ABSTRACT
A gradual decline in renal function occurs even in healthy aging individuals. In addition to aging per se, concurrent metabolic syndrome and hypertension, which are common in the aging population, can induce mitochondrial dysfunction, endoplasmic reticulum stress, oxidative stress, inflammation, altered lipid metabolism, and profibrotic growth factors in the kidney, which collectively contribute to age-related kidney disease. With increasing population of older individuals and the increasing incidence of acute kidney injury and chronic kidney disease, identifying preventable or treatable aspects of age-related nephropathy becomes of critical importance. In this regard we studied the role of the nuclear hormone receptors, the estrogen related receptors (ERRs), whose expression levels are decreased in aging human and mouse kidneys. Our studies have identified the estrogen related receptors ERRα, ERRβ, and ERRγ as important modulators of age-related mitochondrial dysfunction, cellular senescence, and inflammation. Significantly these pathways are also regulated by lifelong caloric restriction (CR), which is known to prevent several age-related complications including kidney disease. ERRα, ERRβ, and ERRγ expression levels are decreased in the aging kidney, and CR and pharmacological treatment with a pan ERR agonist results in increases in expression of ERRα, ERRβ, and ERRγ in the kidney. Remarkably, only a 4-week treatment of 21-month-old mice with the pan ERR agonist reversed the age-related mitochondrial dysfunction, the cellular senescence marker p21, and inflammatory cytokines, including the STAT3 and STING signaling pathways.
INTRODUCTION
The fastest growing group of people in the US with impaired kidney function is the 65 and older age group. This population is expected to double in the next 20 years, while the number worldwide is expected to triple from 743 million in 2009 to 2 billion in 2050. This will result in a marked increase in the elderly population with chronic kidney disease and acute kidney injury. This increase may be further amplified by other age-related co-morbidities including metabolic syndrome and hypertension that accelerate age-related decline in renal function 1–3. Thus, there is increasing need for prevention and treatment strategies for age-related kidney disease.
A gradual decline in renal function occurs even in healthy aging individuals 4–6. In addition to aging per se, metabolic syndrome and hypertension can induce mitochondrial dysfunction, endoplasmic reticulum stress, oxidative stress, inflammation, altered lipid metabolism, and profibrotic growth factors in the kidney, which collectively contribute to age-related kidney disease 4.
There is variation in the rate of decline in renal function as a function of gender, race, and burden of co-morbidities 7–10. Greater glomerular, vascular and tubulointerstitial sclerosis is evident on renal tissue examination of healthy kidney donors with increasing age 11–13. Interestingly, examination of processes leading to sclerosis suggests a role for possible modifiable systemic metabolic and hormonal factors that can ameliorate the rate of sclerosis. With the population of older individuals increasing, identifying preventable or treatable aspects of age-related nephropathy becomes of critical importance.
There is increasing evidence that mitochondrial biogenesis, mitochondrial function, mitochondrial unfolded protein response (UPRmt), mitochondrial dynamics and mitophagy are impaired in aging, and these alterations result in several of the age-related diseases 14–21. In this regard current research efforts are concentrated on modulating these molecular mechanisms to improve mitochondrial function.
Caloric restriction (CR) plays a prominent role in preventing age-related complications. We have previously shown that CR prevents age-related decline in renal function and renal lipid accumulation via inhibition of the sterol regulatory element binding proteins (SREBPs) 22, 23. CR is also an important modulator of mitochondrial function. We have demonstrated that CR prevents age-related mitochondrial dysfunction in the kidney by increasing mitochondrial/nuclear DNA ratio, the mitochondrial transcription factor Nrf-1, AMPK, SIRT1, SIRT3, mitochondrial complex activities, mitochondrial IDH2, and mitochondrial fatty acid β-oxidation 24. In addition, CR also prevents age-related decrease in mitochondrial abundance in the renal tubules. CR increases the renal expression of FXR and TGR5. Treatment of 22-month old mice fed ad lib (AL) for 2 months with the dual FXR-TGR5 agonist INT-767 reversed most of the age-related impairments in mitochondrial function and the progression of renal disease 24. Importantly the FXR-TGR5 dual agonist as well as CR also increased expression of PGC-1α, ERRα and ERRγ, which are important regulators of mitochondrial biogenesis and function.
