Targeted degradation of extracellular mitochondrial aspartyl-tRNA synthetase modulates immune responses

Ethics statement

Mouse studies were conducted under the supervision of the University Laboratory Animal Resource. All studies were approved by the Ohio State University Institutional Animal Care and Use Committee under protocol number 2019A00000019-R1. Human lung tissue was obtained from failed donors under Institutional Research Ethics Board approval at The Ohio State University. We investigated existing banked biospecimens from 40 patients with acute respiratory failure who were admitted to the medical intensive care unit (MICU) at the Ohio State University Wexner Medical Center and The James Cancer Hospital between May 2020 and December of 2021. After obtaining informed consent from patients or their legally authorized representatives, biospecimens and clinical data was collected and stored through the Ohio State University ICU Registry (BuckICU) (IRB protocol #2020H0175). Patients were enrolled within 48 h of ICU admission. Clinical data was coded with sample ID and no uncoded data was shared with the investigators of the study.

Cell culture and reagents

Human cell lines A549, BEAS-2B, HEK293T and THP1 cells were purchased from ATCC. Human CD14+ monocytes were purchased from Lonza. A549 and THP1 cells were cultured in RPMI (Gibco) supplemented with 10% FBS, HEPES, L-glutamine, penicillin/streptomycin, non-essential amino acids and sodium pyruvate (Gibco). BEAS-2B cells were cultured in HITES media: DMEM/F12 supplemented with 10% FBS, insulin, transferrin, hydrocortisone, β-estradiol, HEPES, L-glutamine and penicillin/streptomycin. Tet-inducible Luc and FBXO24 BEAS-2B cells were selected and cultured in HITES media with puromycin (2 μg/mL). HEK293T cells were cultured in DMEM supplemented with 10% FBS, HEPES, L-glutamine, penicillin/streptomycin. CD14+ primary cells were cultured in RPMI with 50 ng/mL Macrophage Colony Stimulating Factor for a week prior to use. Antibodies used for immunoblotting were anti-DARS2 (ProteinTech, 1:2000 dilution), anti-FBXO24 (Novus, 1:1000 dilution), anti-β-actin (1:5000 dilution, Sigma Aldrich) and the following antibodies were all purchased from Cell Signaling Technologies: TOM20 PGC1α, GAPDH, Cyclin-D1, Flag, and V5 all used at 1:1000 dilution. Other antibodies include FBXO45 (Aviva Systems Biology), FBXL2 (Aviva Systems Biology), FBXL18 (Ab-Nova), ubiquitin (Cell Signaling Technologies), IARS2 (ProteinTech), TFAM (Cell Signaling Technologies), Acetyl-Lysine 1:1000 dilution, FLAG-tag 1:1000 dilution, (both from Cell Signaling), HA-Tag 1:2000 dilution (Novus), and Tim23 1:1000 dilution (Santa Cruz Biotechnology). The recombinant human proteins DARS2, AARS2, DARS1 were from Origene and E1, E2 were purchased from Enzo Life Sciences. Plasmids include Skp-1 (GenScript, NM_170679.3), Cullin1 (AddGene, 19896), and Rbx (AddGene, 20717). An aptamer-based protein array was purchased from SomaLogic.

Mouse Precision Cut Lung Slices (PCLS)

Age and sex matched mice were euthanized in a CO₂ chamber, the trachea was exposed, and a blunt needle inserted. Lungs were expanded with 1.5% of UltraPure™ Low Melting Point Agarose (Thermo Fisher, Cat#16520100) in sterile medium (DMEM, Gibco) at 50 °C. The trachea was then ligated, lungs were carefully excised, transferred to a 15 mL conical tube in cold PBS and incubated on ice for 30 min. Lungs were cut axially into 400 µM slices using a vibratome (0.30 mm/sec; Leica VT 1200), suspended in 1 mL sterile DMEM/F12 and incubated at 37 °C with 5% CO₂. Media was changed every hour 4 times to wash off the agarose followed by overnight incubation in DMEM/F-12 medium (Gibco) with 10% FBS (Gibco) and 1% penicillin/streptomycin solution (Gibco).

