Modelling TFE renal cell carcinoma in mice reveals a critical role of WNT signaling

Abstract

TFE-fusion renal cell carcinomas (TFE-fusion RCCs) are caused by chromosomal translocations that lead to overexpression of the TFEB and TFE3 genes (Kauffman et al., 2014). The mechanisms leading to kidney tumor development remain uncharacterized and effective therapies are yet to be identified. Hence, the need to model these diseases in an experimental animal system (Kauffman et al., 2014). Here, we show that kidney-specific TFEB overexpression in transgenic mice, resulted in renal clear cells, multi-layered basement membranes, severe cystic pathology, and ultimately papillary carcinomas with hepatic metastases. These features closely recapitulate those observed in both TFEB- and TFE3-mediated human kidney tumors. Analysis of kidney samples revealed transcriptional induction and enhanced signaling of the WNT β-catenin pathway. WNT signaling inhibitors normalized the proliferation rate of primary kidney cells and significantly rescued the disease phenotype in vivo. These data shed new light on the mechanisms underlying TFE-fusion RCCs and suggest a possible therapeutic strategy based on the inhibition of the WNT pathway.

Keywords: TFEB; WNT pathway; cancer biology; human biology; medicine; mouse; renal cell carcinoma.

Figure 1.. Tfeb overexpressing mice display cystic kidneys.

Morphological analyses were performed on Cdh16^Cre and Cdh16^Cre::Tfeb^fs, and on tam-treated Cdh16^CreErt2 and Cdh16^CreErt2::Tfeb^fs mice. (A) Representative images of the abdominal cavity at P90. (B) Kidney size at different stages (p=days post-natal). (C) Relative ratio of kidney-to-body weight (KW/BW). Data from males (M) and females (F) are shown separately as means of Cdh16^Cre::Tfeb^fs to Cdh16^Cre KW/BW ratio. Three-way Anova was applied (factors: gender, time, genotype). (D) Evaluation of the survival of Cdh16^Cre::Tfeb^fs and tam-treated Cdh16^CreErt2::Tfeb^fs mice. Mantel-Cox test was applied (Cdh16^CreErt2::Tfeb^fs tam P12/tam P14 p-value 0.02; Cdh16^CreErt2::Tfeb^fs tam P12/P30 p-value<0.0001). (E) Haematoxylin and Eosin (HE) staining of kidneys. Enlarged panels show cyst growth over time. (F) Number (left graph) and area (right graph) of kidney cysts in Cdh16^Cre::Tfeb^fs, and Cdh16^CreErt2::Tfeb^fs mice. Number of cysts is shown as an average (± SEM) with bars sub-divided according to the dimension of the cysts. Cyst areas are presented as independent values (dots) with lines representing the means. Three-way (cyst number) and two-way (cyst area) Anova was applied. Cor, cortex; oMed, outer medulla; iMed, inner medulla. (G) Cadherin16 (CDH16) staining of kidneys from P30 mice. (H) Megalin, THP and AQP2 stainings in P90 Cdh16^CreErt2::Tfeb^fs mice. (I) PAS and Sirius Red staining. PAS staining shows the presence of single-layered or multi-layered cysts, and the presence of Clear Cells (CCs). SR staining shows areas of interstitial fibrosis, multi-layered basement membrane and protein casts. Asterisks, protein casts; sCy, simple Cysts; mCy, multilayered Cy; IF, Interstitial Fibrosis; mBM, multi-layered Basement Membrane. (*p<0.05, **p<0.01, ***p<0.001). DOI: http://dx.doi.org/10.7554/eLife.17047.002

Figure 1—figure supplement 1.. Generation of transgenic mouse lines with kidney-specific Tfeb overexpression.

