Targeted inhibition of ubiquitin signaling reverses metabolic reprogramming and suppresses glioblastoma growth

Cell culture

Human glioblastoma cells (U87MG and U87MG-Luc) and human embryonic kidney cells (HEK293) were obtained from the American Type Culture Collection (Manassas, Va.). Cells were cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum (Gibco™ Fetal Bovine Serum South America, Thermo Scientific Fisher-US), 2 mM l-glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin, at 37 °C, 5% CO2 and 95% of humidity.

Plasmid and transfection

Vectors encoding praja2 and HA–ubiquitin were previously described15,19. In particular, praja2 inactive RING mutant (praja2rm) carries cysteines 634/637 changed to alanine19. KSR2-Myc was purchased from Genscript. siRNAs targeting praja2 and KSR2 were transfected using Lipofectamine 2000 (Invitrogen). For praja2 silencing experiments, we used a pool of four siRNAs and sequences (Dharmacon) are the following: sequence 1: 5′-GAGAUGAGUUUGAAGAGUU-3′; sequence 2: 5′-GGGAGAAAUUCCUUGGUUA-3′; sequence 3: 5′-UGACAAAGAUGAAGAUAGU-3′; sequence 4: 5′-UCAGAUGACCUCUUAAUAA-3′. KSR2 siRNA sequence (Thermo Fisher) is 5′-AAAUGCUGAAGAGUCCAAAGUCCGU-3′. Control siRNA was purchased from Ambion (am4637). For the cell growth curve, two independent siRNA mixtures were used: mix 1 (siRNA1 + siRNA3) and mix 2 (siRNA2 + siRNA4).

Immunoprecipitation and immunoblot

Cells were lysed with 1% Triton buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, EDTA 1 mM, 1% Triton, 5 mM MgCl2) supplemented with protease inhibitors, phenylmethylsulfonyl fluoride (PMSF) and phosphatase inhibitors. Lysates were incubated overnight with the indicated antibodies for immunoprecipitation; then, pellets were washed three times with lysis buffer. Precipitates and a quote of lysates were loaded on SDS polyacrylamide gel and blotted on nitrocellulose membrane. Filter were blocked with 5% milk in TBS-Tween 0.1% and incubated with primary antibodies overnight. After incubation with secondary antibodies, proteins were detected with ECL.

In vitro pull-down assay

GST-fusions were expressed and purified from BL21 (DE3) pLysS cells. For in vitro-binding assays, 20 μl of GST and GST-Praja2 beads were incubated in in vitro-translated, [35S]-labeled KSR2 in 200 μl lysis buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.5% Triton X-100) in rotation at 4 °C overnight. Pellets were washed four times in lysis buffer supplemented with NaCl (0.4 M final concentration) and eluted in Laemmli buffer. Eluted samples were size-fractionated on SDS-PAGE and immunoblotted or subjected to autoradiography.

Immunofluorescence and confocal analysis

Cells plated on cover glass were fixed with paraformaldehyde 5% (Merck), permeabilized with 0,3% Triton (Merck), blocked with BSA 5% (SERVA), and finally stained with the indicated primary antibodies. Signals were revealed with rhodamine-conjugated secondary antibodies (1:200, Invitrogen) and nuclei were stained with DAPI. Immunostaining was visualized using Zeiss LSM700 confocal microscope.