Cell functional studies

Cell cycle progression was assayed via BrdU incorporation as a representative of total DNA volume. In brief, cells were pretreated using siRNA of interest for 48 h, followed by 4 h BrdU incorporation. Cells were then collected, fixed, permeabilized, stained and intensity was assayed by flow cytometry using the APC BrdU Flow Kit (BD Pharmingen) according to the manufacturer’s instructions. Flow ctometry data was recorded with BD FACSSymphony A1 (BD Bioscience) and analyzed with FlowJo (FlowJo, LLC). Cell migration/wound healing assays were conducted using cell culture wound healing inserts (Ibidi) according to the manufacturer’s instructions. Mitochondria OCR and dynamics were assayed using a Seahorse XFe96 analyzer (Agilent). A Seahorse XF Cell MitoStress Kit (Agilent) was used to conduct these experiments and were used according to protocols provided by the company. Cytokines were assayed via multiplex ELISA (Meso Scale Discovery) according to the manufacture’s protocol or individual ELISAs for IL-6, IL-1β, or TNF-α (Invitrogen).

Cell-free fractionation

Supernatants for DARS2 secretion studies were performed in cells cultured in FBS free culture media. Protein from supernatant was concentrated using Amicon Ultra-0.5 Centrifugal Filter Devices (Millipore) according to company protocol provided. Supernatant fractions were isolated via serial centrifugation. In brief 500 μL of clarified supernatant was centrifuged at 2000xg for 5 min, supernatant was saved, and the pellet (cellular debris) was washed once in cold PBS collected in protein lysis buffer. Supernatants were spun in new microcentrifuge tubes at 10,000xg for 5 min, the supernatant was collected, and pellet (MV fraction) washed and then collected. Remaining supernatants were next processed using Amicon Ultra-0.5 Centrifugal Filter Devices (Millipore) according to company protocol provided to isolate the exosome fraction.

Protein accumulation and half-life studies

For experiments examining protein stability, cells were treated with the proteasome inhibitor MG132 (10 μM) and lysosome inhibitor Bafilomycin A1 (100 nM) for 6 h or MLN7492 (1 or 5 μM) for variable time points. Protein half-life was determined via treatment of cells with cyclohexamide (40 μg/mL) and lysates collected in a time course study.

Plasmid and siRNA transfections

For cellular plasmid overexpression experiments, plasmid(s) were mixed with X-tremeGENE 9 (SigmaAldrich) at a ratio of 1 μg:2 μL in 300 μL OptiMEM (Gibco) per reaction. For siRNA studies, cells were transfected with a final concentration of 10 nM siRNA in a reaction mixture of 1:50 GenMute: 1x GenMute Buffer, reaction volume was dependent on well size. siRNA sequences are listed in Supplemental Data 1.

Generation of FBXO24 deletion in cell lines and mice

We generated Fbxo24 KO mice using CRISPR/Cas9 technology at the University of Pittsburgh. Briefly, guide RNAs were constructed and tested in blastocysts. The University of Pittsburgh Transgenic and Targeting core facility injected murine fertilized eggs with CRISPR/Cas9 RNA reagents and implanted injected embryos into pseudo-pregnant females. Appropriate gRNAs generated double stranded breaks resulting in a 600 bp deletion producing a non-functional allele. FBXO24 and control gRNAs are provided Supplemental Data 1. Generation of the KO mouse was confirmed by RFLP analysis and DNA sequencing. FBXO24-KO BEAS-2B cells were also generated using CRISPR technology as previously described⁴⁴. In brief BEAS-2B cells were transfected with a plasmid encoding Cas9, and sgRNA against FBXO24 or scrambled sgRNA and GFP. Cells were flow sorted for GFP positive cells and expanded for culture and freezing. To generate doxycycline-inducible FBXO24 cell lines, FBXO24 was cloned into a pSBtet-RP backbone containing a puromycin resistance gene under control of a RPBSA promoter. BEAS-2B cells were transfected with this plasmid and then selected with puromycin for stable integration. Luciferase control cells were generated in the same fashion.