The estrogen-related receptors (ERR) ERRα (NR3B1, ESSRA gene), ERRβ (NR3B2, ESRRB gene), and ERRγ (NR3B3, ESRRG gene) are closely-related members of the nuclear receptor family. Except for one report 25, there are no known endogenous ligands for these orphan receptors. Importantly they do not bind natural estrogens, and they do not directly participate in classic estrogen signaling pathways or biological processes 26–28.
ERRα and ERRγ are strongly activated by their coactivators PGC-1α and PGC-1β 29, 30. In contrast, RIP140 and NCoR1 are important corepressors of ERR activity 31, 32. ERRs are also subject to post-translational modifications including phosphorylation, sumoylation, and acetylation that modulate the receptors’ DNA binding and recruitment of coactivators 33–35.
ERRα and ERRγ regulate the transcription of genes involved in mitochondrial biogenesis, oxidative phosphorylation, tricarboxylic acid (TCA) cycle, fatty acid oxidation and glucose metabolism 28, 36–46. However, in addition to overlapping gene activation there is also ample evidence that ERRα and ERRγ also have differential and opposing effects which can be due to interactions with corepressors, coactivators, posttranslational modification, or differential cell expression 28, 46. Opposing effects for ERRα and ERRγ is seen in breast cancer 47, 48, regulation of gluconeogenesis in the liver 46, 47, skeletal muscle function 46, 47, macrophage function 46, 49, 50, and regulation of LDHA related to anaerobic glycolysis 28. Finally, ERRα and ERRγ are highly expressed in the mouse and human kidney 51–53.
The roles of ERRα and ERRγ in modulating age-related impairment in mitochondrial function and age-related inflammation (inflammaging) are not known. We undertook our current studies to determine if a pan agonist of ERRs, including ERRα, ERRβ, and ERRγ, will improve mitochondrial dysfunction and inflammation in the aging kidney.
RESULTS
ERRα and ERRγ expression is decreased in the aging human kidney
In our previous study, we found that there is a decrease in the expression of ERRα and ERRγ in the aging kidney, and their expression is upregulated by the dual FXR-TGR5 agonist INT-767 treatment or caloric restriction, which correlated with increased mitochondrial biogenesis and function in the treated aging kidneys 24. In view of the role of ERRs in mitochondrial biogenesis, we determined if this observed decreased expression also occurs in the human aging kidney. We performed immunohistochemistry with human kidney sections from young versus old individuals.
The results indicate that both ERRα and ERRγ are expressed in renal tubules and their expression levels are markedly decreased in the aging human kidney samples (Figure 1A). Since ERRs are important modulators of mitochondrial biogenesis, we also stained the human kidney sections for the mitochondrial pyruvate dehydrogenase (PDH) e2/e3 and we found a marked decrease in PDH immunostaining in the aging human kidney samples (Figure 1B).
ERRα and ERRγ RNA distribution in the mouse kidney
To determine where ERRα and ERRγ are expressed in the kidney, we performed single nuclei RNAseq54–57. With 100k read depth and 3000-5000 nuclei sequenced, we were able to identify 12 clusters and assigned them to major cell types known in the mouse kidney (Figure 2A). We found most ERRα is expressed in proximal tubules, intercalated cells and podocytes. For ERRγ, we found proximal tubules and intercalated cells express most of it. Compared to the young kidneys, aging proximal tubules at S1/S2 show decline in ERRα and ERRγ expression (Figure 2B).
Pan ERR agonist treatment increases the mRNA expression of Errα, Errβ and Errγ in the aging kidney
We found that Errα, Errβ and Errγ mRNA abundance is significantly decreased in the kidneys of aging mice. It should be noted however that Errβ expression level is at least 5-fold lower than either Errα or Errγ mRNA abundance. Treatment with the ERR pan agonist induces significant increases in the Errα, Errβ and Errγ mRNA abundance in the kidneys of aging mice, to levels observed in the young mice (Figure 3).
This prompted us to perform bulk RNAseq to determine which pathways are changed by the increased ERR activity. The data showed the main pathways upregulated by the treatment are mitochondrial related and the downregulated pathways are immune related (Figure 4A).
Proteomic analysis revealed that major pathways regulated included mitochondrial electron transport chain (ETC), tricarboxylic acid (TCA) cycle, and mitochondrial fatty acid β-oxidation (Figure 4B).
The multi-omics integration using O2PLS demonstrated a regulation pattern that separated old group from old with pan agonist group. They found pan agonist treatment decreased pathways related to immune system activation in both gene expression and protein abundances (Figure 4C).
