Interleukin-19 is a Novel Lymphangiocrine Factor Inducing Lymphangiogenesis and Lymphatic Junctional Regulation

Abstract

Background:

The lymphatic system functions by removing fluid, macromolecules, and immune cells to maintain tissue homeostasis. The structural organization of junctional protein complexes is vital to lymphatic function where initial lymphatics have permeable button junctions and collecting lymphatics have relatively impermeable zipper junctions. During inflammation, this junctional morphology appears to reverse, contributing to overall lymphatic malfunction. Little is known about the effects of immuno-modulatory cytokines on lymphatic vessel formation and function during inflammation. The purpose of this study is to test the hypothesis that interleukin-19 (IL-19) promotes lymphangiogenesis and proper lymphatic function during inflammation.

Methods:

We used cultured human dermal lymphatic endothelial cells (hdLECs) to determine IL-19 expression and its effects on lymphangiogenesis assays. Immunocytochemistry and electric cell-substrate impedance sensing (ECIS) determined effects on junctional morphology as it relates to permeability in vitro. RNA sequencing determined the effects of IL-19 on gene expression. Il19^−/− Ldlr^−/− double knockout (dKO) mice were used to determine IL-19 effects on lymphatic function and lymphatic vessel visualization in vivo.

Results:

Endogenous IL-19 expression is induced by exogenous IL-19 and vascular endothelial growth factor C (VEGFC) stimulation. IL-19 is lymphangiogenic, increasing hdLEC migration, network formation, and proliferation. IL-19 induces expression of transcription factors and permeability-associated genes. IL-19 induces rapid VE-cadherin phosphorylation, increases permeability of hdLEC monolayers, and mitigates oxidized low-density lipoprotein (oxLDL)-associated decrease in hdLEC permeability. In vivo, dKO mice on a high fat diet have impaired lymphatic drainage, decreased lymphatic branch points, and increased percentage of zippered junctions compared to control mice.

Conclusions:

Taken together, these are the first data to show that IL-19 has potent effects on lymphatic vessel formation and function in vitro, and that IL-19 regulates lymphatic drainage in vivo. IL-19 may represent an immuno-modulatory cytokine with therapeutic potential for improving impaired lymphatic function consequent to inflammation.

Introduction

In contrast to the blood vasculature, the lymphatic vasculature is a unidirectional system which functions by maintaining fluid, macromolecule, and white blood cell homeostasis through lymph transport ¹. The lymphatic system begins as blunt-ended initial lymphatics which contain button-like junctions allowing for passive uptake of lymph ^1–3. Initial lymphatics then converge to form collecting lymphatics which maintain lymph within the system by relatively impermeable zipper junctions ^1–4. To sustain their distinct functions, the two segments of lymphatic vessels maintain their unique structural morphology of junctional proteins, namely vascular endothelial cadherin (VE-cadherin), located between lymphatic endothelial cells ^3–6. The differential regulation of buttons in initial lymphatics and zippers in collecting lymphatics is known to be regulated by several factors, the most studied factor being Angiopoeitin-2 (Angpt2) ^7–11. This structure-function relationship is important in disease states, as lymphatic vessels undergo pronounced morphological changes in inflamed tissue, leading to reduced functionality ¹². During inflammation, the junctional morphology appears to reverse wherein initial lymphatics become zippered and collecting lymphatics become buttoned, contributing to overall lymphatic malfunction ^3,5,13–18. This leads to poor lymphatic drainage of pro-inflammatory molecules and less egression of leukocytes, exacerbating inflammation and preventing resolution ^13,19. During atherosclerotic plaque progression, inflammatory cells progressively accumulate in the adventitia, paralleled by an increased presence of leaky vasa vasorum ²⁰. However, our understanding of the role of adventitial lymphatic vessels during atherosclerotic plaque progression and regression remains controversial, requiring additional study. Studies in both mice and humans suggest that the size and number of adventitial lymphatic vessels increase as atherosclerotic plaques advance in severity ^21–27. Conversely, other studies suggest minimal lymphatic vessels present in the adventitia of both healthy and atherosclerotic human coronary arteries ^28,29. Further, it has been shown that arteries already containing a substantial network of lymphatic vessels seem naturally protected against atherosclerosis when compared with those lacking such a network ²⁴. Despite this discordance, recent evidence seems to suggest that lymphatic vessel malfunction contributes to atherosclerotic plaque formation ^27,30–38. For example, studies in atherosclerosis-prone low-density lipoprotein receptor knockout (Ldlr^−/−) mice suggest that improving lymphatic transport limits plaque formation and macrophage accumulation by mediating reverse cholesterol transport (RCT) ^30,31.

While it is clear that lymphatic vasculature has the potential to dampen atherogenesis through modulation of immune responses and RCT, little attention has been given to the effects of localized vascular inflammatory factors in driving lymphangiogenesis and lymphatic function. As suggested by the Janus phenomenon proposed by Epstein, it is generally accepted that pro-inflammatory factors are pro-angiogenic, whereas anti-inflammatory factors and therapies are anti-angiogenic ^39,40. In this context, a factor that is an exception to the Janus phenomenon could be considered a therapeutic option because such a factor would reduce local inflammation while having positive effects on the localized lymphatic network in terms of expansion (lymphangiogenesis) and function (maintaining appropriate junctional morphology to allow for the removal of fluid and macromolecules) ⁴¹.