Bacterial preparations

Tryptic Soy Agar plates were streaked with 1–5 µL Pseudomonas aeruginosa, strain PA103 (ATCC) glycerol stock and grown overnight in a 37 °C bacterial incubator. An individual colony was placed into 5 mL Tryptic Soy broth and cultures were incubated at 37 °C and shaken at 250 rpm overnight. Cultures were then diluted 1:10-1:20 and cultured an additional hour and an absorbance (OD600) using Nanodrop One (Thermo Fisher Scientific) was used to determine cfu/ml. Liquid cultures were diluted for desired cfu/mL in cell culture media for in vitro or PBS for in vivo experiments. Klebsiella pneumoniae strain NCTC 9633 (ATCC), E. coli strain Seattle 1946 (ATCC 25922) and Staphylococcus aureus were cultured in a similar manner. S. pneumoniae, strain CIP 104225 (ATCC), was grown on Blood TSA plates (McKesson Medical Surgical) and cultured in Brain Heart Infusion Broth (Becton Dickenson) at 37 °C with 5% CO₂.

Experimental pneumonia

All animal experiments were approved by The Ohio State University Institutional Animal Care and Utilization Committee. Mice were housed in the OSU ULAR Vivarium on a 12 h light/dark cycle at a temperature of 65–80 °F at 30–70% humidity. Fbxo24-KO or Wt mice from a C57BL/6 J background were given 5 × 10⁵ cfu PA103/mouse, 3 mg/kg LPS, or PBS i.n., or for the systemic inflammation model, LPS i.p. administration (20 mg/kg) or PBS. Mice were euthanized 24 h post infection. Bronchoalveolar lavage (BAL) was collected via exposing the trachea, creating an incision, and inserting a blunted needle in the trachea. The lungs were then flushed with 1 mL PBS. BAL was then centrifuged at 1500 rpm for 10 min and supernatant transferred to a fresh tube for determination of protein concentration via Lowry assays and cytokine secretion via multiplex ELISA (Meso Scale Discovery). The remaining cell pellet was resuspended in 200 µL ACK solution (Gibco) and incubated on ice for 10 min to lyse red blood cells. To stop the reaction, 1 mL PBS with 0.1 mM EDTA was added, and the solution was centrifuged at 800 g for 10 min, supernatant was aspirated, and cells were resuspended in 400 µL PBS for differentials. Total cell count was performed with trypan blue. 200 µL of cell suspension was spun down on 2 cytology funnel slides (Fisher Scientific) per sample. Slides were processed with Giemsa staining, modified (Sigma-Aldrich) and May-Grünwald staining (Sigma-Aldrich). 100 cells/slide were counted and scored as polymorphonuclear or mononuclear cells; percent average of counts was multiplied by the total cell count and expressed as number of cells. Blood samples for multiplex ELISA were collected by puncturing the heart. Lung and live tissue were harvested for gene expression analysis.

Bacterial loads

To determine bacterial loads mice were grouped and inoculated as described above. Mice were euthanized 24 h later, lungs were excised, collected in 1 mL sterile PBS and homogenized. Lung homogenate was diluted 1:10 8 times and 5 μL from each dilution was plated on marked sections of a TSA plate. The first dilution to display individual colonies was used to calculate bacterial. CFU/mg was calculated as: \(\frac{({colony\;\#\; at\; Dilution\; X} \,*\, (10^{Dilution\; X}) \,*\, 200)}{{mg\; of\; lung\; tissue}}\)

Lung function studies

Fbxo24 KO or WT mice from a C57BL/6 J background were given 1 × 10⁵ cfu PA103/mouse, Lung function was assayed using a FlexiVent FX2 (SciReq). In brief, mice were anesthetized with ketamine/xylazine and received a tracheostomy to install the FlexiVent breathing tube. Once no respiratory effort was detected a mouse lung function program was run to assess several parameters of lung mechanics.