As the hallmarks for aging kidneys are decreased mitochondrial function and increased inflammation 58, 59, we then determined if treatment with the pan ERR agonist improved those defects in the aging kidneys.
Pan ERR agonist treatment restored mitochondrial function in aging kidneys
The canonical action of ERRs is to induce mitochondrial biogenesis. We have seen the increased expression in aging kidneys of master mitochondrial biogenesis regulator PGC1α/β and Tfam1 with pan ERR agonist treatment. As a result, mitochondrial DNA/nuclear DNA ratio is increased in the aging kidneys following treatment (Figure 5), an indicator of increased mitochondrial biogenesis.
ERR activation also increased the expression of genes related to mitochondrial electron transport chain (ETC) and tricarboxylic acid (TCA) cycle, such as complex I subunit Ndufb8, complex II subunit Sdhc, complex III subunit Uqcrb, complex IV subunit Cox6a2, complex V subunit Atp5b (Figure 6A), and Pdhb, Mdh1, Idh3b, and Sucla2, important enzymes for TCA cycle (Figure 6B). This is consistent with the increased succinic acid level, one of TCA intermediate metabolites, following the treatment (Figure 6C). The interrelationship between ETC and TCA cycle is illustrated in Figure 6D.
Mitochondrial ETC is also regulated by ERR agonism. Native blue gel showed the increased level of assembled complex II, III, IV and V after the treatment (Figure 7A). This results in the increased maximum respiration capacity in mitochondria isolated from aging kidneys (Figure 7B).
In addition, enzymes that mediate mitochondrial fatty acid β-oxidation including CPT-1a and MCAD mRNA is upregulated by the pan ERR agonist (Figure 8), suggesting the involvement of ERR agonism in promoting fatty acid oxidation.
Pan ERR agonist treatment altered mitochondrial dynamics in aging kidneys
Transmission electron microscopy showed alterations in the mitochondria of aging kidneys including decreases in the width and length of mitochondria, which were restored to levels seen in young kidneys upon treatment with the pan ERR agonist (Figure 9A).
Since these mitochondrial changes are reminiscent of alterations in mitochondrial fusion and fission, we also determined expression of proteins that regulate mitochondrial fusion and fission.
Opa1 protein localizes to the inner mitochondrial membrane and helps regulate mitochondrial stability, energy output, and mitochondrial fusion 60, 61. While mRNA level was decreased, there was no significant change in the protein level in the aging kidney. However, upon treatment there was an increase in the mRNA level and a tendency for the protein level to increase in the aging kidneys (Figure 9B). Mitofusin 2 is found in the outer membrane that surrounds mitochondria and participates in mitochondrial fusion 62. There was a significant decrease in the aging kidney and the ERR pan agonist increased the protein abundance in both the young and the old kidneys, with the resulting levels in the old kidneys being the same as in the young kidneys (Figure 9B). In addition, there were also significant decreases in Mitoguardin 2 and MitoPLD mRNA levels in the aging kidneys that were normalized upon treatment with the pan ERR agonist (Figure 9B).
Drp1, is a member of the dynamin superfamily of proteins and is a fundamental component of mitochondrial fission 62–64. We found that there were significant increases in Drp1 and phospho-Drp1 protein in the kidneys of aging mice, which were restored back to levels seen in young mice following treatment with the pan ERR agonist (Figure 9C).
Pan ERR agonist treatment decreased inflammation in aging kidneys
The cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) pathway has been reported to lead mitochondrial dysfunction to the activation of senescence and innate immune system 65, 66. In the aging kidneys we found increased expression of STING mRNA and pan agonist can significantly lower its expression (Figure 10A).
These changes are in parallel to the expression of senescence marker p21 which has been induced in the aging kidney and reduced by the treatment. Another marker p16 was found increased in the aging kidneys but no remarkable change after the treatment (Figure 10A). Both markers have been found downregulated in the aging kidneys with life-long CR (Figure 10B). To find out which inflammatory factors are regulated by the pan agonist treatment we searched the RNAseq data and verified by the real-time PCR that the proinflammatory cytokine IL-1β and cell adhesion molecule ICAM1 were targeted by ERR agonist (Figure 10C). In addition, we observed the correspondent changes occurring to STAT3, a cytokine activated signaling pathway (Figure 10D). As a marker for STAT3 activation, we found a significant increase in p-Tyr705-STAT3 protein and total STAT3 protein that was decreased upon treatment with the pan ERR agonist.