Materials and antibodies

The following chemicals were used: 1,2-dioleoyl-3-trimethylammonium-propane chloride (DOTAP), N-palmitoyl-sphingosine-1 {succinyl[methoxy (polyethylene glycol)2000]} (cer-PEG), 3ß-[N-(N‘,N‘-dimethylaminoethane)-carbamoyl]cholesterol hydrochloride (DC-chol), cholesterol (CHOL), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[amino(polyethylene glycol)-2000] (DSPE-PEG2000), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[amino(polyethylene glycol)-2000]-maleimide (DSPE-PEG-Mal), human transferrin (Tf), ammonium ferrithiocyanate, sodium borate, sodium chloride, sodium citrate, sodium phosphate, HEPES, ethylenediaminetetraacetic acid (EDTA), citric acid, Sepharose G-25, and 2-iminothiolane (Traut’s reagent) 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl ammonium salt) (Rhod), were purchased from Spectra2000 s.r.l. (Rome, Italy). Sodium chloride, sodium phosphate dibasic, calcium chloride, potassium chloride, were obtained from Merck Life Science s.r.l (Milan, Italy). Forskolin (#F3917) was purchased from Merck; AICAR (#A611700) was obtained from Toronto research chemicals; AMPK inhibitor SBI-0206965 was purchased from SIGMA (SML1540). The following primary antibodies were used: Flag (1:2000 immunoblot, 1:200 immunoprecipitation; #F3165, Merck); HA.11 (1:1000; #16B12, Biolegend); praja2 (1:1000 immunoblot, 1:200 immunofluorescence; #A302-991A, Bethyl Laboratories); phospho-Thr172-AMPKα (1:1000; #2535 S, Cell Signaling); AMPKα (1ː1000; ♯07350, Merck); Myc (1:1000 immunoblot, 1:500 immunofluorescence; #M4439, Merck); KSR2 (1:50 immunoprecipitation; #ab173377, Abcam); KSR2 (1:1000 immunoblot, 1:50 immunoprecipitation; #sc100421, Santa Cruz); HSP90 (1:5000 #60318-1-Ig, Proteintech); MOB1 (1:50, #ab236969, Abcam). Antibody-antigen complexes were detected by HRP-conjugated antibodies (Biorad) and ECL (Euroclone).

Proteomic analysis of praja2 protein complexes

HEK293 cells overexpressing Flag-praja2rm were harvested and lysed with a buffer containing 1% Triton X-100, 150 mM NaCl, 1 mM EDTA, and 50 mM Tris-HCl (pH 7.5) and supplemented with protease inhibitors, PMSF and phosphatase inhibitors. Lysates were immunoprecipitated with anti-Flag M2 affinity gel (Merck, # A2220), for 3 h. After three washes using lysis buffer, proteins were eluted by incubation with 3xFlag-peptide (Thermo Fisher, #A36805) 150 ng/μl in PBS, for 2 h. As a control, immunoprecipitation of lysates from HEK293 cells overexpressing Flag-empty vector was performed. Immunopurified proteins were analyzed by 10% T SDS-PAGE. After staining with colloidal Coomassie blue, whole gel lanes were cut into 15 slices, minced, and washed with water. Corresponding proteins were separately in-gel reduced, S-alkylated with iodoacetamide, and digested with trypsin, as previously reported58. Individual protein digests were then analyzed with a nanoLC-ESI-Q-Orbitrap-MS/MS platform consisting of an UltiMate 3000 HPLC RSLC nano system (Thermo Fisher Scientific, USA) coupled to a Q-ExactivePlus mass spectrometer through a Nanoflex ion source (Thermo Fisher Scientific). Peptides were loaded on an Acclaim PepMap™ RSLC C18 column (150 mm × 75 μm ID, 2 μm particles, 100 Å pore size) (Thermo Fisher Scientific), and eluted with a gradient of solvent B (19.92/80/0.08 v/v/v water/acetonitrile/formic acid) in solvent A (99.9/0.1 v/v water/formic acid), at a flow rate of 300 nl/min. The gradient of solvent B started at 3%, increased to 40% over 40 min, raised to 80% over 5 min, remained at 80% for 4 min, and finally returned to 3% in 1 min, with a column equilibrating step of 30 min before the subsequent chromatographic run. The mass spectrometer operated in data-dependent mode using a full scan (m/z range 375–1500, a nominal resolution of 70,000, an automatic gain control target of 3,000,000, and a maximum ion target of 50 ms), followed by MS/MS scans of the ten most abundant ions. MS/MS spectra were acquired in a scan m/z range 200–2000, using a normalized collision energy of 32%, an automatic gain control target of 100,000, a maximum ion target of 100 ms, and a resolution of 17,500. A dynamic exclusion value of 30 s was also used. Triplicate analysis of each sample was performed to increase the number of identified peptides/protein coverage.