Lung inflammation analysis

Mouse lungs were fixed in 10% formalin and sent to the Histowiz for sectioning and H&E staining. A lung injury score was quantified based on assigning 0 (none), 1 (moderate) or 2 (severe) based on individual criteria (in parenthesis) to number of infiltrating neutrophils (none, 1-5, >5), number of hyaline membranes, (none, 1, >2), proteinaceous debris in alveoli (none, 1, >2) and alveolar thickening (<2x, 2-4X, >4x). Individual scores were averaged for a final lung injury score. The scoring system used is based on the Official American Society Workshop⁴⁵. All images were quantified by a minimum of two blinded scorers and their assessments were averaged.

In vivo administration of recombinant protein

Recombinant human DARS2 and AARS2 were prepared in protein transfection buffer (ProJect). 5μg of recombinant human DARS2 or AARS2 was given to mice either i.t. or i.p., protein transfection reagent was used as vehicle. Mice were euthanized 4 h or 24 h after administration. BAL was collected for differential analysis and multiplex ELISA for cytokine profile. Lung tissue was collected for histology. The peritoneal cavity was washed with 10 ml of PBS/ EDTA (0.1 mM) solution and total cell counts were measured.

In vivo administration of BC-1293

BC-1293 was prepared in PBS with 5% DMSO with a final concentration of 20 ng/μL and 50 ng/μL. C57BL6 mice were assigned to three experimental groups: DMSO, 2 μg or 5 μg BC-1293 (n = 5/group) and received 100 μL of respective stocks i.t. Mice were euthanized 16 h post-treatment and BAL was collected for differential analysis of immune cells and multiplex ELISA for cytokine concentration.

Immunoblotting

Cells were collected and lysed in Protein Lysis Buffer A: PBS with 0.2% SDS, 0.05% 100X-Triton and 1 Pierce Protease Inhibitor Tablets/10 mL, (Thermo Scientific, A32963) then sonicated for 20 sec at 25% on a VibraCell Sonicator (Sonics). Protein concentration was measured by Lowry Assay. Samples were diluted with Laemmli buffer and SDS-PAGE was performed with precast gradient mini- or midi- gels and transferred to nitrocellulose membranes using a Trans-Blot Turbo Transfer apparatus. All immunoblotting supplies were from Bio Rad. Densitometry was performed using ImageJ software (NIH) or Image Lab software (Bio Rad).

Immunoprecipitation (IP)

Prior to lysis cells were treated for 6 h with the proteasome inhibitor MG132 (100 μM) and lysosome inhibitor Bafilomycin A1 (200 nM). Cells were lysed in protein lysis Buffer A with deubiquitinase inhibitors 1,10-phenanthroline, PR-619 and N-ethylmaleimide. For experiments examining acetylation, cells were pretreated as described, but Protein Lysis Buffer A was made with 100x Deacetylase Inhibitor Cocktail (ApexBio) in place of DUB inhibitors. Lysates were sonicated and protein was normalized as described above. Cells were then incubated while rotating at RT for 30 min with anti-FlagM2 Magnetic Beads (Millipore-Sigma), washed with lysis buffer, washed twice with IP Wash buffer (0.2% Triton-X100 in PBS), and then eluted in 2x Laemmli via boiling. IP of biotinylated proteins for the TurboID screen was done as previously described⁴⁶. For post-translational modifications studies, DARS2-Flag constructs were expressed in HEK293T cells. Cells were lysed and DARS2 was subjected to IP as described above and samples processed for analysis by MS.