DISCUSSION
Our studies have identified the nuclear hormone receptors the estrogen related receptors ERRα, ERRβ, and ERRγ as important modulators of age-related mitochondrial dysfunction, cellular senescence, and inflammation, that are also regulated by lifelong CR. ERRα, ERRβ, and ERRγ expression levels are decreased in the aging kidney, and caloric restriction results in increases in expression of ERRα, ERRβ, and ERRγ in the kidney.
Remarkably, only a 4-week treatment of 21-month-old mice with the pan ERR agonist reversed the age-related mitochondrial dysfunction, the cellular senescence marker p21, and inflammation. These effects were comparable with those achieved with lifelong CR, which is known to protect against age-related co-morbidities, including loss of renal function 22, 23, 67.
Recent evidence indicates that mitochondrial dysfunction of one of the mediators of cellular senescence and the associated senescence associated secretory phenotype (SASP) that includes pro-inflammatory cytokines and pro-fibrotic growth factors 68–74. This process may also be involved in the age-related inflammation that has been termed inflammaging or senoinflammation, which is also prevented by CR 75–79.
Recently mitochondrial dysfunction has also been linked to activation of the cGMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) signaling pathway, which plays key roles in immunity, inflammation, senescence and cancer 65, 66, 80–83. In addition to the recent identification of the importance of this signaling pathway in mouse models of chronic kidney disease and fibrosis 66, our studies also determine increased expression of STING in the aging kidneys, and its downregulation following treatment with the pan ERR agonist.
In addition to cGAS-STING signaling mitochondrial dysfunction is also associated with activation of STAT3 84. Increased STAT3 signaling is associated with senescence 85, 86 as well as kidney disease 87, 88. We have found increased STAT in the aging kidneys both at the mRNA and protein level, including increased phospho-STAT3 (Tyr705) and normalization after treatment with the pan ERR agonist.
At this time, it is not known whether the ERR induced decreases in STAT3 and STING mediate the decrease in the senescent cell marker p21 followed by decreases in IL-1β and ICAM1. However, our studies identify the ERRs as modulators of mitochondrial function, senescence, and inflammation in the aging kidney.
METHODS
Mice
Studies were performed in 4-month-old and 21-month-old male C57BL/6 male mice obtained through the NIA aging rodent colony. The mice were treated with 3% DMSO or the pan ERR agonist SLU-PP-332 at 25 mg/kg body weight/day administered intraperitoneally. The mice were treated for 4 weeks, following week they underwent anesthesia and the kidneys were harvested and processed for a) histology, b) transmission electron microscopy, c) isolation of nuclei, d) isolation of mitochondria, and e) biochemical studies detailed below.
Immunohistochemistry
Formalin-fixed paraffin-embedded tissue sections were subjected to antigen retrieval with EDTA buffer in high pressure heated water bath and staining was performed using either mouse monoclonal ERRα (1:2500, Abcam, Cambridge, MA) or ERRγ (1:400, Abcam) antibodies for 90 minutes or pyruvate dehydrogenase E2/E3bp (1:1000, Abcam) antibody for 45 minutes. Then UnoVue HRP secondary antibody detection reagent (Diagnostic BioSystems, Pleasanton, CA) was applied followed by DAB chromogen. The imaging was taken with Nanozoomer (Hamamatsu Photonics, Japan).
Transmission Electron Microscopy
One mm3 cortex kidney tissues were fixed for 48 hrs. at 4°C in 2.5% glutaraldehyde and 1% paraformaldehyde in 0.1M cacodylate buffer (pH 7.4) and washed with cacodylate buffer three times. The tissues were fixed with 1% OsO4 for two hours, washed again with 0.1M cacodylate buffer three times, washed with water and placed in 1% uranyl acetate for one hour. The tissues were subsequently serially dehydrated in ethanol and propylene oxide and embedded in EMBed 812 resin (Electron Microscopy Sciences, Hatfield, PA, USA). Thin sections, approx. 80 nm, were obtained by utilizing the Leica ultracut-UCT ultramicrotome (Leica, Deerfield, IL) and placed onto 300 mesh copper grids and stained with saturated uranyl acetate in 50% methanol and then with lead citrate. The grids were viewed with a JEM-1200EXII electron microscope (JEOL Ltd, Tokyo, Japan) at 80kV and images were recorded on the XR611M, mid mounted, 10.5M pixel, CCD camera (Advanced Microscopy Techniques Corp, Danvers, MA).
