ABSTRACT
Nephronophthisis (NPH) is an orphan recessive kidney disease mostly caused by mutations in NPHP1 and 20 other genes encoding proteins that localize to primary cilia. To date the pathways linking altered primary cilia function to progressive kidney scarring in NPH remain poorly defined and therapeutic options allowing NPH patients to escape end-stage kidney disease are lacking.
Distinct proteins mutated in NPH interact with components of the Hippo pathway, an important regulator of cell fate. YAP (Yes-associated protein) overactivation has been shown to induce renal scarring while YAP inhibition showed protective effect in kidney diseases unrelated to NPH. Yet, the therapeutic potential of YAP inhibition in NPH has not been formerly assessed.
Here we studied the impact of both genetic and pharmacologic YAP inhibition on the NPH-like phenotype caused by a bi-allelic mutation of Lkb1, a ciliary kinase interacting with NPHP1. Contrary to non NPH renal disease, our results reveal an unexpected protective role of YAP in Lkb1 mutant kidneys. Indeed, YAP genetic disruption drastically increase kidney disease burden in Lkb1 deficient mice, while pharmacologic inhibition of YAP failed to improve their phenotype.
Collectively these results suggest that YAP inhibition is not a valid therapeutic strategy in NPH and suggest that LKB1 and YAP are parallel negative regulators of a yet uncharacterized pathway detrimental for kidney health.
INTRODUCTION
Nephronophthisis (NPH) is an orphan genetic disease affecting the kidney. This recessive affection usually manifests with polyuria followed by a gradual reduction in kidney function related to progressive renal scarring. To date, no treatment is available for this affection, which is nonetheless the leading genetic cause of end-stage kidney disease in children1,2.
NPH is mostly caused by mutations affecting proteins that localize to primary cilia, solitary antenna-like organelles that protrude from the apical surface of most mammalian epithelial cells. Primary cilia emerged from the extension of tubulin doublets originating from the triplets forming the core of the mother centriole. Protein cargoes enter to and exit from the cilia through the transition zone, a complex protein sorting process taking place at the base of the cilium3. The most frequently mutated genes in NPH encode proteins that localize to the transition zone. Indeed, NPHP1, which mutations account for 25% of NPH cases, as well as NPHP4 and RPGRPI1L/NPHP8 are all core proteins of the transition zone4. Loss of function of NPHP genes does not impede ciliogenesis but perturbs cilia organization and/or signalling. Unfortunately, the dysregulated pathways responsible for kidney degeneration in NPH have not yet been solved, precluding the development of efficient therapies for the children and young adults affected by the disease.
One of the main difficulty faced studying the pathophysiologic events driving kidney damage in NPH is the overall lack of orthologous rodent models recapitulating all the disease features. Indeed, Nphp1 inactivation does not lead to significant kidney fibrosis5,6 in mice and the same is true for the majority of NPH genes. One notable exception is Glis2, which inactivation consistently leads to kidney fibrosis in mice7,8, but accounts for only 0.1% of NPH cases in human.
Liver Kinase B1 (LKB1; encoded by STK11) is a ciliary kinase involved in the control of polarity and metabolism9. LKB1 interacts with NPHP1 and the bi-allelic disruption of Stk11 in renal tubules of mice recapitulates the NPH phenotype10. Combining this model with the analysis of Glis2 mutant mice and material derived from NPH patients, we further demonstrated that NPH is associated with a specific inflammatory signature that likely participates to kidney damage11. While NF-ΚB signalling has been identified as a contributor to kidney damage in Glis2 deficient mice12, the molecular machinery driving kidney inflammation in NPH remains poorly defined.
The Hippo signalling pathway is an evolutionarily conserved kinase cascade that plays a fundamental role in several biologic processes such as embryonic development, organ size control, cell proliferation and apoptosis13. The main function of Hippo kinase is to phosphorylate the transcription co-activator YAP (Yes-associated protein) or its paralog TAZ/WWTR1 (Transcriptional coactivator with PDZ-binding domain). Unphosphorylated YAP and TAZ bind to transcriptional enhanced associate domain (TEAD1-4) transcription factors to regulate the expression of multiple genes in a cell and context specific fashion. Upon their phosphorylation by Hippo kinase, YAP and TAZ are targeted to degradation and/or sequestrated in the cytoplasm, shutting down the transcription of their target genes. In mammals, Hippo signalling consists in four serine/threonine kinases: the two upstream mammalian STE20-like protein kinases 1 and 2 (MST1/2; encoded by STK4 and STK3, respectively) phosphorylate the effector large tumor suppressor kinases 1 and 2 (LATS1/2), which in turn phosphorylate YAP and TAZ causing their exclusion from the nuclear compartment14,15.