Proximity dependent biotinylation assay

As before⁴⁶, mCherry FBXO24-Wt and a catalytically inactive FBXO24-LPAA variant were cloned into a Sleeping Beauty backbone conjugated with a TurboID construct. Constructs were stably integrated into A549 cells and then treated with Biotin (50 μM) for 4 h. Samples were then subjected to IP as described above. Biotinylated proteins were identified by mass spectrometry (MS) analysis conducted by the MS and Proteomics Facility at the OSU Campus Chemical Instrument Center.

Mass spectrometry sample processing and analysis

Beads were washed with 50 mM ammonium bicarbonate three times. Then 5uL of DTT was added and the sample is incubated at 65^oC for 15 min, then 5uL of iodoacetamide (15 mg/ml) added and the sample is kept in dark at room temperature for 30 min.1ug of sequencing grade-modified trypsin (Promega, Madison WI) was added followed by incubation at 37^oC for overnight. The reaction is quenched the next morning by adding acetic acid for acidification. Supernatant was taken out, concentrated for LC/MSMS analysis. Nano-liquid chromatography-nanospray tandem mass spectrometry (Nano-LC/MS/MS) of protein identification was performed on a Thermo Scientific orbitrap Fusion mass spectrometer equipped with an nanospray FAIMS Pro™ Sources operated in positive ion mode. Samples (6.4 µL) were separated on an easy spray nano column (PepmapTM RSLC, C18 3 µ 100 A, 75 µm X150mm Thermo Scientific) using a 2D RSLC HPLC system from Thermo Scientific. Each sample was injected into the µ-Precolumn Cartridge (Thermo Scientific) and desalted with 0.1% Formic Acid in water for 5 min. The injector port was then switched to inject and the peptides were eluted off of the trap onto the column. Mobile phase A was 0.1% Formic Acid in water and acetonitrile (with 0.1% formic acid) was used as mobile phase B. Flow rate was set at 300nL/min. mobile phase B was increased from 2% to 16% in 105 min and then increased from 16-25% in 10 min and again from 25-85% in 1 min and then kept at 95% for another 4 min before being brought back quickly to 2% in 1 min. The column was equilibrated at 2% of mobile phase B (or 98% A) for 15 min before the next sample injection. MS/MS data was acquired with a spray voltage of 1.95 KV and a capillary temperature of 305 °C is used. The scan sequence of the mass spectrometer was based on the preview mode data dependent TopSpeed™ method: the analysis was programmed for a full scan recorded between m/z 375-1500 and a MS/MS scan to generate product ion spectra to determine amino acid sequence in consecutive scans starting from the most abundant peaks in the spectrum in the next 3 seconds. Three compensation voltage (cv = -50, -65 and -80v) were used for samples acquisition. The AGC Target ion number for FT full scan was set at 4 x 105 ions, maximum ion injection time was set at 50 ms and micro scan number was set at 1. MSn was performed using HCD in ion trap mode to ensure the highest signal intensity of MSn spectra. The HCD collision energy was set at 32%. The AGC Target ion number for ion trap MSn scan was set at 3.0E4 ions, maximum ion injection time was set at 35 ms and micro scan number was set at 1. Dynamic exclusion is enabled with a repeat count of 1 within 60 s and a low mass width and high mass width of 10ppm. Data were searched using Mascot Daemon by Matrix Science version 2.7.0 (Boston, MA) via ProteomeDiscoverer (version 2.4 Thermo Scientific,) and the database searched against the most recent Uniprot databases. A decoy database was also searched to determine the false discovery rate (FDR) and peptides were filtered according at 1% FDR. Proteins identified with at least two unique peptides were considered as reliable identification. Any modified peptides are manually checked for validation. Label free quantitation1 was performed using the spectral count approach. The normalization scheme in Scaffold adjusts the sum of the selected quantitative value for all proteins in the list within each MS sample to a common value: the average of the sums of all MS samples present in the experiment. Student-t test was performed by scaffold (Proteome Software, Portland, OR) to evaluate if the folder change for certain proteins is significant (p < 0.05).