The minimum diameter of the mitochondrial short axis was measured in a minimum of 6 TEM images from each mouse (n = 3-4 each group). All mitochondria that were completely within the image field were measured. For each group, the image with the median diameter of the mitochondrial short axis closest to the median diameter for the whole group was chosen as the representative image. If two images were equally close, the image with a more normal distribution was chosen as the representative image for the group.
RNA extraction and real-time quantitative PCR
Total RNA was isolated from the kidneys using Qiagen RNeasy mini kit (Valencia, CA), and cDNA was synthesized using reverse transcript reagents from Thermo Fisher Scientific (Waltham, MA). Quantitative real-time PCR was performed as previously described 89–92, and expression levels of target genes were normalized to 18S level. Primer sequences are listed in Supplementary Table 1.
RNA-seq
Approximately 300-500ng of kidneys RNA were used to generate barcoded RNA libraries using Ion AmpliSeq™ Transcriptome Mouse Gene Expression Panel, Chef-Ready Kit. Precise library quantification was performed using the Ion Library Quantitation Kit (Thermo Fisher Scientific). Sequencing was performed on an Ion Proton with signal processing and base calling using Ion Torrent Suite (Thermo Fisher Scientific). Raw sequence was mapped to Ampliseq supported mm10 transcriptome. Quality control metrics and normalized read counts per million were generated using the RNA-seq Analysis plugin (Ion Torrent Community, Thermo Fisher Scientific).
RNA-seq of single nuclei
Mouse kidney single nuclei were isolated 54–57 and counted using the EVE Automated Cell Counter (NanoEnTek, VWR). The resulting mixture was provided to the Genomics and Epigenomics Shared Resource (GESR) at Georgetown University, and further processed by the Chromium Controller (10X Genomics, Pleasanton, CA) using Single Cell 3’ GEM Kit v3, Single Cell 3’ Library Kit v3, i7 multiplex kit, Single Cell 3’ Gel Bead Kit v3 (10X Genomics) according to the manufacturer’s instructions with modifications for single nuclei. Libraries were sequenced on the Illumina Novaseq S4 System (Illumina, San Diego, CA) to an average depth of >300 M reads PF per sample. Data Analysis: The 10XGenomics BCL data was loaded into the Cellranger Makefastq pipeline, which demultiplexes raw base call (BCL) files generated by Illumina sequencers into FASTQ files which are further analyzed by the Cellranger Count pipeline.
Inside the Cellranger Count pipeline, SART was used to align sequencing reads to Mouse MM10 genome reference. The Cloupe files were obtained and were further visually investigated by the Loupe Cell Browser.
Proteomics
Proteomics studies were conducted with Udayan Guha and Yue Qi at CCR, NCI, NIH. 200 microgram of cortical kidney sections were homogenized and lysed by 8M urea in 20mM HEPES (pH=8.0) buffer with protease and phosphatase inhibitors using Tissue Lyser II (QIAGEN). Samples were reduced and alkylated followed by MS-grade trypsin digestion. The resulting tryptic peptides were labeled with 11 plex tandem mass tag (TMT). After quench, the tagged peptides were combined and fractionated with basic-pH reverse-phase high-performance liquid chromatography, collected in a 96-well plate and combined for a total of 12 fractions prior to desalting and subsequent liquid chromatography−tandem mass spectrometry (LC−MS/MS) processing on a Orbitrap Q-Exactive HF (Thermo Fisher Scientific) mass spectrometer interfaced with an Ultimate 3000 nanoflow LC system 93. Each fraction was separated on a reverse phase C18 nano-column (25μm × 75cm, 2μm particles) with a linear gradient 4∼45% solvent B (0.1% TFA in Acetonitrile). Data dependent mode was applied to analyze the top 15 most abundant peaks in one acquisition cycle.
MS raw files were mapped against Uniprot mouse database (version 20170207) using the MaxQuant software package (version 1.5.3.30) with the Andromeda search engine 94, 95. Corrected intensities of the reporter ions from TMT labels were obtained from the MaxQuant search. For the TMT experiment, relative ratios of each channel to the reference channel (channel11, pooled from 20 samples) were calculated. Perseus (version 1.5.5.3) was used to further analyze and visualize the data. Hierarchical clustering of proteins was obtained in Perseus using log ratios or log intensities of protein abundance.
