After UPEC superinfection challenge, DEFA1A3 is expressed in gene dose–dependent fashion, and higher mRNA expression correlates with a lower bacterial burden

We evaluated the α-Defensin 1-3/DEFA1A3 gene dose–dependent mechanism(s) of protection by inducing UTIs with double transurethral inoculation of UPEC; CFT073, 3 h apart, as previously described in the superinfection challenge (36). In our previous study, kidney DEFA1A3 mRNA expression inversely correlated with bacterial burden under single and superinfection UPEC challenges (9). Because both PMNs and collecting duct epithelial cells can express α-Defensin 1-3, the inducible cellular source of DEFA1A3 expression that inversely correlates with the lower bacterial burden phenotype remains to be dissected. Previous studies have characterized an increased period of bacteriuria, infection of urinary tract organs, and acute inflammation after the UPEC superinfection challenges under the C57BL/6J background (36). Comparing the superinfection-challenged mice with absent, low, and high human DEFA1A3 DNA copy numbers (DEFA^0/0, DEFA^4/0, and DEFA^4/4), our results indicate transgenic DEFA^4/4 mouse bladders displayed significantly lower average mean bacterial CFUs at 6 hpi compared with DEFA^4/0 and DEFA^0/0 mice (Fig 1A). Kidney tissues from DEFA^4/0 and DEFA^4/4 mice showed a decrease in CFU burden compared with DEFA^0/0 (Fig 1B). After superinfection challenges at 6 hpi, DEFA^4/4 (8 total copies) kidneys significantly increased DEFA1A3 mRNA expression by an average of 20-fold, far exceeding the twofold increase in infected DEFA^4/0 (four total copies) mouse kidneys (Fig 1C). Collectively, these results suggest transgenic DEFA1A3 DNA copy number–dependent mRNA expression is inversely proportional to bacterial burdens in the kidney and bladder.

Induction of intercalated cell α-Defensin 1-3/DEFA1A3 expression contributes to reduced urinary tract bacterial burden under the UPEC superinfection challenge

Because transcription and translation of PMN-derived α-Defensin 1-3/DEFA1A3 occur during promyelocytic stages in the bone marrow, we evaluated the possibility that α-Defensin 1-3 from other cellular sources represent the increased DEFA1A3 mRNA expression in the infected kidney (37, 38). Collecting duct epithelial–derived ICs were postulated as the source of DEFA1A3 induction in the kidney because of shared physiology with other constitutive and inducible renal AMPs (22, 39, 40). After enrichment of kidney-derived PMNs, collecting duct epithelial ICs, and kidney cells that are not intercalated cells (non-ICs), we performed quantification of DEFA1A3 gene expression that reveals the increased mRNA expression of 13.2-fold ± 0.71 in ICs after the UPEC superinfection challenge compared with other cell types (Fig 1D). Our results indicate an inverse relationship between kidney bacterial CFUs and IC-DEFA1A3 mRNA expression upon induction (r = −0.570), but not in DEFA1A3-expressing PMN sources (r = 0.254) (Fig S1C). Non-IC–derived DEFA1A3 mRNA expression lacked correlation relationships to bacterial CFUs in the kidney (r = −0.063), although non-IC mRNA expression increased to a lesser extent than ICs (2.1 ± 0.33 mRNA fold expression) (Figs 1D and S1D and E). To confirm expression at the protein level, we analyzed immunofluorescence co-localization of α-Defensin 1-3 with structural kidney cells by staining human pyelonephritis, and mouse post-challenged kidney sections with the IC marker (V-ATPase G3) (Fig 1E–G). We confirmed α-Defensin 1-3 co-localization with V-ATPase E1 marker for ICs in both human and mouse kidney sections (Fig S1A and B). The results suggest collecting duct epithelial intercalated cells primarily induce α-Defensin 1-3/DEFA1A3 expression to a lower bacterial burden in the setting of UPEC-induced superinfection.