(A) Map of the transgene: Tfeb-3xFLAG was inserted after the chicken actin promoter (CAG) and the chloramphenicol acetyltransferase (CAT)-SV40pA flanked by loxP sites. The latter can be removed by CRE recombinase, resulting in the overexpression of the Tfeb gene under the control of the strong CAG promoter. Two different CRE lines were used: (1) a constitutive kidney-specific Cdh16^Cre(Cadherin 16) and (2) a tamoxifen-inducible Cdh16^CreErt2. (B, C) Representative genotypes of littermates. (B) Lanes 1 and 4 indicate double heterozygous Cdh16^Cre::Tfeb^fs mice as they carry both the Cdh16^Cre (420 bp CRE band, 200 bp wt band) and the Tfeb^fs transgenes (700 bp) (M, marker; B, Blank). (C) Lane 1 indicates a double heterozygous Cdh16^CreErt2::Tfeb^fs mouse as it carries both the Cdh16^CreErt2 (507 bp Cre band, 388 bp wt band) and the Tfeb^fs transgenes (700 bp). (D) Real-time PCR analysis of Tfeb-3xFLAG mRNA levels performed on Cdh16^Cre and Cdh16^Cre::Tfeb^fs mice at different stages (P0, P12, P30) and on Cdh16^CreErt2and Cdh16^CreErt2::Tfeb^fs mice induced with tamoxifen at P12 and sacrificed at P90. Values are shown as the average (± SEM) of at least three animals per time point and genotype (*p<0. 05, **p<0.01, two-sided, Student’s t test). (E) Immunoblot analysis using an anti-Flag antibody to determine the expression of Tfeb-3xFLAG protein in Cdh16^Cre::Tfeb^fs and tam-treated Cdh16^CreErt2::Tfeb^fs mice. Each replicate is a different biological sample. DOI: http://dx.doi.org/10.7554/eLife.17047.003

Figure 1—figure supplement 2.. Renal-specific Tfeb overexpression results in kidney enlargement and failure.

(A) Tamoxifen injection of Cdh16^CreErt2::Tfeb^fs mice at P12, P14 and P30 with representative images at the time of sacrifice (tam, tamoxifen). (B) Relative ratio of kidney to body weight (KW/BW) evaluated in tam-treated Cdh16^CreErt2::Tfeb^fs mice at P90. Two-way Anova was applied (factors: treatment, genotype). (C) Blood urea and albuminuria levels in Cdh16^Cre::Tfeb^fs and Cdh16^CreErt2::Tfeb^fs mice. Values are shown as means (± SEM) of at least three Cdh16^Cre::Tfeb^fs and Cdh16^CreErt2::Tfeb^fs mice and are normalized versus the control animals (Cdh16^Creand Cdh16^CreErt2). (D) High-frequency ultrasound (HFUS) images of kidneys from P30 Cdh16^Cre::Tfeb^fs mice. (E) Images and PAS staining from tam-treated Cdh16^CreErt2::Tfeb^fs mice at P90. DOI: http://dx.doi.org/10.7554/eLife.17047.004

Figure 1—figure supplement 3.. Characterization of cyst origin in Cdh16^Cre::Tfeb^fs and Cdh16^CreErt2::Tfeb^fs mice.

IHC staining of megalin, THP and AQP2 at different time points. Insets are enlargements of representative areas of interest. Larger cysts (denoted by an asterisk) are negative for all the markers tested. DTcy = Distal Tubules cysts; CDcy = Collecting Ducts cysts. DOI: http://dx.doi.org/10.7554/eLife.17047.005

Figure 2.. Kidney-specific Tfeb overexpression is associated with cancer development in Cdh16^Cre::Tfeb^fs and Cdh16^CreErt2::Tfeb^fs mice.

(A) ¹⁸F-FDG PET/CT scan on P30 Cdh16^Cre::Tfeb^fs mice. (B) HE and Ki67 staining performed on Cdh16^Cre::Tfeb^fs mice at P1, P12, P30 and 5 months. Beginning at P12 the increase in cyst size is associated with an increase in papillary proliferation that becomes completely neoplastic by 5 months. NP, Neoplastic Papillae. (C–H) Representative images of neoplastic lesions at different stages: (C) neoplastic nodules (arrows) in P12 Cdh16^Cre::Tfeb^fs mice; (D) micropapillae (arrows) and (E) hobnail-like cells (arrows) in P30 Cdh16^Cre::Tfeb^fs mice; (F) mitotic spindles (arrows) in 5-month-old Cdh16^Cre::Tfeb^fs mice; (G) microcalcifications (asterisk) in tam-treated Cdh16^CreErt2::Tfeb^fs mice induced at P14 and sacrificed at 5 months; (H) neoplastic nests (NN) and clear cells (CCs) in tam-treated Cdh16^CreErt2::Tfeb^fs mice induced at P12 and sacrificed at P90. (I) HE staining of neoplastic lesions invading the surrounding stroma (arrows) in Cdh16^Cre::Tfeb^fs and in tam-treated Cdh16^CreErt2::Tfeb^fs mice. (L) Liver metastases in 5 month-old Cdh16^Cre::Tfeb^fs mice stained for HE, Ki67, PAX8 and CK7. DOI: http://dx.doi.org/10.7554/eLife.17047.006