While Hippo pathway plays fundamental roles in kidney development16,17, several lines of evidence suggest that its deregulation drives kidney damage. Indeed, tubule specific inactivation of Mst1 and 2 promotes renal inflammation and scarring largely through the induction of YAP target genes18. YAP and TAZ activation has been observed in autosomal dominant polycystic kidney disease (ADPKD)19 which is a genetic renal cystic disease caused by mutations affecting the ciliary proteins polycystins 1 and 2. In this context, the disruption of YAP and TAZ has been shown to reduce cystic disease burden in mice20,21. The idea that YAP/TAZ exerts pro-fibrotic function in the kidney is further supported by studies showing that verteporfin, a compound dissociating YAP/TEAD complex, reduces kidney fibrosis in genetic or acquired model of kidney diseases22,23.
Experimental data has linked NPH proteins to Hippo/YAP pathway. In vitro, NPHP4 binds and inhibits MST124, while NPHP9/NEK8 has been shown to facilitate TAZ nuclear translocation25. On the other hand, NPHP9/NEK8 and NPHP16/ANKS6 pathogenic variants were reported to increase YAP transcriptional output in severe syndromic cystic dysplasia patients derived fibroblasts26 and late onset chronic kidney disease27. Verteporfin treatment rescues the developmental phenotype caused by nek8 inhibition in zebrafish embryo26. In the same line, an unbiased siRNA screen identified LKB1 as a suppressor of YAP transcriptional activity28, while Yap inactivation has been shown to prevent liver growth caused by Lkb1 disruption28.
Taken together these data suggest that NPH and Hippo/YAP signalling intersects, but the signification of this interaction for NPH kidney disease has not formerly been assessed. The growing list of drugs allowing the manipulation of Hippo and/or YAP activation prompted us to investigate this question.
RESULTS
Depletion of LKB1 or NPHP1 impairs Hippo signalling
To determine how LKB1 and NPHP1 intersect with Hippo pathway, we used cultured renal epithelial cells (MDCK) in which LKB1 or NPHP1 were depleted by shRNA-mediated knockdown10. Western blot experiments revealed that while nuclear YAP accumulated in Lkb1 and Nphp1 knockdown cells as compared to control cells (Figure 1A), cytoplasmic and nuclear TAZ levels were not modified. In parallel, the expression of Ankrd1 (Ankyrin repeat domain 1), a prototypic YAP/TEAD transcriptional target, increased upon Lkb1 or Nphp1 shRNA-mediated knockdown (Figure 1B). To obtain further mechanistic insights, we examined the interaction between LKB1, NPHP1 and the upstream regulators of YAP nuclear localization. As previously described by others28,29, both LKB1 and NPHP1 co-immunoprecipitated with MST1 and LATS1 kinases in vitro (Figure 1C). Collectively, these data suggest that both LKB1 and NPHP1 interact with Hippo kinases to repress the nuclear translocation of YAP, thereby reducing the expression of YAP target genes.
Loss of Lkb1 in the mouse kidney results in Hippo pathway dysregulation
In order to verify whether LKB1 could modulate Hippo pathway also in vivo, we took advantage of a mouse model characterized by Lkb1 depletion specifically in the distal tubules (Lkb1ΔTub) causing a NPH-like phenotype10. Quantitative RT-PCR analysis revealed increased levels of known YAP target genes including Cyr61 (cysteine-rich angiogenic inducer 61), Ctgf (connective tissue growth factor), Ankrd1, Edn1 (endothelin 1) and Areg (amphiregulin) mRNA in the Lkb1ΔTub mice compared to control mice as soon as 5 weeks after birth (Figure 2A). Moreover, the expression of these genes increased over time as higher levels were found in 23-week old Lkb1ΔTub mice as compared to controls (Figure 2A). Immunohistochemistry revealed that YAP is strongly expressed in nuclei of Lkb1ΔTub mice, especially in epithelial cells belonging to dilated tubules similarly to what was found in ADPKD19 (Figure 2B). Together, our in vivo and in vitro data revealed that LKB1 and NPHP1 promoted Hippo pathway through inhibition of YAP/TEAD transcriptional activity.