PMNs, but not PMN-derived α-Defensin 1-3/DEFA1A3 expression, are critical to urinary tract defense after the UPEC superinfection challenge

Upon uropathogen invasion, PMNs migrate from the bloodstream to the kidney to release antimicrobial effectors stored in granules (41). To determine whether neutrophils are necessary to mediate DEFA1A3-dependent antimicrobial effects, we depleted Gr-1⁺ cells in DEFA^4/4 and DEFA^0/0 littermate mice before superinfection challenges. After depletion of neutrophils and challenge, CFU analysis shows infected urinary tract tissues between mouse groups were not different in the bladder nor kidney tissues (Fig 2A and B). Urinary tract tissues from challenged DEFA^4/4 mice with Gr1+ cell depletion had increased bladder and kidney bacterial burdens but to a similar degree as the WT counterpart (P = 0.528 for bladder and P = 0.291 for kidneys). In the infected urinary tract tissues of DEFA^4/4 and mice after UTI challenges, Gr1⁺ cell depletion led to increased bacterial burdens compared with IgG isotype treatment (Fig 2). Thus, neutrophils are needed to confer urinary tract protective phenotypes against UPEC independent of their α-Defensin 1-3 production. Interestingly, the DEFA^4/4 mouse kidneys had lower bacterial burdens in the kidneys compared with WT after Gr1⁺ cell depletion, indicating some protective effect potentially from renal α-Defensin 1-3 production similar to our transplant mouse data in previous work (Fig 2B) (9).

α-Defensin 1-3 permeabilize bacterial membranes and agglutinate E. coli strains in time- and dose-dependent manners

α-Defensin 1-3 agglutinate Escherichia coli in addition to permeabilizing and blebbing bacterial membranes (26, 32, 42). To test whether these effects were dependent on time exposure and dosage, we established time-kill antimicrobial assays that co-incubate α-Defensin 1-3 peptides with three increasing sub-inhibitory concentrations against laboratory; K12 and pyelonephritis; CFT073 E. coli strains (Fig 3A and B). α-Defensin 1-3 at a minimum of 10 μg/ml concentration significantly decreased bacterial counts as early as 3 and 24 h post-incubation after being sub-cultured overnight for CFU measurements in both strains. α-Defensin 1-3 had minimal bacteriostatic and bactericidal capacity against pyelonephritis multi-drug–resistant strain; MDR58 at 3 h, and viable bacterial counts recovered to baseline after 24 h post-incubation (Fig 3C). We used fluorescent dyes to visualize and quantify direct antimicrobial effects and agglutination. Propidium iodide (PI) identifies damaged/permeabilized bacteria, and SYTO9 stains all bacteria (Fig S2); hence, these can be used to see live versus dead bacteria by comparing SYTO9 with PI. At 3 h post-incubation, the in vitro assays indicated bacterial membrane permeabilization and agglutination at 10 μg/ml of α-Defensin 1-3 compared with vehicle-treated E. coli strains (Fig S2A–F). When the concentration was increased to 100 μg/ml, α-Defensin 1-3 peptides were able to induce higher significant agglutination effects similarly visualized across E. coli strains as evidenced by immunofluorescence (Fig S2C and F). Furthermore, MDR58 co-incubation with α-Defensin 1-3 at 10 μg/ml lacked membrane permeabilization and agglutination effects after 3 h (Fig S2G and H).

We then quantified the total surface area of agglutination in all bacteria (SYTO9) and damaged (PI) bacterial images across E. coli laboratory and pyelonephritis strains with escalating concentrations of α-Defensin 1-3 (Fig 3D and E). α-Defensin 1-3 concentrations led to a dose-dependent positively correlated increase in bacterial agglutination with significantly larger “clumps” of live and damaged bacteria in the laboratory strain and pyelonephritis strain; CFT073. Significant agglutination was quantified for both channels using the 100 μg/ml concentration of α-Defensin 1-3 peptides (P = <0.0001 and 0.006, respectively) compared with vehicle co-incubations in the pyelonephritis strain. Co-incubations of multi-drug–resistant UPEC; MDR58 and α-Defensin 1-3 peptides did not result in significant bacterial agglutination (P = 0.999 and 0.705) in neither of recorded SYTO9 nor PI at 100 μg/ml of α-Defensin 1-3 peptides (Fig S2I). The agglutination and antimicrobial activity were eliminated after co-incubation with anti-α-Defensin 1-3 blocking antibody (Figs 3F and S2J and K). Cumulatively, α-Defensin 1-3 direct membrane permeability and agglutination are bactericidal effects that occur in a time- and dose-dependent manner against E. coli and are dependent on pathogen-specific factors such as virulence or antibiotic resistance mechanisms.