Quick coupled transcription and translation and in vitro ubiquitylation

In vitro ubiquitination studies were conducted using Ubiquitylation kits (Enzo Life Sciences) according to manufacturer’s instructions. Recombinant proteins not included in the kit were produced using a TnT Quick Coupled Transcription/Translation Systems Kit (Promega) according to the protocol included.

Protein-compound interaction assay

In brief, Flag-tagged FBXO24 was transiently overexpressed in HEK293T cells subjected to IP. FBXO24 bound to beads suspended in PBS was then incubated with increasing concentrations of BC-1293 or a control compound, BC-1395, for 24 h at 4 °C. Recombinant DARS2 or DARS1 were then added to the bead-FBXO24-compound reactions for 4 h at RT. Following this, beads were washed thoroughly with IP wash buffer to remove any excess unbound protein. Finally, captured protein was eluted in 2x Laemmli via boiling and samples underwent immunoblot analysis.

Plasmids

DARS2 primary gene sequence was isolated from cDNA from HEK293T cells and cloned into a pcDNA-3.1 backbone under the control of a EF1α promoter with a Flag tag. FBXO24 primary gene sequence was cloned into a pcDNA-3.1 backbone under the control of a EF1α promoter with a V5 tag. FBXO24-LPAA mutants were generated by site directed mutagenesis. DARS2 deletion and K → R and K →Q substitution mutants were generated through site directed mutagenesis from the original pcDNA-DARS2-Flag vector. The carboxyl-terminal deleted residues are shown in Table 1.

Ubiquitin mutants were generated using site directed mutagenesis from a pRK5 plasmid harboring ubiquitin WT with a HA tag. All plasmids were sequence confirmed by a core at OSU and protein expression was validated in vitro.

Confocal microscopy

For confocal microscopy cells were cultured on culture slides (Lab-Tek). Media was aspirated, cells were fixed in 10% formalin and permeabilized with 0.5% triton and blocked with 1% BSA. Cells were then stained with primary antibody for V5 (Invitrogen), Flag (Cell Signaling Technologies) and TOM20 (Cell Signaling Technologies) at ^o4 C overnight, then fluorescent secondary antibody at RT for 1 h, and counterstained with DAPI. MitroTracker Deep Red (Invitrogen) was performed according to the manufacture’s protocol Images were obtained using an Olympus FV3000 Spectral Confocal System microscope. Image analysis and image preparation was done with ImageJ (NIH).

Gene expression

To analyze changes in gene expression cells were lysed and processed using the RNeasy Plus Kit (Qiagen) according to manufacturer’s protocol. Mouse and human tissue were processed using a miRNeasy Mini Kit (Qiagen). Reverse transcription reactions produced cDNA with the High-Capacity RNA-to-cDNA™ (Applied Biosystems). Expression levels were quantified by RT-qPCR using the SYBR Green system with a CFX96 Real-Time System (Bio Rad). Primers are listed in a Supplementary Table 2.

Administration of recombinant proteins

DARS2 and AARS2 were added to Project Protein transfection reagent (ProJect) or Lipofectamine (ThermoFisher) according to manufacturer’s instructions. For cell-based studies, these reactions were suspended in optiMEM and added to culture.

Molecular docking studies and compound design

The docking experiments were carried out by using software from Discovery studio 4.1. A custom library containing 100,000 diverse molecules were first used to screen potential ligands for FBXO24. Based on the docking and best-fit analysis of suitable ligands, several compounds were tested for FBXO24 inhibitory activity including BC-1293.