MS and MS/MS raw data files per lane were merged for protein identification into Proteome Discoverer v. 2.4 software (Thermo Scientific), enabling the database search by Mascot algorithm v. 2.6.1 (Matrix Science, UK) with the following parameters: UniProtKB human protein database (11/2020, 214889 sequences) including the most common protein contaminants; carbamidomethylation of Cys as fixed modification; oxidation of Met, deamidation of Asn and Gln, and pyroglutamate formation of Gln as variable modifications. Peptide mass tolerance and fragment mass tolerance were set to ±10 ppm and ±0.05 Da, respectively. Proteolytic enzyme and a maximum number of missed cleavages were set to trypsin and 2, respectively. Protein candidates assigned on the basis of at least two sequenced peptides and a Mascot score ≥30 were considered confidently identified. Definitive peptide assignment was always associated with manual spectra visualization and verification. Results were filtered to a 1% false discovery rate. A comparison with results from the corresponding control allowed to identify contaminant proteins in each experiment that, nonetheless their abundance, were removed from the list of praja2-interacting partners (Supplementary Data 1).

Protein–protein interaction (PPI) network analysis

The protein network representing praja2 interactors was built starting from three sets of interaction data. The first set represents praja2 interactors directly identified in this study by proteomics. The 1538 interactors identified were filtered for proteins involved in metabolic pathways by using the gene IDs reported in KEGG pathways (KEGG:hsa04152, KEGG:hsa00190, KEGG:has_M00087, KEGG:hsa01212, KEGG:hsa00010), thus obtaining a network of 55 nodes and 54 edges. The second set of interactions was built by using all the 24 praja2 direct interactors reported in the Intact database []. The third set of interactions was built starting from CPTAC-GBM discovery cohort protein assay data, composed of 99 samples and 10,409 proteins. Data used in this work were generated by the Clinical Proteomic Tumor Analysis Consortium (NCI/NIH)59. Starting from the protein abundance matrix, the ARACNE algorithm was applied from the MINET package with default parameters and an interaction network composed of 10,409 nodes and 61,116 edges was generated60. Then, the subnetwork of 13 ARACNE interactors of praja2 was extracted. The three networks were loaded into Cytoscape version 3.7.2 and the STRINGify function from the stringApp tool was applied to each network61,62. Then, networks were merged using the Cytoscape merge function and performing graph union.

Total RNA extraction, libraries preparation, and sequencing with Illumina technologies

U87MG glioblastoma cells have been transfected, in biological triplicate, with a siRNA targeting the praja2 gene or control siRNA. Libraries preparation was performed as described previously63. Cells were collected and homogenized in Trizol Reagent. Total RNA extraction was performed with an RNA Clean and Concentrator Kit (Zymo Research Corp., Irvine, CA, USA), following the manufacturer’s instructions. RNA purity and integrity were assessed with a Nanodrop 2000c spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and a 4200 TapeStation instrument (Agilent Technologies, Santa Clara, CA, USA), respectively. For RNA purity, an A260/280 ratio of 2.0 and an A260/230 ratio of 2.0–2.2 were considered acceptable; for RNA integrity, an RNA Integrity Number (RIN) of 9.0–10.0 has been obtained for all samples, indicating the absence of degradation and high integrity of RNA samples. For a precise estimation of the RNA concentration, a Qubit 2.0 fluorometer assay (Thermo Fisher Scientific, Waltham, MA, USA) was used. For RNA sequencing, 1 μg of high-quality total RNA was used for library preparation with a TruSeq Stranded Total RNA Sample Prep Kit (Illumina, San Diego, CA, USA) and sequenced (paired-end, 2 × 75 cycles) on the NextSeq 500 platform (Illumina, San Diego, CA, USA).

RNAseq data and Gene Set Enrichment Analysis (GSEA) analysis

Data analysis was performed as described previously64. In detail, the raw sequence files generated (.fastq files) underwent quality control analysis using FASTQC ( and adapter sequences were removed using Trimmomatic version 0.3865. Filtered reads were aligned on the human genome (assembly hg38) considering genes present in GenCode Release 36 (GRCh38.p12) using STAR v2.7.5a with standard parameters66. Quantification of expressed genes was performed using featureCounts67 and differentially expressed genes were identified using DESeq268. A given RNA was considered expressed when detected by at least ≥10 raw reads. Differential expression was reported as |fold-change|(FC) ≥1.5 along with associated adjusted p value ≤0.10, which was computed according to Benjamini–Hochberg. About 10,200 transcripts were expressed in each biological replicate of the two conditions, setting a normalized read-count cutoff ≥10, while DESeq2 highlighted 725 transcripts differentially expressed with 526 transcripts upregulated (fold-change ≥1.5 and adjusted p value ≤0.10) and 199 downregulated (fold-change ≤−1.5 and adjusted p value ≤0.10)69,70,71. GSEA was performed to examine pathway enrichment for the differential expressed genes with the Molecular Signature Database “Hallmarks” gene set collection44. Only those with an FDR ≤0.25 were selected. MA-plot was generated using R ggpubr library, while heatmaps were generated using MultiExperiment Viewer tool72. The RNAseq raw data are publicly available in the ArrayExpress repository under accession number: E-MTAB-11137.