α-Defensin 1-3 have complementary antimicrobial effects with LL-37, RNase7, and DMBT1 peptides against multi-drug–resistant UPEC

Prior studies have initially elucidated how different AMPs orchestrate damaging responses against bacteria (9, 32, 34). We evaluated the ability of α-Defensin 1-3 to induce direct bacterial membrane permeability and agglutination effects in combination with known human AMPs LL-37, RNase7, and DMBT1. Prior research efforts have elucidated the antimicrobial activities of these AMPs against bacterial species including E. coli (9, 43, 44, 45). Alone, LL-37 and DMBT1 peptides induced membrane damage and agglutination after 3 h of co-incubation with pyelonephritis strain; CFT073 (Fig S3A and E). Alternatively, RNase7 only induced membrane damage but lacked agglutinating activity against bacteria (Fig S3C). By comparing SYTO9 and PI quantification of mean total surface area after co-incubations of different AMPs, we were able to assess agglutination versus bactericidal activity in a cooperative fashion against the pyelonephritis strain; CFT073 (Figs 4A–C and S3D). When co-incubated with α-Defensin 1-3, cooperative effects were observed with LL-37, DMBT1, and RNase7 based on bacterial agglutination (Fig S3B, D, and F). Quantification of agglutination indicates a significant increase in the mean total surface area of all bacteria when used in combination with LL-37 and DMBT1 peptides (P = 0.0008 and <0.0001) (Fig 4A and C). When compared to the PI staining, agglutination of viable bacteria increased with the combination of α-Defensin 1-3 and LL-37 (Fig 4A). A mild increase in the mean total surface area of damaged agglutinated bacteria for RNase7 and α-Defensin 1-3 mixtures was observed compared with the RNase7 co-incubations (P = 0.0318) (Fig 4B). The most significant effect on damaged bacterial agglutination was recorded with α-Defensin 1-3 in combination with DMBT1 (Fig 4C). We proceeded to evaluate the activity of these AMP combinations against pyelonephritis multi-drug–resistant strain; MDR58, using respective fluorescent dyes (Fig S4A–H). Although less mean bacterial agglutination total surface area was recorded, significant cooperative effects were quantified and compared with co-incubation of the candidate AMP mixtures against pyelonephritis multi-drug–resistant strain (Fig 4D–F). The most significant difference between pyelonephritis and MDR strains is the blunting of LL-37 and α-Defensin 1-3 combinatorial agglutination, whereas DMBT1 agglutination with killing seems rescued with the addition of α-Defensin 1-3 (Fig 4D and F). Our results indicate α-Defensin 1-3 work in concert with other AMPs to neutralize E. coli strains in relation to strain-specific factors.

DEFA1A3 gene dosage associates with reduced pro-inflammatory immune cell recruitment after UPEC superinfection

Previous studies have characterized the effects of the human DEFA1A3 gene in the resident immune cell composition of the transgenic mouse urinary tract at baseline (9). We evaluated a comprehensive flow cytometry panel of leukocyte immune cell surface markers in the kidneys of the infected DEFA^4/4, DEFA^4/0, and DEFA^0/0 mice (Fig 5A–J). After UPEC superinfection challenges, we found pro-inflammatory immune cell populations to be mostly differentially recruited to the kidney between DEFA1A3-carrier and non-carrier mice (Fig 5A). The significant differences corresponded to the less immune influx of neutrophils (P = 0.004), eosinophils (P = 0.008), inflammatory macrophages (P = 0.007), and inflammatory monocytes (P = 0.015) comparing infected DEFA^4/4 and DEFA^0/0 kidneys (Fig 5C–F). Comparing DEFA^4/4 and DEFA^4/0 mice, the inflammatory macrophages were differentially recruited (P = 0.019) to the infected kidneys (Fig 5E). Our results suggest that high α-Defensin 1-3/DEFA1A3 copy number dampens pro-inflammatory macrophage recruitment during the cellular immune recruitment response to UPEC.