Loss of tubular Yap causes premature death, urine concentration defect and rapid renal function decline in Lkb1ΔTub mice
To determine if the deletion of Yap together with Lkb1 would prevent the NPH-like renal disease, we produced an allelic series by crossing Lkb1ΔTub mice with mice bearing Yap floxed alleles and compared them with littermate controls. At 4 weeks of age, kidneys of Lkb1ΔTub mice displayed reduced LKB1 expression measured by RT-PCR and immunohistochemistry (Supplementary Figure 1A-B). Similarly, Yap mRNA and protein expression were decreased in YapΔTub kidneys and associated with enhanced Cyr61 mRNA expression without affecting Taz mRNA level (Supplementary Figure 1C-F). Macroscopic inspection of the kidneys from control and mutated mice revealed that YapΔTub kidneys resemble the control ones, while Lkb1ΔTub kidneys showed a discrete surface granular irregularities, as previously described10. Surprisingly, kidneys carrying a bi-allelic inactivation of Yap and Lkb1 showed drastic parenchymal thinning with medulla atrophy resulting in a balloon-like aspect (Figure 3A). Importantly, renal pelvis from double mutant animals were not distended excluding ureteral obstruction. As a consequence Lkb1ΔTub; YapΔTub mice displayed smaller kidneys compared to controls (Figure 3B and Supplementary Figure 2A-B). Contrary to other animals, Lkb1ΔTub; YapΔTub mice did not survive for more that 8 weeks (Figure 3C). As soon as 4 weeks of age, all the animals carrying a bi-allelic inactivation of either Lkb1 or Yap displayed a dramatic urine concentration defect as compared to littermate controls (Figure 3 D). Remarkably, only Lkb1ΔTub mice with allelic dosage of Yap inactivation showed a reduced kidney function proportional to Yap inactivation (Figure 3E). By contrast, YapΔTub mice showed a normal kidney function. Collectively, these results demonstrate that the combination of Lkb1 and Yap inactivation in distal tubules worsen the Lkb1ΔTub phenotype.
Genetic inhibition of tubular Yap promotes tubular lesions and fibrosis in Lkb1 mutant mice
Thickened tubular basement membranes, tubular dilatation and atrophy, interstitial cell infiltration and renal fibrosis are the main features that characterize Lkb1ΔTub kidneys. Renal histology revealed that all these features were exacerbated in Lkb1ΔTub in proportion to the reduction of Yap gene dosage (Figure 4A-B and Supplementary Figure 2). Consistently, Yap deletion in Lkb1ΔTub kidneys proportionally increased the mRNA expression of two tubular injury markers, Lcn2 and Kim1 (Figure 4C-D). Sirius red staining confirmed the increase in renal fibrosis in Lkb1ΔTub kidneys, which was more pronounced in Lkb1ΔTub; YapΔTub kidneys. Markers of interstitial fibrosis, such as collagens or pro-fibrotic cytokine levels (Tgfb1, Pdgfb) were increased proportionally to Yap gene dosage reduction (Figure 4E-G and Supplementary Figure 2). By contrast, Yap inactivation in wild-type Lkb1 animals (YapΔTub) did not lead to kidney damage at least in this time frame. Overall, our data indicate that the mid-term maintenance of Lkb1 deficient kidney structure and function is critically dependent of YAP.
Tubular Yap inactivation promotes renal inflammation in Lkb1ΔTub mice
We previously showed that NPH is characterized by a complex and specific cytokine signature that is associated with an early kidney infiltration by macrophages, neutrophils and T cells11. Thus, we evaluated the inflammatory status of Lkb1ΔTub mice when Yap was inactivated. Quantitative RT-PCR analysis revealed a global upregulation of the pro-inflammatory cytokines identified in NPH that directly correlated with the level of Yap deletion (Figure 5A and Supplementary Figure 3). As these secreted factors are involved in the recruitment of different immune cell populations, we then performed immunohistochemistry staining for macrophages (F4/80), T cells (CD3) and neutrophils (Ly6B.2). Supporting the molecular results, Lkb1ΔTub; YapΔTub kidneys presented increased kidney infiltration by macrophages, T cells and neutrophils (Figure 5B-F and Supplementary Figure 4). Of note, Yap inactivation did not result in renal inflammation in animal carrying at least one functional copy of Lkb1 (YapΔTub and Lkb1fl/+; YapΔTub). These results confirmed that the specific inflammatory signature previously identified is associated with the development of the disease and indicated that tubular YAP limits kidney inflammation in NPH.