Figure 3.. Activation of ErbB and WNT signaling pathways in kidneys from Cdh16^Cre::Tfeb^fs mice.

Transcriptional and biochemical analyses were performed on Cdh16^Cre and Cdh16^Cre::Tfeb^fs mice. (A,B) Tables show the relative increase of genes related to the ErbB (A) and WNT (B) pathways in the microarray analyses performed on kidneys from P0 Cdh16^Cre::Tfeb^fs mice. Graphs show real-time PCR validations performed on kidneys from Cdh16^Cre::Tfeb^fs mice at different stages (P0, P12, P30). Data are shown as the average (± SEM) of at least three Cdh16^Cre::Tfeb^fs mice normalized versus wild-type mice. (C,D) Immunoblot analyses performed on (C) P30 kidney tissues and (D) primary kidney cells isolated from Cdh16^Cre::Tfeb^fs mice to evaluate ErbB and WNT activation status. Each replicate is a distinct biological sample. ErbB signaling was assessed by looking at phosphoAKT (Ser473) to total AKT ratio, and phosphoERK1 (T202/Y204)/ERK2(T185/Y187) to total ERK ratio; WNT signaling was assessed by quantifying β-catenin and CCND1 (Cyclin D1) protein levels. Graphs represent the densitometry quantification of Western blot bands. Values are normalized to actin when not specified and are shown as an average (± SEM) (*p<0.05, **p<0.01, ***p<0.001, two-sided, Student’s t test). DOI: http://dx.doi.org/10.7554/eLife.17047.007

Figure 3—figure supplement 1.. ErbB and WNT transcriptional profiles in Cdh16^CreErt2::Tfeb^fs mice.

Transcriptional analyses performed on Cdh16^CreErt2::Tfeb^fs mice. (A,B) mRNA levels of previously validated genes belonging to the WNT (left graphs) and ErbB (right graphs) signaling pathways assessed in P90 Cdh16^CreErt2::Tfeb^fs mice induced at (A) P14 and at (B) P30 with tamoxifen respectively. Data are shown as the average (± SEM) of at least Cdh16^CreErt2::Tfeb^fs mice and values are normalized to the wild-type line. (*p<0.05, **p<0.01, ***p<0.001, two-sided Student’s t test). DOI: http://dx.doi.org/10.7554/eLife.17047.010

Figure 3—figure supplement 2.. Biochemical analysis of ErbB signaling.

Immunoblot analysis performed on P90 kidneys from Cdh16^Cre::Tfeb^fs mice (A) and P90 Cdh16^CreErt2::Tfeb^fs animals induced with tamoxifen at P14 (B) and at P30 (C), respectively. Each replicate is a different biological sample. ErbB was analyzed by quantifying phosphoAKT (Ser473) to total AKT, and phosphoERK1 (T202/Y204)/ERK2(T185/Y187) to total ERK; graphs are the densitometry quantifications of Western blot bands normalized to wild-type line and are shown as an average (± SEM) (*p<0.05, **p<0.01, ***p<0.001, two-sided Student’s t test). DOI: http://dx.doi.org/10.7554/eLife.17047.011

Figure 4.. Molecular and histological analysis of WNT signaling.