DEFA1A3 DNA gene dose–dependent antimicrobial effects lead to down-regulation of Toll-like receptor expression in infected kidneys

We then explored Toll-like receptor signaling as the possible innate immune mechanistic cue(s) responsible for differential pro-inflammatory responses against UPEC in DEFA1A3-expressing mice (6, 46). Extensive research has suggested α-Defensin 1-3 can inhibit bacterial cell wall synthesis and SOS repair mechanisms via interactions with different bacterial components and Toll-like receptor ligands such as cell wall precursor lipid II, RNAs, and RecA-ssDNA nucleoprotein interactions in vitro (26, 32, 47). To investigate whether these interactions confer a differential immune pathogen recognition response in vivo, we evaluated the induction of Toll-like receptor expression for E. coli cellular bacterial ligands under the UPEC superinfection model. Upon assessment of acute Toll-like receptor genes induced at 6-hpi CFT073 superinfection challenge, we demonstrate a differential significant up-regulation of Tlr2, Tlr4, Tlr6, and Tlr9 genes at the transcription level after challenges between DEFA1A3 copy number–carrier and non-carrier mice (Fig S5A–F). Although the DEFA^4/0 and DEFA^0/0 mice similarly increased Tlr2, Tlr4, and Tlr6 mRNA gene expression (P = 0.252, 0.713, and 0.997, respectively), the DEFA^4/4 mice show the diminished inducible expression of these Toll-like receptor genes (P = 0.044, 0.036, and 0.012, respectively). Similarly, Tlr9 mRNA induction was elevated in the DEFA^0/0 and DEFA^4/0 mice to a comparable extent (P = 0.995). However, infected DEFA^4/4 kidneys lacked significant induction of Tlr9 mRNA expression levels compared with DEFA^0/0 and DEFA^4/0 mice (P = 0.04 and 0.05, respectively). On the contrary, inducible Tlr7 mRNA expression trends were similar across challenged mouse groups. Overall, the results suggest high DEFA1A3 DNA copy numbers drive selective blunting of Toll-like receptor mRNA expression after UPEC invasion in the kidney.

High DEFA1A3 DNA gene dosage kidney anti-inflammatory effects are independent of UPEC antimicrobial activity

To discern whether lower urinary tract bacterial burden drives reduced kidney Toll-like receptor signaling in a bacterial viability–dependent manner, we performed UPEC superinfection challenge with α-Defensin 1-3–resistant UPEC (MDR58) and analyzed inducible gene expression trends. After MDR58 challenges, kidney DEFA^4/4 bacterial burdens between infected mice did not show differences compared with DEFA^4/0 and DEFA^0/0 mice (Fig 6A). Evaluation of Toll-like receptor downstream effector gene targets (Il1β, Ifnβ, and Il6) was further assessed to confirm acute protective immunomodulatory phenotypes in DEFA^4/4 in the setting of MDR58 infections. Superinfection-challenged kidneys showed down-regulation of Il1β gene expression in a DEFA1A3 copy number–dependent manner (Fig 6B). However, infected DEFA^4/4 kidneys showed significant Ifnβ and non-significant Il6 reduction in gene expression when compared to non-carrier mice (Fig 6C and D). We further confirmed anti-inflammatory effects are viability-dependent because of the lack of reduced Tlr2 and Tlr4 gene expression by comparing phenotypes with CFT073 and MDR58 superinfection challenges (Fig 6E and F). Similarly, a mild effect was also observed for induction of Tlr5 gene expression (Fig 6G). After MDR58 challenge, DEFA^4/4 kidneys show significantly reduced Tlr6 expression and down-regulation of Tlr9 transcript induction to baseline levels compared with littermate counterparts (Fig 6H and I). Furthermore, mice with DEFA^4/4 copies exert a protective phenotype that is both bacterial viability-dependent and -independent of the transgenic DEFA1A3 gene copies to decrease DAMP-associated inflammatory pathways. These results suggest α-Defensin 1-3 interactions with UPEC-associated Toll-like receptor ligands lead to differential Toll-like receptor signaling that results in a subsequent reduction in pro-inflammatory signaling to ameliorate both infection burden and inflammation in the kidney.

Transgenic mouse generation

A mouse genetically engineered to express DEFA1A3 was generated by Dr. Tomas Ganz when human α-Defensin 1-3/DEFA1A3 gene copies were inserted into the murine genome using a bacterial artificial chromosome (BAC-10) construct as previously described (20). Our group characterized this mouse on a C57BL/6 background using long-range DNA sequencing and determined that each mouse possessed four copies of each gene per chromosome (9, 20, 21). To generate littermate mice with various DEFA1A3 genome copies, WT C57BL/6J (Cat# 000664) mice were purchased from the Jackson Laboratory and bred in-house. Mouse breeding and experimental procedures were performed under approved IACUC protocol #20105 at the Laboratory Animal Resource Center at Indiana University School of Medicine.