Metabolic assays

The real-time oxygen consumption rate (OCR) of human immortalized glioblastoma cells (U87MG) was measured at 37 °C using a Seahorse XF Analyzer (Seahorse Bioscience, North Billerica, MA, USA). U87MG cells were transfected with praja2 siRNA, with praja2 and KSR2 siRNAs or with a control, scrambled siRNA (siRNAc). Forty-eight hours after transfection, cells were plated into specific cell culture microplates (Agilent, USA) at the concentration of 3 × 104 cells/well, and cultured for the last 12 h in DMEM, 10% FBS. OCR was measured in XF media (non-buffered DMEM medium, containing 10 mM glucose, 2 mM l-glutamine, and 1 mM sodium pyruvate) under basal conditions and after the sequential addition of 1.5 µM oligomycin, 2 µM FCCP, and rotenone + antimycin (0.5 µM all) (all from Agilent). Indices of mitochondrial respiratory function were calculated from the OCR profile: basal OCR (before the addition of oligomycin), basal OCR, maximal respiration (calculated as the difference between FCCP rate and antimycin+rotenone rate), spare respiratory capacity (calculated as the difference of FCCP-induced OCR and basal OCR), ATP production (calculated as the difference between basal OCR and oligomycin-induced OCR). Reported data were the mean values ± SEM of four measurements deriving from four independent experiments. The acidification rate (ECAR) of human immortalized glioblastoma cells, U87 cells, was measured at 37 °C using a Seahorse XF Analyzer (Seahorse Bioscience, North Billerica, MA, USA). U87 were transfected with praja2 siRNA, with both praja2 and KSR2 siRNAs, and with a scrambled siRNA. 48 hours after transfection, 3 × 104 cells were reseeded in triplicate into specific cell culture microplates (Agilent, USA), previously coated with polylysine. ECAR was measured in glucose-deprived XF media (non-buffered DMEM medium, 2 mM l-glutamine, and 1 mM sodium pyruvate) under basal conditions and after the sequential addition of Glucose (10 µM), Olygomycin (1 µM), and 2 DG (50 mM).