(A,B) Western blot analysis performed on (A) P30 and (B) P90 kidneys from Cdh16^Cre::Tfeb^fs mice to assess WNT signaling activation by looking at different proteins related to this pathway. Each replicate is a distinct biological sample. p-LRP6 (Ser1490)/LRP6, active β-catenin, β-catenin and p-GSK3β (Ser9)/GSK3β protein levels were quantified by densitometry analysis of the Western blot bands. Values are normalized to actin when not specified, and are shown as an average (± SEM) (*p<0.05, **p<0.01, ***p<0.001, two-sided Student’s t test). (C) Immunohistochemistry staining of CDH16, β-catenin and active β-catenin proteins performed on P30 kidney tissues from Cdh16^Cre::Tfeb^fs mice. DOI: http://dx.doi.org/10.7554/eLife.17047.014

Figure 4—figure supplement 1.. Molecular analysis of WNT signaling pathway in Cdh16^CreErt2::Tfeb^fs animals.

(A,B) Immunoblot analysis of WNT-related proteins performed on P90 Cdh16^CreErt2::Tfeb^fs animals induced with tamoxifen at P14 (A) and at P30 (B). Each replicate is a different biological sample. Graphs show densitometry analysis of the Western blot bands. Values are normalized to actin when not specified, and are shown as an average (± SEM) (*p<0.05, **p<0.01, ***p<0.001, two-sided Student’s t test). DOI: http://dx.doi.org/10.7554/eLife.17047.015

Figure 5.. Inhibition of WNT signaling rescues the hyper-proliferative phenotype of kidney cells from Cdh16/Tfeb mice.

(A,B) Activity of the TCF/LEF reporter TOP-FLASH. Luciferase activity after co-transfection of β-catenin and TCF plasmids in HEK293 (A) and HK2 (B) cells with and without Tfeb overexpression. Values are shown as an average (± SEM) of each point in duplicate, normalized to the Renilla values and to the basal condition. Data are representative of three independent experiments. (C) MTT tetrazolium reduction assay (MTT) was used to evaluate proliferation of primary kidney cells derived from Cdh16^Cre::Tfeb^fs mice. Values are shown as an average (± SEM) of each point in triplicate and normalized versus wild-type mice. Data are representative of three independent experiments. (D) MTT proliferation assays of primary kidney cells treated independently with two WNT signaling inhibitors, PKF118-310 and CGP049090, added at different dosages for 24 hr. 0 µm represents the basal proliferation of cells. Values are shown as means (± SEM) of three replicates per point normalized to the vehicle (DMSO), added at the same concentration, and versus the Cdh16^Cre cells without drug treatment. Results are representative of three independent experiments. Two-way Anova was applied (factors: cell genotype, treatment). (E) Immunoblot analysis on primary kidney cells treated with Drug (PKF118-310) or Vehicle (DMSO) for 24 hr at 1.6 µM. Graphs show the densitometry quantifications of Western blot bands. Values are normalized to actin and are shown as averages (± SEM) (Cor, cortex; Med, medulla). (*p<0.05, **p<0.01, ***p<0.001). DOI: http://dx.doi.org/10.7554/eLife.17047.016

Figure 6.. Treatment with WNT inhibitor attenuates cystic and neoplastic phenotypes.

Morphological, histological and biochemical analyses performed on Cdh16^Cre::Tfeb^fs mice treated with Vehicle (DMSO) or Drug (PKF118-310). (A,B) Kidney images (A) and sizes (KW/BW) (B) from Cdh16^Cre::Tfeb^fs mice injected intraperitoneally (IP) either with vehicle or drug at 0.85 mg/kg. KW/BW ratios are shown as means (± SEM) and values are normalized to the Cdh16^Cre animals treated with vehicle. Two-way ANOVA was applied (factors: treatment, genotype). (C) Ki67 staining of kidneys from Cdh16^Cre::Tfeb^fs mice after treatment with vehicle or drug. Insets are enlargements of a single cyst. (D) Quantification of several parameters related to cysts and papillae performed on kidney sections from vehicle- and PKF118-310-treated Cdh16^Cre::Tfeb^fs mice. (E) Gpnmb mRNA fold activation in kidneys from Cdh16^Cre::Tfeb^fs and tam-treated Cdh16^CreErt2::Tfeb^fs mice at different stages. Values are shown as means (± SEM) of at least three mice and each group is normalized to the proper control (respectively Cdh16^Cre and tam-treated Cdh16^CreErt2). (F) Gpnmb fold activation in kidneys from Cdh16^Cre::Tfeb^fs mice treated with vehicle or PKF118-310. Values are shown as means (± SEM) of at least three animals per group and are all normalized versus the Cdh16^Cre mice treated with vehicle. (*p<0.05, **p<0.01, ***p<0.001, two-sided Student’s t test). DOI: http://dx.doi.org/10.7554/eLife.17047.017