Clinical Cohort Sample Collection

Biospecimens, including peripheral blood collected in sodium citrate tubes, are collected on days 1, 7, and 21 of ICU admission. Following collection, blood is centrifuged to separate plasma and cells. Plasma is stored in aliquots at -80^oC. Clinical data is adjudicated by at least 2 physician-scientists certified in critical care medicine.

Analysis of human plasma samples

We analyzed banked plasma samples from 40 distinct patients on days 1, 7, and 21 of ICU admission. Plasma protein abundance was analyzed by an aptamer-based array using the SomaScan® platform by SomaLogic, Inc. We used a R (version 4.2.2) statistical program to analyze data. To perform multiple comparisons of numerical variables, we utilized ANOVA for normally distributed data, and the two-sample Wilcoxon test (Mann-Whitney) or Kruskal-Wallis tests for non-normally distributed data. Chi-square tests were used to compare categorical variables among the different groups.

RNA-Seq

RNA was extracted from cells using RNeasy Plus mini kit according to manufacture protocol (Qiagen). RNA quality, mRNA library preparation (polyA enrichment), sequencing using NovoSeq P150 (6 G raw data per sample) and data quality control was done by Novogene. ROSALIND® (https://rosalind.bio/) was used to analyze the RNAseq raw data. The HyperScale architecture has been developed by ROSALIND, Inc. (San Diego, CA). Cutadapt was used to trim the reads⁴⁷ and FastQC was used to obtain quality scores⁴⁸. STAR was used to align the reads were to the Homo sapiens genome build GRCh38⁴⁹. The quantification of individual sample reads was performed using HTseq⁵⁰ and subsequent normalization was carried out using the Relative Log Expression (RLE) method implemented in the DESeq2 R package⁵¹. Read distribution percentages, violin plots, identity heatmaps, and sample multidimensional scaling (MDS) plots were generated as part of the quality control (QC) process using the RSeQC software⁵². The DEseq2 package was utilized for the computation of fold changes and p-values, as well as for the implementation of optional covariate correction. The clustering of genes for the final heatmap of differentially expressed genes was performed using the PAM (Partitioning Around Medoids) approach, utilizing the fpc R library⁵³. The hypergeometric distribution was employed to assess the enrichment of pathways, gene ontology, domain structure, and other ontologies. The topGO R library⁵⁴ was employed to assess local similarities and dependencies among Gene Ontology (GO) concepts for the purpose of implementing Elim pruning correction. Differential expression analysis was performed to identify genes that exhibited significant changes in expression levels. The criteria for determining differentially expressed genes (DEGs) included a fold-change threshold of 1.5 and a statistically significant adjusted P value of less than 0.05. The heatmaps were constructed using the ROSALIND software, displaying the log₂ expression values that have been standardized across the heatmap. The downstream investigation of cellular function and pathway enrichment was conducted using Ingenuity Pathway investigation, an advanced bioinformatics software tool developed by Qiagen in Ann Arbor, MI (www.ingenuity.com). Cell function enrichment of DEGs was shown as number of DEGs for each cell function cluster. Pathway enrichment was shown as percent of pathway specific DEGs vs. total DEGs with the highest predicted up- or downregulation using a cut-off of absolute |2.5| for the activation Z-score. A positive Z score predicts overall activation of the pathway, and a negative Z score predicts downregulation of the pathway, based on the number of DEGs that are up- and downregulated in each pathway cluster.

Statistical analysis

Student’s t-tests were used for experiments with two groups. One-way ANOVA with multiple comparisons were used for experiments exceeding two groups with a single independent variable. Two-way ANOVAs were used for analysis of groups with multiple independent variables, and repeated measures was used when appropriate. Recommend tests for variance of standard deviations were used and associated corrections for non-parametric data were used as needed. Outliers were removed using Data with Robust regression and Outlier removal (ROUT) method with Q = 0.1% as the threshold for removal. Statistical analysis was performed in GraphPad Prism (GraphPad) unless otherwise noted.

Reporting summary

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