Preparation of hybrid self-assembling nanoparticles (SANPs) encapsulating siRNA

SANPs were prepared as reported previously45. Thus, calcium-phosphate colloidal dispersion (CaP NPs) was prepared. Briefly, an aqueous solution of dibasic hydrogen phosphate (10.8 mM, pH 9.5) was added 1:1 v/v, drop by drop and under magnetic stirring, to an aqueous solution of calcium chloride (18 mM, pH 9.5) for 10 min, and filtered with RC filter (0.22 µm filter membranes of regenerated cellulose). CaP NPs were prepared before the use, and then mixed with an aqueous solution of siRNA scramble (Sc) or praja2 by vortex for 10 sec in a ratio of 8:1 v/v, followed by incubation for 10 min, finally resulting in CaP/siRNA NPs. Different PEGylated cationic liposomes (PLs) consisting of DOTAP/chol/DSPE-PEG2000 (mM ratio 1:1.8:0.125), DOTAP/DSPE-PEG2000 (mM ratio 1:0.125), DC-chol/DSPE-PEG2000 (mM ratio 1:0.125), DC-chol/CER-PEG (1:0.125 mM), DOTAP/chol/CER-PEG (mM ratio 1:1.8:0.125), and DOTAP/CER-PEG (mM ratio 1:0.125) were prepared by hydration of a thin lipid film followed by extrusion (Supplementary Table 3). In the case of formulations prepared with the fluorescent marker, Rhod (1% w/w of total lipids) was added to the lipid mixture. The lipid film was obtained by a rotary evaporator (Laborota 4010 digital, Heidolph, Schwabach, Germany) and hydrated with RNAse-free water, for 2 h. Then, liposome suspension was extruded by a thermobarrel extruder system (Northern Lipids Inc., Vancouver, BC, Canada) passing repeatedly the suspension under nitrogen through polycarbonate membranes with decreasing pore sizes from 400 to 100 nm (Nucleopore Track Membrane 25 mm, Whatman, Brentford, UK). Finally, for transferrin-targeted PLs, Tf was firstly thiolated using 2-iminothiolane (Traut’s reagent). Briefly, Tf was dissolved in 0.1 M Na-borate buffer pH 8, followed by the addition of Traut’s reagent (1:5 mol/mol). Thereafter, thiolated Tf was incubated with preformed PLs DOTAP/CER-PEG/DSPE-PEG2000-Mal (CER-PEG/DSPE-PEG2000-Mal 95:5 w/w), overnight, at room temperature. The unconjugated Tf was removed by molecular exclusion chromatography, using a Sepharose G-25 column. PLs were stored at 4 °C. Finally, SANPs-siRNA were prepared by mixing, in equal volume, CaP/siRNA NPs and different PLs by vortex, for 10 s. Plain SANPs, without siRNA, were prepared similarly. Each formulation was prepared in triplicate.

SANPs-siRNA characterization

Plain SANPs (without siRNA) and SANPs-siRNA were characterized in terms of mean diameter, polydispersity index (PI), and zeta potential (ζ). Measurements were performed by dynamic light scattering after sample dilution by the Nanosizer Ultra (Malvern, UK). For each formulation, the mean diameter, the PI, and the ζ values were calculated as the mean of measurements carried out on at least three different batches. Moreover, in the case of SANPs-siRNA, the siRNA encapsulation efficiency was also determined in the different formulations by an indirect measure of unencapsulated siRNA, separated by ultracentrifugation (Optima Max E, Beckman Coulter, USA) at 80,000 rpm, 4 °C, for 40 min. The supernatants were analyzed by UV (UV-1800, UV Spectrophotometer) at the wavelengths of 260 nm, and the concentration of siRNA was calculated by a calibration curve of siRNAc and praja2 (R2 = 0.999) in H2O. Each analysis was carried out in duplicate.


CD1 4–6 weeks old male nude mice (20–22-g body weight; Charles River, Calco, CO, Italy) were kept under controlled conditions (temperature, 22 °C; humidity, 40%) on a 12-h light/dark cycle with food and water ad libitum. Mice received standard rodent pelleted chow 4RF21 (Mucedola, Settimo Milanese, Italy) and water ad libitum. All animal experiments were performed in accordance with guidelines approved by the Italian Ministry of Health (589/2017-PR; July 21, 2017).

Orthotopic models of glioblastoma and SANPs treatments

U87MG cells were stereotaxically implanted into the left caudate nucleus (by using the following coordinates: 0.6 mm anterior to the bregma; 1.7 mm lateral to the midline; and 3.5 mm ventral from the surface of the skull of male mice under ketamine (100 mg/kg, i.p.)/xylazine (10 mg/kg, i.p.) anesthesia29. Cells (0.5 × 106 cells/5 μl) were implanted at an infusion rate of 1 μl/min. The needle was left in place 5 min after cell infusion before it was withdrawn. After 1 week from implantation, treatments with SANPs-siRNAs particles were performed as follows. Mice were treated every 48 h for 3 weeks by tail vein injection with SANPs-siRNAc (rhodamine-conjugated nanoparticles encapsulating siRNA scramble, control group) or SANPs-siPraja2 (rhodamine-conjugated nanoparticles encapsulating siRNA for praja2, treated group). At the end of treatment, mice were sacrificed and isolated brains were frozen in isopentane at −80 °C until the inclusion in OCT. For each brain, 10-μm-thick serial sections, from the beginning of the striatum to the hippocampus (1 section every 400 μm), were sliced. Subsequently, the sections were stained with Mayer’s hematoxylin and eosin (both from Diapath, Bergamo, Italy) and subjected to analysis for the quantification of tumor volume. The volumetric analysis was performed using software that measured the tumor area in each section and calculated the volume of the tumor according to Cavalieri’s method using the following formula: V = ∑(A)i × TS × n, where (A)I is the area of the tumor in level i, TS is the section thickness, and n is the number of sections disposed between the two levels73. Where indicated, the mouse model of glioblastoma was generated by intracerebral injection of 3 × 105 U87 MG-Luc cells in 5 μL of saline solution (NaCl 0.9%) in 4–6 weeks old CD1 mice deeply anesthetized. To verify the successful injection, bioluminescent intensity (BLI) was measured after 3 h with IVIS Spectrum (Perkin-Elmer, Waltham, MA, USA), after intraperitoneal injection of 150 mg/kg d-Luciferin potassium salt (Perkin-Elmer). According to BLI measurement after 7 days, mice were randomized into three experimental groups of 12 animals, and each group was treated with SANPs-siRNAc or SANPs-siPraja2 by tail vein injection. Treatments were repeated twice a week for 4 weeks; then four mice for each group were sacrificed and organs were collected. Eight animals for each group were monitored until they showed a weight loss between 10 and 20%.