PCR of transgenic DEFA mouse genotyping

∼2 mm of mouse tail biopsy was subjected to DNA extraction and amplification with KAPA HotStart Mouse Genotyping Mix, and oligonucleotide sequences DEFA1A3_set 1 to generate a WT allele band forward primer 5′-TCTCACCTGAGGTTCCTGCT-3′ and reverse 5′-CCTGATGAGCTGATTGCAGA-3′; and DEFA1A3_set 2 used as forward 5′-GATGTTCACAGCAGGGGATT-3′ and reverse 5′-CCTGATGAGCTGATTGCAGA-3′ for the transgenic allele were combined and amplified using thermocycler conditions according to the manufacturer’s protocol (Cat# 2GFHSGKB; Roche). As previously demonstrated, 513- and 729-bp PCR products were analyzed in 2.5% electroporated agarose gel based on 1-kb DNA Hyper Ladder (Cat# N3232; New England Biolabs) (9).

Murine pyelonephritis challenge experiments

E. coli pyelonephritis strain (UPEC; CFT073 or UPEC; MDR58) was grown overnight on Luria–Bertani (LB) media at 37°C in a orbital shaker at 5g. 8- to 16-wk-old anesthetized female C57BL/6J mice were infected with a 50-μl suspension containing 1 × 10⁸ CFU/ml of E. coli pyelonephritis strain in PBS or vehicle control as described (71). After 3 h of initial infection, transurethral infections were repeated for superinfection challenges (36). Prior studies have characterized the increased DEFA1A3 mRNA expression and robust UTI burden after superinfection challenges (9). Mice were euthanized by CO₂ inhalation at respective timepoints, and urinary tract organs were harvested and processed for consecutive experiments.

Quantitative real-time PCR mRNA expression assays

Total RNA from organ or cell lysates was isolated and purified using RNeasy Plus Micro Kit. Following the manufacturer’s instructions, cDNA was prepared with the high-capacity cDNA reverse transcription kit (Applied Biosystems). The list of gene targets and catalog numbers used throughout the study can be found in Table S1. Singleplex qRT–PCR protocol was performed using Luna Universal qRT-PCR Master Mix (New England Biolabs). Target gene expression cycle thresholds were normalized to the housekeeping gene and calculated into relative expression using the 2^−ΔΔCq method (73).

Magnetic-activated cell sorting of kidney leukocytes, intercalated cells (ICs), and non-intercalated cells (non-ICs)

Kidneys from mice of different conditions and genotypes were aseptically collected in ice-cold 1x sterile PBS. Using gentleMACS dissociation (Miltenyi Biotec), single-cell suspensions were digested in Accumax enzymatic solution (Cat# AM105; Innovative Cell Technology) as previously described. Cells were filtered and supplemented with 1xDMEM before centrifugation for 10 min at 300g. Red blood cell lysis buffer (Cat# 420301; BioLegend) was applied to resuspend cells and incubated for 5 min on ice. After lysis, the cell suspension was supplemented with DMEM and subjected to filtering through a 70-μm nylon mesh (Cat# 22-363-548; Thermo Fisher Scientific). Cells were washed again and resuspended in EasySep buffer (Cat# 20144; Stem Cell Technologies) for subsequent annexin V separation using EasySep Dead Cell Removal Kit (Cat# 17899; Stem Cell Technologies). To isolate kidney-derived leukocytes, anti-mouse CD45 microbeads (Cat# 130-052-301; Miltenyi Biotec) were applied and incubated, and CD45⁺ cells were flushed out with 1 ml of magnetic-activated cell sorting buffer (PBS containing 0.5% BSA and 2 mM EDTA). The process was repeated using anti-mouse CD117 microbeads (Cat# 130-097-146) on CD45⁻ cell suspensions to flush out ICs (CD45⁻/CD117⁺). Non-ICs were recovered from CD117⁻ cells (9, 59).

Enrichment of PMNs and flow cytometry studies

Whole-kidney single-cell suspensions were prepared as described above. Briefly, washed filtered cells were centrifuged without deceleration in 33% Percoll gradient (Cat# P1644; Millipore Sigma) for 20 min at 300g under 4°C. One million total PMNs were then subjected to the Fc binding blockade with anti-CD16/32 (Cat# 553142; BD Biosciences) for 15 min. After non-specific binding blocking, the mouse antibodies further described in Table S2 were incubated on ice for 30 min. Fixed viable cells were counted using Live/Dead cell dye and analyzed in the Attune NxT flow cytometer (Thermo Fisher Scientific). Supplemental gating strategy analysis of immune cells was applied as previously characterized (74, 75). The relative percent (%) change of immune cell populations in infected kidneys was calculated by dividing the mean counts from gated cell populations of two genotype groups and multiplication of the ratio by 100%.

