Circulating extracellular vesicles in human cardiorenal syndrome promote renal injury in a kidney-on-chip system

BACKGROUND Cardiorenal syndrome (CRS) — renal injury during heart failure (HF) — is linked to high morbidity. Whether circulating extracellular vesicles (EVs) and their RNA cargo directly impact its pathogenesis remains unclear. METHODS We investigated the role of circulating EVs from patients with CRS on renal epithelial/endothelial cells using a microfluidic kidney-on-chip (KOC) model. The small RNA cargo of circulating EVs was regressed against serum creatinine to prioritize subsets of functionally relevant EV-miRNAs and their mRNA targets investigated using in silico pathway analysis, human genetics, and interrogation of expression in the KOC model and in renal tissue. The functional effects of EV-RNAs on kidney epithelial cells were experimentally validated. RESULTS Renal epithelial and endothelial cells in the KOC model exhibited uptake of EVs from patients with HF. HF-CRS EVs led to higher expression of renal injury markers (IL18, LCN2, HAVCR1) relative to non-CRS EVs. A total of 15 EV-miRNAs were associated with creatinine, targeting 1,143 gene targets specifying pathways relevant to renal injury, including TGF-β and AMPK signaling. We observed directionally consistent changes in the expression of TGF-β pathway members (BMP6, FST, TIMP3) in the KOC model exposed to CRS EVs, which were validated in epithelial cells treated with corresponding inhibitors and mimics of miRNAs. A similar trend was observed in renal tissue with kidney injury. Mendelian randomization suggested a role for FST in renal function. CONCLUSION Plasma EVs in patients with CRS elicit adverse transcriptional and phenotypic responses in a KOC model by regulating biologically relevant pathways, suggesting a role for EVs in CRS. TRIAL REGISTRATION ClinicalTrials.gov NCT03345446. FUNDING American Heart Association (AHA) (SFRN16SFRN31280008); National Heart, Lung, and Blood Institute (1R35HL150807-01); National Center for Advancing Translational Sciences (UH3 TR002878); and AHA (23CDA1045944)


Supplemental
) Box and whisker plots showing differential expression of genes in the "Injury" and "No Injury" groups.Box plots represent the first quartile, median, and third quartile, with whiskers indicating minimum and maximum values.Supplemental methods RNAse A Treatment Isolated EVs (210 μL) were incubated with RNAse A at a 0.5µg/µL final concentration to degrade any extracellular RNAs not protected within EVs (Thermo Fischer Scientific, Waltham, MA, USA) for 20 minutes at 37 °C followed by an addition of RNAse inhibitor (Thermo Fischer Scientific) at a 20U/µL concentration to inactivate RNAse A prior to RNA extraction.

Microfluidic resistive pulse sensing (MRPS)
EVs were diluted at 1:100 to prevent saturation of the upper limit of detection or aggregation, subjected to MRPS using the Spectradyne's nCS1 (Spectradyne,Signal Hill, CA, USA), and analyzed with both high and low sensitivity settings (NP100; voltage, 0.60 V; stretch, 46.0 mm and NP400; voltage, 0.40 V; stretch, 43.5 mm respectively).The pressure was preset at 7.0 mbar.Minimum2000 particles were analyzed for each sample.

Transmission electron microscopy (TEM) of plasma EVs
A drop containing 5 μL of purified plasma EVs (for both c-DGUC and SEC) was placed on parafilm, and a carbon-coated copper grid was placed on top of the drop for 30 minutes.Carbon-coated copper grids were previously glow discharged for 30 seconds to turn into an overall hydrophilic surface.For immune gold labeling, grids were blocked with 1% BSA in 1X PBS for 10 minutes, and IgG primary antibody anti-CD81 (Santa Cruz Biotechnology, Dallas, TX, USA; 1:30 dilution in blocking reagent) was used for 30 minutes at room temperature.Grids were then washed three times in 1X PBS and incubated with Protein A, conjugated with 10 nm gold particle (1:30 dilution) for 20 minutes at room temperature.Grids were washed twice with 1XPBS (for 5 minutes total) and 4 times with water (for 10 minutes total).Grids were stained with 0.75% uranyl formate for 1 minute and visualized on a JEOL 1400 electron microscope outfitted with an Orius SC1000 CCD camera (Gatan, Inc.Pleasanton, CA, USA).Immunofluorescence analysis Epithelial cells, grown on coverslips were treated with EVs from either Healthy Control or HFpEFCRS patients and then stained with the primary antibodies (NGAL, IL-18 and KIM-1) followed by incubation with Alexa Fluor 555-labeled secondary antibodies (Molecular Probes) (Supplemental Table 5).After the slides were mounted with Vectashield (with DAPI [4′,6-diamidino-2-phenylindole]) (Vector Laboratories, Newark, CA), the slides were examined under a fluorescence microscope (BioRad).