Figure 6—figure supplement 1.. In vivo treatment of Cdh16^Cre::Tfeb^fs mice with the WNT inhibitor PKF118-310 partially rescues cystic and neoplastic phenotypes.

Measurements of different parameters related to the cystic and papillary phenotype on six animals treated with vehicle (DMSO) and six animals treated with drug (PKF118-310). Values are shown as means (± SEM) when appropriate and are represented separately for each animal. DOI: http://dx.doi.org/10.7554/eLife.17047.019

Figure 6—figure supplement 2.. In vivo treatment of Cdh16^Cre::Tfeb^fs mice with the PKF118-310 drug inhibits WNT pathway overactivation.

(A) Tfeb, Cyclin D1, Myc and Axin2 mRNA levels in kidneys from Cdh16^Cre and Cdh16^Cre::Tfeb^fs mice treated with vehicle or PKF118-310. Values are shown as the average (± SEM) of four animals per group, and are all normalized to the Cdh16^Cre mice treated with vehicle. (B) Biochemical analysis performed on Cdh16^Cre and Cdh16^Cre::Tfeb^fs mice treated with vehicle or PKF118-310. Each replicate is a distinct biological sample. Cyclin D1 and MYC protein levels were quantified by densitometry analysis of Western blot bands. Values are normalized to actin when not specified, and are shown as an average (± SEM) of at least three animals per time-point and genotype. Two-way ANOVA was applied (factors: cell genotype, treatment) (*p<0.05, **p<0.01, ***p<0.001). (C) Cyclin D1 staining performed on drug- and vehicle- treated Cdh16^Cre::Tfeb^fs mice before and after the hematoxylin counterstaining. DOI: http://dx.doi.org/10.7554/eLife.17047.020

Figure 6—figure supplement 3.. Inhibition of autophagy in Tfeb overexpressing mice (Atg7^flox/flox::Cdh16^Cre::Tfeb^fs) does not affect the cystic phenotype.

(A) Real-time PCR validation of well-known TFEB direct gene targets whose function is related to the lysosomal and autophagic pathways performed on P30 Cdh16^Cre::Tfeb^fs mice. Values are shown as the average (± SEM) of at least three Cdh16^Cre::Tfeb^fs mice and are normalized to wild-type mice (*p<0.05, **p<0.01, ***p<0.001, two-sided Student’s t test). (B) P14 and P30 kidney lysates from Cdh16^Cre::Tfeb^fs animals were evaluated by LC3I/II immunoblot. LC3 active / LC3 total ratios were quantified via densitometry of the Western blot bands (graph). Each replicate is a different biological sample. Values are shown as the average (± SEM) of at least three Cdh16^Cre::Tfeb^fs mice and are normalized to wild-type lines. (*p<0.05, **p<0.01, ***p<0.001, two-sided, Student’s t test). (C) Renal images of (Atg7^flox/flox::Cdh16^Cre::Tfeb^fs) double transgenic mice and controls at P30. (D) Kidney-to-body weight ratio (KW/BW) from the different genotypes obtained. Values are normalized to the Atg7^flox/flox::Cdh16^Cre line and are shown as the average (± SEM) of at least three mice per genotype. One-way Anova was applied (factors: genotype) (*p<0.05, **p<0.01, ***p<0.001). (E) PAS and ATG7 staining of kidneys from the different mouse lines. DOI: http://dx.doi.org/10.7554/eLife.17047.021

Author response image 1.

DOI: http://dx.doi.org/10.7554/eLife.17047.024