Immunofluorescence microscopy

Tissues were dissected and fixed in 4% PFA overnight. After fixation, tissue blocks were prepared for histology in 70% ethanol. Slides were deparaffinized using Pro-Par Clearant (Cat# NC9537734; Anatech Ltd) three times in total before switching to a different container every 3 min. After rehydration and 1xPBS wash, slides were permeabilized at RT. Antigen retrieval was applied and washed again. Superblock (Cat# 37537; Thermo Fisher Scientific) was pipetted on the tissue for 5 min to block non-specific binding. Primary antibodies chicken anti-V-ATPase-E1 (Cat# GW22284F; Sigma-Aldrich), rabbit anti-V-ATPase-G3 (Cat# AB1220012; Abcam), and α-Defensin 1-3 (Cat# N5377-03P; US Biologicals) at 15 μg/ml were then added to the tissues and incubated at 4°C overnight. The next day, slides are washed off to remove residual antibodies. Secondary Alexa Fluor 488 anti-chicken, Alexa Fluor 488 anti-rabbit, and Cyanine3 anti-goat antibodies were respectively added and incubated in the dark for 30 min at RT at 1:600 dilution. After washing excess reagents, slides were mounted with Vectashield mounting solution with DAPI (Cat# H-1200-10; Vector Laboratories) and allowed to harden in the dark for at least 20 min before imaging in a Keyence BZ9000 microscope (Keyence Corporation).

Organ bacterial colony-forming unit analysis

Mouse tissues were homogenized and serially diluted in a 96-well microplate containing sterile 1xPBS. Samples were added to LB agar plates and incubated statically at 37°C overnight for CFU quantification as previously demonstrated (9).

Time/dose–response-kill UPEC assays

E. coli strains (K12; SMG123, UPEC; CFT073, and UPEC; MDR58) were cultured in LB media at 37°C with 5g orbital shaker conditions overnight. Mid-log bacterial suspensions (OD₆₀₀ = 0.1) were further grown for the consecutive kinetic experiments. Sub-bacteriostatic concentrations of α-Defensin 1-3 (Cat# HC4014; Hycult Biotech) were added to each prepared 1–2 × 10⁶ CFU/ml E. coli inoculum in sterile 1xPBS (1, 10, 100 μg/ml). After OD₆₀₀ reading at baseline, the cultures were further incubated and aliquoted for 1-, 3-, and 24-h post-inoculation readings in a round-bottom 96-well microplate (Cat# 3799; Corning). Control wells and testing conditions were performed in technical duplicates.

Bacterial agglutination assays

E. coli strains under various conditions were visualized for viability under the microscope using Live/Dead BacLight Bacterial Viability Kit (Cat# L7012; Thermo Fisher Scientific). Human AMPs: cathelicidin peptide, LL-37 (Cat# AB140725; Abcam), Ribonuclease 7, RNase7, (Cat# AB164629), and recombinant Deleted in Malignant Brain Tumors 1, DMBT1 (kindly provided by Dr. Jan Mollenhauer), were added to media containing 1–2 × 10⁶ CFU/ml of mid-log bacterial suspensions, incubated, and imaged in a 96-well round-bottom microplate (Cat# 3799; Corning). An equal amount of SYTO9 and propidium iodide (PI) viability kit dyes was applied to each well separately and imaged under the EVOS microscope at a 20X objective fluorescent lens (Thermo Fisher Scientific). Quantification of surface area (μm²) for the largest 20 clumps was performed and compared across conditions tested in imaged wells using Celleste Imaging Image Software (Thermo Fisher Scientific).

Figure design and statistical software applications

Data visualization in figures plotted was designed using GraphPad Prism 10 (GraphPad Software). Measured probability values at <0.05*, <0.01**, and <0.001*** represent significant difference symbols between groups were determined by a specified statistical test described in the figure legend. Supplemental scientific abstract Fig S6 was created with BioRender.com (license agreement #EW26LANXWC, Toronto, Canada).