Pharmacological inhibition of YAP enhances renal inflammation in Lkb1 mutant mice
YAP can be targeted pharmacologically by verteporfin, which inhibits the formation of the transcriptomic complex YAP-TEAD. We treated control and Lkb1ΔTub mice with verteporfin (i.p. 100mg/kg) from 8 to 12 weeks of age. While verteporfin treatment had no impact on urine concentration defect or kidney function in Lkb1ΔTub mice, we observed some lethality and higher mRNA expression of Lcn2 tubular injury marker in verteporfin-treated Lkb1ΔTub animals (Figure 6A-B and Supplementary 5A-G). Verteporfin treatment did not reduce interstitial fibrosis nor fibrotic genes expression in Lkb1ΔTub mice (Supplementary Figure 5H-J). However, verteporfin treatment in Lkb1ΔTub animals tends to increase the expression of some inflammatory cytokines and exarcerbate infiltration of the kidney by immune cells (Figure 6D-Q). Overall, our genetic and pharmacologic inhibition of YAP demonstrated that reducing YAP gene dosage or YAP/TEAD transcriptional activity is not a viable strategy to modulate NPH progression.
DISCUSSION
The aim of this study was to determine if YAP inhibition could represent a valid therapeutic target in NPH using Lkb1ΔTub mouse model. Our results clearly demonstrate that, in this model, YAP activation does not contribute to kidney lesions. On the opposite, we observed that tubule specific Yap inactivation drastically worsened kidney damage associated with Lkb1 loss of function. Similarly, treating mice with the systemic YAP/TEAD inhibitor verteporfin did not slow the progression of NPH-like renal disease and even tends to worsen its inflammatory aspect.
As we used a non-orthologous NPH model of the disease that recapitulates the fibrotic feature observed in human bearing bi-allelic NPHP1 deletion, we cannot rule out that YAP inhibition may exert some positive effects on the less frequent forms of NPH characterized by a cystic phenotype reminiscent of ADPKD. In this regard, it would be interesting to perform similar studies in jck mice, which bear a spontaneous mutation in Nphp9/Nek8 that cause cystogenesis associated with YAP activation26.
Interestingly, while YAP genetic disruption drastically worsened the kidney phenotype of Lkb1ΔTub mice, verteporfin treatment was mostly neutral. This discrepancy may reflect an incomplete inhibition of YAP transcriptional activity by verteporfin. Alternatively, this observation could also pinpoint a TEAD independent functions of YAP in Lkb1ΔTub kidneys. Besides its function as a co-activator of TEAD, YAP has been shown to bind other transcription factors such as RUNX2/3, p73, SMADs, ERBB30. In addition, YAP also modulates important signaling pathways independently of transcription factors. Such regulation has been notably demonstrated for important pro-inflammatory pathways such as NF-κB, which has also been implicated in the kidney disease caused by Glis2 inactivation12.
While our results do not support YAP inhibition as a promising therapy for NPH, they unveil an unforeseen protective role of YAP in this disease. Our results suggest a model in which LKB1 and YAP are parallel negative regulators of a molecular pathway responsible for the development of kidney lesions in NPH. Further experiments are needed to validate this model and may lead to the identification of molecular events responsible for NPH.
MATERIALS and METHODS
Mice
Mice were housed in a specific pathogen-free facility, fed ad libitum and housed at constant ambient temperature in a 12-hour day/night cycle. Breeding and genotyping were done according to standard procedures.
