Endothelial SOCS3 maintains homeostasis and promotes survival in endotoxemic mice

SOCS3 is the main inhibitor of the JAK/STAT3 pathway. This pathway is activated by interleukin 6 (IL-6), a major mediator of the cytokine storm during shock. To determine its role in the vascular response to shock, we challenged mice lacking SOCS3 in the adult endothelium (SOCS3iEKO) with a nonlethal dose of lipopolysaccharide (LPS). SOCS3iEKO mice died 16–24 hours postinjection after severe kidney failure. Loss of SOCS3 led to an LPS-induced type I IFN–like program and high expression of prothrombotic and proadhesive genes. Consistently, we observed intraluminal leukocyte adhesion and neutrophil extracellular trap–osis (NETosis), as well as retinal venular leukoembolization. Notably, heterozygous mice displayed an intermediate phenotype, suggesting a gene dose effect. In vitro studies were performed to study the role of SOCS3 protein levels in the regulation of the inflammatory response. In human umbilical vein endothelial cells, pulse-chase experiments showed that SOCS3 protein had a half-life less than 20 minutes. Inhibition of SOCS3 ubiquitination and proteasomal degradation led to protein accumulation and a stronger inhibition of IL-6 signaling and barrier function loss. Together, our data demonstrate that the regulation of SOCS3 protein levels is critical to inhibit IL-6–mediated endotheliopathy during shock and provide a promising therapeutic avenue to prevent multiorgan dysfunction through stabilization of endothelial SOCS3.


Flow cytometry
Mice were euthanized with an overdose of pentobarbital. The chest cavity was opened, and 100 µl of blood were collected from a nicked aorta into EDTA-containing vials. Lungs were excised, washed several times with PBS, and minced finely with scissors. Then, the minced lungs were transferred to a 50ml conical tube containing 6 ml of pre-warmed collagenase type I/dispase II/DNAse I mix (CDD) and incubated at 37 ºC with gentle agitation for 1 hour. Single cell suspensions were obtained after the suspensions were passaged multiple times through 20gauge cannulas and filtered through a 70 µm mesh. Cells were then spun at 400 g for 8 minutes at 4 ºC and resuspended in 3 ml of cold PBS containing 0.1% BSA.
In parallel, bone marrow cells were collected by flushing the right femur with 10ml PBS onto a 70 µm cell strainer and then centrifuged at 6000 rpm for 5 min 4ºC. Red blood cells from blood and bone marrow cells were lysed using RBC lysis buffer according to manufacturer's instructions. Live and dead cells from lungs, blood and bone marrow were distinguished using Pacific blue Annexin V Apoptosis Detection kit with 7-AAD. Lung single cell suspensions (300 µl / sample) were then stained with ly6G-APC-Cy7 to label neutrophils. Cells were then subjected to flow cytometry analysis using a BD FACS symphony. 30,000 events were acquired for each sample and the data was analyzed using FlowJo v10.7.1. Single cells were gated using FSC-H vs FSC-A, dead cells were excluded based on 7AAD-Annexin-V and only live cells were gated for further analysis. Results are expressed as percentage of total cells.

Isolation of RNA from spleen CD45+ cells
Control, SOCS3iEKO or heterozygous mice were treated with tamoxifen to induce Cre activation as described in the main text. Two weeks after, mice were sacrificed to collect spleens. Organs were then washed in PBS. One third of each spleen was transferred into a tube with Trizol to have RNA of total organ isolated later following the Trizol protocol. The remaining portion of each spleen was pressed with a plunger from a 10 cc syringe. Released cells and tissue remains were collected and pushed through 70 μm cell strainers to obtain single cells isolates. Cells were centrifuged at 400g for 8 min at 4 ºC. The cell pellet was then resuspended in 1.5 ml ice-cold PBS+0.1%BSA. The cell suspension was incubated with anti-CD45 conjugated Dynabeads M280 streptavidin beads (3 μg ab,300 μg beads/sample) for 10 minutes at RT under gentle rotation. CD45+ cells were separated using a magnetic separator (Dynal MPC-S) and washed six times with PBS+0.1%BSA. Cells were placed on magnet for 3 minutes, the supernatant was removed and cells were lysed in Trizol reagent. RNA was extracted following a modified protocol. Briefly, 0.1 ml chloroform was added to 0.5 ml of Trizol per sample. After a 2 min incubation, tubes were centrifuged for 15 min at 12,000 g at 4 ºC. The Supplemental Figure 1. Enrichment of CD45 cells from spleen. Spleens from control, heterozygous or SOCS3iEKO mice (corresponding to those shown in Figure 1C) were collected 14 days after tamoxifen treatment. CD45 cells were isolated using magnetic beads prior to RNA extraction and RT-qPCR. Enrichment was confirmed by a 5-10-fold increase in CD45 expression and >10-fold decrease in VWF expression compared to total RNA from the same spleens. NS: not statistically significant (Kruskal-Wallis). Data combined from three independent experiments.
Supplemental Figure 2. Mild liver dysfunction in endotoxemic SOCS3 iEKO . A, H&E staining showing a reduction in glycogen content but no other overt histological features in endotoxemic SOCS3 iEKO mice. B, Plasma albumin levels remained similar among all experimental groups (two-way ANOVA). Data combined from at least three independent experiments. Supplemental Figure 3. Gene expression changes in HUVEC treated with IL-6+R. A, IL-6+R-mediated loss of barrier function requires continuous protein synthesis. B, RT-qPCR of IL-6+R-treated HUVEC. Data was obtained as described in Figure 7. Data combined from three independent experiments. C, RT-qPCR of cells treated with or without IL-6+R after infection with lentivirus to overexpress WT SOCS3 or K6Q-SOCS3. An empty vector lentivirus was used as control. Two-way ANOVA. Combined data from three independent experiments performed in duplicate each.
Supplemental Figure 4. Multiple significant parameters correlate with the severity score and plasma IL-6 levels. A, Cross-correlation analysis for all parameters measured in LPStreated mice (full data available as Supplemental Table 6). B, Graph showing all significantly correlated parameters to the severity score or plasma IL-6. C, Examples of the correlation of expression of specific genes (expressed as a base 2 logarithm of fold changes) with the severity score for each mouse. Data combined from three independent experiments. Supplemental Figure 5. Type I IFN-like, adhesive and prothrombotic gene expression in endotoxemic SOCS3 iEKO mice. RT-qPCR analysis of whole lung (A) or liver (B) RNA levels. Two-way ANOVA and Holm-Sidak post-hoc tests comparing het and SOCS3 iEKO mice to control within saline or LPS-treated groups. Asterisks denote p < 0.05. Data combined from at least three independent experiments. Supplemental Figure 6. Neutrophil and monocyte infiltration in endotoxemic organs. A, flow cytometry analysis of Ly6G+ cells in lungs. Asterisks denote p < 0.05. Data from two independent experiments (n=7 saline-and n=9 LPS-treated mice). B, Immunohistochemical staining of FFPE sections for myeloperoxidase (MPO, a neutrophil marker) and F4/80 (a monocyte marker) in lungs and kidneys. Bars, 50 μm. Images are representative from three independent experiments.  Table 2 Gene symbol Forward