Collection of human glial tumors

The immunohistochemical analysis for praja2 was carried out on low-grade and high-grade glioma samples (from Neurosurgery patients of IRCCS Neuromed, Pozzilli, Italy) classified according to histopathological WHO 2016 classification of CNS tumors. A total of 20 cases of glial tumors were used: ten gliomas carrying wild-type IDH1 and ten gliomas with IDH1-R132 mutation. The material used in this study represents waste material and all patients have given their informed consent.

Immunohistochemistry and biodistribution of rhodamine-conjugated SANPs-siRNA in brain sections

Brain OCT embedded sections were fixed with 70% ethanol and immunostained for the proliferation marker Ki67 (Roche prediluted), praja2 (1:100, Bethyl), MOB1 (1:50, Cell signaling), and KSR2 (Santa Cruz Biotech) with a Benchmark Ultra XT (Roche). Antibody detection was performed by using ultraview DAB detection kit (Roche). Negative control samples were incubated with secondary antibodies only. The number of Ki67-positive and total cells was determined in five random 100x magnification fields by using Zeiss Epifluorescence Microscope Nikon Eclipse 50i equipped with a Nikon DS-Ri1 camera. To demonstrate the internalization of the rhodamine-conjugated nanoparticles, a separate experiment was carried out. The nude mice inoculated stereotaxically were treated for 3, 6, 9, and 16 h with rhodamine-conjugated nanoparticles. At the end of the treatment, the animals were sacrificed and the fresh brains were included in OCT. The OCT embedded sections were fixed in 70% ethanol and nuclei counterstained with DAPI (VECTASHIELD Antifade Mounting Medium for Fluorescence) were examined under a fluorescence microscope (Zeiss Epifluorescence Microscope Nikon Eclipse 50i). For praja2 staining, formalin-fixed, paraffin-embedded tissues from the tumors were selected. Representative slides of each tumor were stained with hematoxylin and eosin. Immunohistochemistry for praja2 was performed automatically with a Nexes instrument (Ventana, Tucson, Ariz.). Antibody detection was performed using a multilink streptavidin–biotin complex method, and antibodies were visualized by a diaminobenzidine chromagen method. Negative control samples were incubated with primary antibodies only. The number of praja2-positive cells was determined in fifty random fields (75 × 100 µm2 each) by using the Image Pro Plus 6.2 software and a stereological microscope Zaiss Image M11. Our aim was to obtain objective baseline data for the study of the expression of praja2 in these tissues. The results were expressed in numerical densities of positive cells in low-grade glioma and high-grade tissue and positive cells for fields at 100x magnification.

Statistics and reproducibility

Each experiment was repeated three to five times as described in the legend figures. All results are expressed as mean ± SEM in dot plots. Data distribution and gene expression statistical analyses were performed using GraphPad Prism software (v5.0; GraphPad Software Inc., San Diego, CA), Microsoft Excel 2016 (v16.04471; Microsoft Office 2016) and Interactive Dotplot ( Comparisons of two groups were performed using a Student’s t-test. A p value of <0.05 was considered to be statistically significant.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Read more here: Source link