Lkb1ΔTub mice were previously described10. Yapflox/flox mice (C57BL/6J) were kindly provided by Prof. Eric Olson (University of Texas South western Medical Center, Dallas, USA)31 and were backcrossed for 2 generations with Lkb1ΔTub mice. The progeny was then intercrossed to generate mice with allelic dosage of Lkb1 into tubule-specific Yap knockout: Yap knockout with wild-type Lkb1 (further referred to as Lkb1+/+; YapΔTub) or heterozygous floxed Lkb1 allele (further referred to as Lkb1fl/+; YapΔTub); and mice with allelic dosage of Yap into tubule-specific Lkb1 knockout: Lkb1 knockout with wild-type Yap (further referred to as Lkb1ΔTub; Yap+/+), heterozygous Yap (further referred to as Lkb1ΔTub; Yapfl/+) or homozygous floxed Yap alleles (further referred to as Lkb1ΔTub; YapΔTub). Littermates lacking KspCre transgene were used as controls. Experiments were conducted on both females and males.
For Verteporfin experiment, only control and Lkb1ΔTub male mice were used. Verteporfin (Sigma, SML0534) was dissolved in DMSO (100 mg/mL), aliquoted, and stored at −80°C. Working solution was prepared at 10 mg/mL in PBS freshly before use. Eight-week old mice were administered i.p. with vehicle or verteporfin solution at a dose of 100 mg/kg every other day for 1 month. Animals were sacrificed 4 hours after the last injection.
Urine and Plasma Analyses
Spot urine samples were obtained from mice the day of sacrifice. Urine osmolality was measured with a freezing point depression osmometer (Micro-Osmometer from Knauer). For Verteporfin protocol, 8-hour urine samples were obtained from mice housed in individual boxes without access to water and food. Urine excretion was measured.
Retro-orbital blood was collected from anesthetized mice. Plasma blood urea nitrogen (BUN) was measured using a Konelab 20i Analyzer (Thermo Scientific).
Morphological Analysis
Mouse kidneys were fixed in 4% paraformaldehyde, embedded in paraffin, and 4µm sections were stained with periodic acid-Schiff (PAS) or Picrosirius Red. Stained full size images were recorded using a whole slide scanner Nanozoomer 2.0 (Hamamatsu) equipped with a 20x/0.75 NA objective coupled to NDPview software (Hamamatsu). The severity of renal pathology was graded in a blinded fashion by two independent observers assessing the overall lesions on the whole kidney section stained with PAS. A 7-point scale was defined, in which 0 indicated normal kidney architecture, 3 indicated tubular atrophy, tubular basement thickening and cell infiltration and 6 characterized by loss of medulla and complete disorganization of the remaining renal parenchyma. For verteporfin experiment, only four scores were defined ranging from score 0 normal kidney architecture to score 3 associating tubular atrophy, tubular basement thickening and cell infiltration.
Cell culture
Madin-Darby canine kidney cells (MDCK, kind gift from Prof. Kai Simons, MPI-CBG, Dresden, Germany) were cultured using Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin (Sigma). Generation of MDCK cell lines for tetracycline-inducible knockdown of target genes was previously described10. The shRNA targeting sequences used were as follows: Lkb1-i (5’-GCTGGTGGACGTGTTATAC-3’), Luciferase (5’-CGTACGCGGAATACTTCGA-3’) and Nphp1-i (5’-GGTTCTCAGTAGACATGTA-3’). For nucleo-cytoplasmic experiment, 150,000 cells/cm2 were seeded with or without tetracycline (5 µg/ml) for 6 days on 6 well plate (Greiner, 657 160). Cells were lysed in Pierce NE-PER™ buffer (Thermo Fisher Scientific) to extract both nuclear and cytoplasmic proteins. In parallel, 150,000 cells/cm2 were seeded with or without tetracycline (5 µg/mL) for 6 days on polycarbonate membranes (COSTAR, 3401). Total RNAs were extracted after 5 hours of serum deprivation and qRT–PCR was performed.
Human embryonic kidney (HEK 293T) cells (ATCC Promocell; CRL-11268) were cultured using Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (Biochrom). Plasmids were transiently transfected using calcium. Tagged constructs for CD2AP, NPHP1, and LKB1 were previously described4. Cells were lysed in IP buffer (20 mM Tris-HCl pH 7.5, 50mM NaCl, 1% Triton X100, 15mM Na4P2O7 and 0.1mM EDTA) supplemented with 2 mM Na3VO4, 50 mM NaF, 5 mM b-glycerophosphate, and protease inhibitor cocktail tablet (Roche) using a 20-G needle. Equal amounts of proteins were incubated with anti-FLAG M2 affinity gel (Sigma Aldrich, A2220) and processed by Western blotting.
