cell proliferation rate was defined as the number of KI-67-positive nuclei relative to the total number of nuclei in the urothelium of the ureter as detected by DAPI counterstaining.
Image documentation
Images of ureter sections stained by immunofluorescence or RNA in situ hybridization as well as of ureter cultures were acquired using a DM5000 microscope with a DFC300FX digital camera or a DM6000 microscope with a DFC350FX digital camera (all Leica Microsystems, Wetzlar, Germany). HE-stained sections were scanned using an Aperio AT2 Scanner (Leica Microsystems). Digital images were processed using ImageScope v11.2.0.780 software (Leica Microsystems) and Adobe Photoshop CS6 (Adobe Systems Incorporated, San José, CA, USA).
Gene expression microarray and network analysis
Two independent pools of 20 ureters each, one derived from male embryos, the other from female embryos, were collected from E18.5 Lifr−/− embryos and Lifr+/+ littermates, respectively. The total RNA from each pool was extracted using the peqGOLD RNApure Kit (PeqLab/VWR International, Radnor, PA, USA), and RNA was labeled and hybridized to Agilent mRNA Microarrays (Ag4x180K_1M_DC) assaying the expression levels of around 30 000 genes by the Research Core Unit Transcriptomics at Hannover Medical School, Hannover, Germany. Gene expression data were normalized, and transcripts showing a ≥1.5 fold change in transcriptomes of both sex-matched pools of Lifr−/− versus Lifr+/+ ureters were identified using Microsoft Excel 2010 software (Microsoft Corporation, Redmond, WA, USA). Networks of the genes showing a ≥1.5 fold change were identified using IPA (Qiagen, Hilden, Germany). To evaluate microarray results on a cellular level, RNA in situ hybridization analysis on transverse ureter sections was performed for selected genes as described in the results section.
Quantification of S-cells in the urine of CAKUT patients
To determine the number of S-cells in the urine of the CAKUT patient (A004) carrying the LIFR frameshift variant and of the age- and sex-matched control CAKUT patient (A018B) without LIFR variants, at least three samples of the second morning urine of both patients were collected over a time period of 1 year. The volume of each sample was determined for normalization, cells were harvested by centrifugation (400 × g, 10 min), washed in PBS, mounted on a microscope slide using a Shandon Cytospin 2 centrifuge (Marshall Scientific, Hampton, NH, USA) and treated with 4% PFA in PBS overnight for fixation.
Statistical analysis
Data were expressed as mean ± standard deviation. Statistical analysis was performed using a two-tailed student’s t-test. Differences were considered significant (P ≤ 0.05, *), highly significant (P ≤ 0.01, **) or extremely significant (P ≤ 0.001, ***).
Acknowledgements
The authors wish to thank the patients and their families for participating in this study. We wish to acknowledge Wolfgang H. Ziegler, Birga Sötje and Anja Ziolek for their excellent support in urine preparation, Jan Hinrich Bräsen for providing human ureter sections from nephrectomy samples, and the Research Core Unit Transcriptomics for performing microarray analyses and for assistance in data evaluation, all at Hannover Medical School, Hannover, Germany.
Conflict of Interest Statement. None declared.
Funding
This work was supported by grants from the Deutsche Forschungsgemeinschaft (DFG) to Anne Christians [KO5614/2-1] and Andreas Kispert [KI728/12-1].
References
Author notes
The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors.
The authors wish it to be known that, in their opinion, the last two authors should be regarded as joint Last Authors.