Long RNA sequencing of plasma EV samples
Long RNA sequencing on the exRNAs isolated from the plasma of HFpEF patients with high creatinine (n = 9, 1.1-1.6 mg/dL) and low creatinine (n=9, 0.7-0.9mg/dL) were performed using the miRNeasy Serum/Plasma kit (Qiagen).cDNA libraries were constructed using the SMARTerStranded Total RNA-Seq Kit v2 Pico Input Mammalian (Takara Bio, San Jose, CA, USA) and sequenced using NextSeq 2000 platform.

Deconvolution analysis for the identification of source organs:
A similar approach to identify the source tissue has been previously (1) performed to generate the log transformed gene expression values for each sample in the Creatinine high and low groups, the DESeq2 ( https://academic.oup.com/database/article/doi/10.1093/database/baz046/5427041?login=false).This The median tissue enrichment score for each tissue and sample was then calculated from these 1000 repeats.process of tissue specific gene sampling and tissue enrichment score calculation was repeated 1000 times.
transformed gene expression values and tissue specific genes, following the algorithm of PangoDB was selected for each tissue.The tissue enrichment score for each sample was calculated using the log genes (corresponding to the bottom 25% percentile of the number of tissue-specific genes across tissues) website(https://pubmed.ncbi.nlm.nih.gov/25613900/), and a random sample of up to 172 tissue-specific Tissue-specific genes were obtained from the protein atlas (https://genomebiology.biomedcentral.com/articles/10.1186/s13059-014-0550-8)rlog function was used.

1 :
Immunofluorescence study of kidney injury markers: Immunofluorescence 58 study showing increased expression of HAVCR1, LCN2 and IL18 in the epithelial cells of HFPEFCRS groupcompared to Healthy Control group (Magnification= 100µm).Representative images of three independent experiments conducted.61 Supplemental Figure 2: (A-N) Box and whisker plots showing significant higher expression (reads per million) of different miRNAs in the EVs of the HFpEFCRS group compared to the HFpEFCRS.Box plots represent the first quartile, median, and third quartile, with whiskers indicating minimum and maximum values.Results were analyzed with unpaired t test and expressed as ± SEM ( n = 4 -5 ).ns, non significant; **, p < 0.01; ***, p < 0.001.68 Supplemental Figure 3. MiRNA cocktail 1 and cocktail 2 attenuate or mimic the effects of HFpEFCRS EVs: (A) CST3 expression was significantly decreased in the "HFpEFCRS+miRNAs cocktail 1 treated group" compared to "HFpEFCRS+Control cocktail 1 treated group"; n = 3-4 for each group.(B) CST3 expression was markedly increased in the "Healthy Control+MiRNAs cocktail 2 treated group" compared to "Healthy Control+Control cocktail 2 treated group".; n = 3 for each group.GAPDH was used as internal loading control.Results were analyzed by unpaired t test and expressed as ± SEM of three independent experiments.**, p < 0.01.78 Supplemental Figure 4: (A) Volcano plot was created by all differentially expressed genes.Y axis shows Log10 FDR and X axis displays the log2-fold change value.The red dots represent the differentially expressed genes with FDR adjusted p value ≤ 0.05 and absolute fold change ≥ 1.5, while green dots represent non significantly modulated genes.