All cells were regularly tested for mycoplasma contamination and were mycoplasma-free.
Quantitative RT-PCR
Total RNAs were obtained from cells or mouse kidneys using RNeasy Mini Kit (Qiagen) and reverse transcribed using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) according to the manufacturer’s protocol. Quantitative PCR were performed with iTaq™ Universal SYBR® Green Supermix (Bio-Rad) on a CFX384 C1000 Touch (Bio-Rad). Hprt, Ppia and Rpl13 were used as normalization controls32. Each biological replicate was measured in technical duplicates. The primers used for qRT-PCR are listed in Supplementary Table 1.
Heatmap was generated using R v4.1.0 and the R package pheatmap v1.0.12 (note: heatmaps displays Z-scores computed on the expression levels of the identified cytokines, measured by qPCR).
Western blot
Cells were lysed in IP buffer or Pierce NE-PER™ buffer (Thermo Fisher Scientific) using a 20-G needle. Protein content was determined with Pierce BCA protein assay kit (Thermo Fisher Scientific). Equal amounts of proteins were resolved on 4–15% Mini-PROTEAN® TGX™ gel (Bio-Rad) under reducing conditions, transferred, incubated with primary and secondary antibodies, and visualized on film according to standard protocols. Band density was calculated and normalized using LabImage 1D L340 software. Antibodies used are listed in Supplementary Table 2.
Immunohistochemistry
Four-micrometer sections of paraffin-embedded kidneys were submitted to antigen retrieval and avidin/biotin blocking (Vector, SP-2001). Sections were incubated with primary antibody followed by biotinylated antibody, HRP-labeled streptavidin (Southern Biotech, 7100-05, 1:500), and 3,30-diaminobenzidine-tetrahydrochloride (DAB) revelation. Antibodies used are listed in Supplementary Table 2. Full size images were recorded using a whole slide scanner Nanozoomer 2.0 (Hamamatsu) coupled to NDPview software. For macrophage quantification, stained area was measured with ImageJ software from full sized kidney images and visualized as the ratio of stained DAB surface to total kidney section area. For neutrophil quantification, we counted manually the number of foci in the whole kidney section. Foci were defined as 4 or more neutrophils surrounding a tubule. For T cell quantification, 10 randomly selected fields of view were measured with ImageJ software. For all quantification, glomerular and intra-tubular positive stainings were removed from the analysis.
Statistical analysis
Data were expressed as means. Differences between groups were evaluated using Mann-Whitney test when only two groups were compared. When testing more comparisons, Kruskal-Wallis test was used. The statistical analysis was performed using GraphPad Prism V8 software. All image analyses and mouse phenotypic analyses were performed in a blinded fashion.
Study approval
All animal experiments were conducted according to the guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals, as well as the German and French laws for animal welfare, and were approved by regional authorities (Regierungspräsidium Freiburg G-13/18 and Ministère de l’Enseignement, de la Recherche et de l’Innovation APAFIS#2020051216078531).
CONFLICT OF INTEREST STATEMENT
The authors declare that they have no conflict of interest.
ACKNOWLEDGEMENTS
We are grateful to Prof. Eric Olson for sharing Yapfl/fl mice, and to Dr Frank Bienaimé for critically reading the manuscript. We thank the members of the mouse histology and breeding facilities (S.F.R Necker INSERM US24, Paris, France), and of the mouse renal physiology facility (CRC, Paris, France) for technical assistance.
Marceau Quatredeniers and Amandine Viau were supported by a public grant “RHU-C’IL-LICO” overseen by the Agence Nationale de la Recherche (grant number: ANR-17-RHUS-0002), E. Wolfgang Kuehn was supported by Deutsche Forschungsgemeinschaft KU1504/7-1 and KU1504/8-1, Sophie Saunier was supported by the Institut National de la Santé et de la Recherche Médicale (INSERM), the Ministère de l’Education Nationale, de la Recherche et de la Technologie (MRT), by a State funding from the Agence Nationale de la Recherche (grant references: ANR-10-IAHU-01, ANR-17-RHUS-0002).