Metabolomic and Lipidomics analysis
These studies were conducted with Dr. Frank J. Gonzalez, Shogo Takahashi and Thomas J. Velenosi, and Daxesh P. Patel, CCR, National Cancer Institute, NIH. In these studies, we determined metabolites and lipid profiles in kidneys using Q-TOF-MS 96–101. Samples were analyzed by HILIC (Hydrophilic-interaction-liquid-chromatography) separation and mass spectrometry using a Waters Acquity H-class UPLC coupled to a Xevo G2 Q-Tof mass spectrometer for metabolomics. For lipidomics, samples were analyzed by UPLC-ESI-QTOFMS using a Waters Acquity CSH 1.7 µm C18 column (2.1×100 mm) (92, 93). For metabolomic and lipidomic data analysis, centroided and integrated chromatographic mass data was processed by Progenesis QI (Waters Corp., Milford, MA) to generate a multivariate data matrix. MS-DIAL (http://prime.psc.riken.jp/Metabolomics_Software/MS-DIAL/) and MS-FINDER (http://prime.psc.riken.jp/Metabolomics_Software/MS-FINDER/index.html) software was used for metabolites and lipid annotation 102, 103. The matrices were analyzed by principal components analysis (PCA) and supervised orthogonal projection to latent structures (OPLS) analysis using SIMCA. The OPLS, loadings scatter and S-plot analysis by SIMCA software were used to determine those ions that contribute to separation between groups.
Mitochondrial isolation
Kidney mitochondria were isolated using the kit from Sigma (St. Louis, MO) and followed the instruction accordingly.
Mitochondrial Biogenesis
We measured mitochondrial and nuclear DNA by RT-QPCR.
Mitochondrial Respiration
We measured basal respiration, ATP turnover, proton leak, maximal respiration and spare respiratory capacity using the Seahorse XF96 Analyzer on equally loaded freshly isolated mitochondria. We also measured mitochondrial complex I, II, III, IV, and V protein abundance by Native Blue Gel Electrophoresis (Thermo Fisher Scientific) with equally loaded mitochondrial fractions.
Multi-omics data analysis and integration bioinformatics methods
The project includes several types of omics data (proteomics, transcriptomics, metabolomics, and lipidomics) measured in the same set of biological samples. Taking the advantage of this multi-omics set of measurements that reflects the same biological processes in the samples from different perspectives, we performed Two-way Orthogonal Partial Least Square (O2PLS) integration 104 in pairs of the transcriptomics, proteomics, metabolomics, and lipidomics datasets. As result of each omics pair integration, matching orthogonal components in dataset-members of the pair were found that show what major mutually coordinated regulations were found in this particular pair of the omics datasets.
Majority of the gene expression and protein abundance profiles can be presented as combinations in different proportions of these three o2PLS components. An association of a gene/protein profile with one of individual components is defined by “loading” of the profile on this particular component – how high is a projection of the profile as a vector of the sample-values on the component as also a vector of the sample values. Subsets of the most associated with individual components genes and proteins were determined.
Western blotting
Western blotting was performed as previously described 89–92. Equal amount of total protein was separated by SDS-PAGE gels and transferred onto PVDF membranes. After HRP-conjugated secondary antibodies, the immune complexes were detected by chemiluminescence captured on Azure C300 digital imager (Dublin, CA) and the densitometry was performed with ImageJ software. Primary antibodies used for western blotting were listed in Supplementary Table 1.
ACKNOWLEDGEMENTS
The above study is supported by NIA R01 AG049493 and NIDDK R01 DK116567 to M.L. and AHA postdoctoral fellowship to K.M. (19POST34381041) and A.E. L. (19POST34430001).
REFERENCES
- 1.↵
- 2.
- 3.↵
- 4.↵
- 5.
- 6.↵
- 7.↵
- 8.
- 9.
- 10.↵
- 11.↵
- 12.
- 13.↵
- 14.↵
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.
- 35.↵
- 36.↵
- 37.
- 38.
- 39.
- 40.
- 41.
- 42.
- 43.
- 44.
- 45.
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.
- 53.↵
- 54.↵
- 55.
- 56.
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.
- 70.
- 71.
- 72.
- 73.
- 74.↵
- 75.↵
- 76.
- 77.
- 78.
- 79.↵
- 80.
- 81.
- 82.
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.
- 91.
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.
- 98.
- 99.
- 100.
- 101.↵
- 102.↵
- 103.↵
- 104.↵