Fig 2: LIF inhibits BAE cell growth through the JAK-STAT3 pathway ARecombinant human LIF inhibited growth of BAE cells in a dose-dependent manner. BAE cells were cultured in the presence of vehicle and indicated concentrations of recombinant human LIF (rhLIF). Cell proliferation was analyzed after 6 days, n = 3.BJAK inhibitor baricitinib blocked activation of STAT3 by LIF. BAE cells preincubated with DMSO and inhibitors for 1 h were treated with vehicle and LIF (10 ng/ml) for 15 min. Whole-cell lysates were subjected to Western blotting with indicated antibodies. Ctrl, no preincubation with inhibitors; Ba, baricitinib (2 µM); Co, cobimetinib (150 nM); BE, BEZ235 (5 nM).CThe JAK inhibitor baricitinib reversed LIF-induced BAE growth inhibition. BAE cells preincubated with inhibitors for 1 h were treated with vehicle and LIF (10 ng/ml, Sigma). Cell proliferation was analyzed after 6 days using alamar blue, n = 3.D, EKnockdown of STAT3 in BAE cells. BAE cells were transfected with siRNAs targeting STAT3. qRT-PCR was performed to examine STAT3 mRNA levels. STAT3 level in siNegative was set as 1. Data from three independent experiments were averaged and shown in D. In E, cells transfected with siRNAs were treated with LIF (10 ng/ml, Sigma) and vehicle for 15 min. Whole-cell lysates were subjected to Western blotting with indicated antibodies.FLIF-induced BAE cell growth inhibition was abolished by knockdown of STAT3. BAE cells with STAT3 knockdown were cultured with LIF (10 ng/ml, Sigma) and vehicle. Cell proliferation was analyzed after 3 days. Fluorescence reading for each vehicle group was set as 1, n = 3. Data information: Bars and error bars represent mean ± SD. All experiments were carried out in three independent studies. siNegative, negative control siRNA not targeting any known genes. Two-way ANOVA was used as statistical test. Source data are available online for this figure.

Fig 3: Identification of LIF as the EC mitogen in LN-229-conditioned medium (CM) LN-229 CM stimulated growth of BCE cells, n = 3.VEGF neutralizing antibody (B20) failed to suppress BCE cell growth induced by LN-229 CM, n = 3.Reverse-phase fractions of LN-229 CM induced BCE cell growth. BCE cells were incubated with fractions (2 µl/well) as indicated in the figure, n = 3.Candidate proteins generated from mass-spectrometry analysis of LN-229 CM reverse-phase fractions. Candidates were identified by excluding intracellular proteins and proteins showing higher abundance in inactive fractions compared to those in mitogenic factions. Proteins were ranked for relative abundance as described in Materials and Methods.The anti-LIF neutralizing antibody abolished BCE cell growth induced by reverse-phase fractions, n = 3.Recombinant human LIF proteins stimulated growth of BCE cells in a dose-dependent manner. BCE cells were cultured in the presence of vehicle, VEGF (10 ng/ml), and the indicated concentrations of recombinant human LIF (rhLIF, Sigma), n = 3.LIF and VEGF synergistically stimulated BCE cell growth. Cell proliferation was analyzed after 6 days using alamar blue as described in Materials and Methods, n = 3. Data information: Bars and error bars represent mean ± SD. All experiments were carried out in three independent studies. Two-way ANOVA was used as statistical test. ns, not statistically significant. Source data are available online for this figure.

Fig 4: LIFR, but not LIF, is highly expressed in human choroidal ECs by single-cell transcriptomic analysis A–EViolin plots of the log scale normalized expression of LIF receptor subunits, IL6ST (A), LIFR (B), LIF (C), IL6R (D), and KDR (E) in all cell types in the human choroid (based on Voigt et al, 2019). Note the wide distribution of IL6ST (gp130), the common subunit to all receptors of the IL-6 family (A). In contrast, LIFR is strongly expressed in ECs and to a lesser extent in fibroblasts and pericytes (B). LIF is expressed in fibroblasts, macrophages/monocytes, and pericytes and minimally in ECs (C). The specific IL-6 receptor (IL6R) is highly expressed in macrophage/monocytes (D), while LIFR has little expression in these cells (B). KDR shows, as expected, high expression selectively in EC (E).FCell clusters in uniform manifold approximation and projection (UMAP) space.

Fig 5: LIF promotes angiogenesis in in vivo models A, BIntravitreal injection of LIF increases blood vessel density in the mouse eyes. Adult mice were intravitreally injected with VEGF (10 ng) or LIF (10–100 ng, Sigma). Seven days after injection, PFA-fixed choroid–sclera complexes and retina were subjected to CD31 IF. Representative images of CD31-positive vessels are shown in A. Scale bar = 100 µm. Vascular density, determined with ImageJ software, is shown in B), n = 5–8.C, DOCT-A imaging of LIF-treated mouse retina. Adult mice were intravitreally injected with 1 µl of LIF (50 ng, Sigma) or vehicle solution (PBS). Retinal OCT-A images were obtained 7 days after the injection and representative images are shown. Blood vessel density was determined as percentage of vessel-covered area/total area surface using ImageJ software and shown in D), n = 7–8.ESTAT3 was phosphorylated in retinal ECs after LIF treatment. Fifty nanogram of our in-house LIF with low endotoxin levels was injected. Phospho-Stat3 was examined on cryosection of retina after 2 h by IHC. Magnification in the dashed box is shown in the lower panel. n = 5. Scale bar = 100 µm.FFive-day-old neonatal mice were intravitreally injected with LIF (50 ng) or vehicle solution (PBS). After 3 days, retinas were subjected to IF staining with Dyight-488-labeled lectin. Representative images for similar ocular loci and quantification of lectin-labeled area using ImageJ software were shown, n = 4. Scale bar = 100 µm. Data information: Bars and error bars represent mean ± SEM. All experiments were carried out in three independent studies. Two-way ANOVA was used as statistical test. Source data are available online for this figure.