Interleukin-19 (IL-19) is an immuno-modulatory cytokine and a member of an IL-10 sub-family containing IL-20, IL-22, and IL-24, but it is distinct from these sub-family members in terms of cell-specific expression and function ⁴². IL-19 has direct anti-inflammatory effects on resident vascular cells (endothelial and vascular smooth muscle), mainly through post-transcriptional mechanisms by decreasing stability of pro-inflammatory mRNA transcripts ^43–51. In direct contrast with the Janus phenomenon, IL-19 also has angiogenic effects on endothelial cells ^{41,43,47,52,53}. Paradoxically, in addition to being pro-angiogenic, IL-19 is also anti-atherosclerotic ^50,54,55. However, IL-19’s effects on lymphangiogenesis and lymphatic function, particularly through junctional regulation, have never been reported, leaving major gaps in our knowledge. Current therapeutics for atherosclerosis focus on lowering cholesterol levels but fail to mediate inflammation or promote lymphatic function.

In this study, we test the hypothesis that IL-19 can induce lymphangiogenesis and increase the function of lymphatics through junctional regulation and promoting appropriate levels of permeability. We show that IL-19 and its receptor complex are expressed in cultured human hdLECs and induce a lymphangiogenic transcriptional program epitomized by Prospero homeobox 1 (Prox1) upregulation and subsequent proliferation, migration, and lymphatic network formation. IL-19 appears to have both acute and chronic effects on hdLEC permeability and reduces lymphatic impermeability caused by oxLDL. In vivo, IL-19 deletion is characterized by reduced lymphatic function, decreased lymphatic branching, and increased junctional zippering. Taken together, these are the first data to show that IL-19 is expressed in and has potent lymphangiogenic and functional effects on hdLECs. IL-19 might be considered as a potential therapeutic for inflammatory conditions in which lymphatic function is in deficit.

Materials and Methods

All data and materials have been made publicly available at the TUScholarshare and can be accessed at https://scholarshare.temple.edu/handle/20.500.12613/11017. Additional information supporting the findings of this study are available from the corresponding author upon request.

Cells and Culture.

Primary human dermal lymphatic endothelial cells (hdLECs) were obtained as cryopreserved secondary culture from Cell Biologics (Chicago, Illinois, USA) Lot 030414F14 (Male, Unspecified Age “Adult,” Back Dermis), and PromoCell (Heidelberg, Germany) Lots 467Z001.2 (Male, Age 2, Foreskin), 481Z001.2 (Female, Age 18, Labia), and 483Z001.2 (Male, Age 6, Foreskin). Cells were maintained in their respective culture media (Cell Biologics, Cat. H1168; PromoCell, Media Cat. C-22221, Endothelial Cell Supplements Cat. C-39221) within passages 2–6 for all experiments, as recommended. For all subsequent experiments, cells were treated with sterile Phosphate Buffered Saline (PBS) (Corning, Cat. 21–040-CV), recombinant human IL-19 (rhIL-19) at concentrations of 100ng/mL or 200ng/mL (R&D Systems, Cat. 1035-IL), recombinant human vascular endothelial growth factor C (rhVEGFC) at a concentration of 100ng/mL (Peprotech, Cat. 100–20CD-20UG), human medium oxLDL at concentrations of 20μg/mL or 100μg/mL (Kalen Biomedical, Cat. 770202–4). Cells were transfected with 1.25μM of Prox1, STAT3, KLF4, IL-19, FOXO1, or ON-TARGETplus siCONTROL non-targeting (Scram) siRNA from Dharmacon (Lafayette, CO, USA). Amaxa HCAEC Nucleofactor Kit was used per manufacturer’s protocol from Lonza (Basel, Switzerland).

Cell Proliferation Assays.

For proliferation by cell number, cells were seeded at 10,000 per well in 12-well trays. Cells were starved for 24 hours with 1% fetal bovine serum (FBS) (HyClone, Cat. SH30396.03) in basal media, treated, and counted after 6 days using Nexcelom T4 Cellometer. For BrdU, cells were seeded at 5,000 per well in 96-well trays in 0.05% FBS for overnight starvation. Media was removed and new 0.05% FBS media with respective treatments was added for 24 additional hours. BrdU was added for the final 4 hours and all steps of the BrdU cell proliferation protocol was followed per BrdU Cell Proliferation Assay Kit (Cell Signaling Technology, Cat. 6813). Absorbance was read at 450nm on VersaMax from Molecular Devices (San Jose, CA, USA). For STAT3 inhibition, cells were treated with 1μM WP1066 (Santa Cruz, Cat. sc-203282) or an equivalent volume of DMSO (Sigma, Cat. D2650) as a control ^47,54. Each data point represents one biological replicate, determined by an average of three technical replicates within the experiment. Experiments were performed using three independent biological replicates.

Cell Migration Assays.

For scratch assay, cells were seeded at 30,000 per chamber in chamber slides and grown to 80% confluency before starving (0.5% FBS) for 24 hours and scratched with a 2mm-apparatus. Starvation media was replaced after scratching the monolayer and cells were treated. Images were taken using ThermoFisher Scientific EVOS M5000 at 0 hours and 20 hours and percent change in scratch area over 20 hours was calculated using ImageJ Software. For transwell migration assay, cells were seeded at 50,000 in the upper chamber of transwells (Millicell, Cat. PI8P01250) while treatments were added to the bottom chamber. After 24 hours, cells were washed with PBS and removed of the bottom of the transwell using Trypsin (Corning, Cat. 25–052-CI) and counted after using Nexcelom T4 Cellometer. Each data point represents one biological replicate, determined by an average of three technical replicates within the experiment. Experiments were performed using three independent biological replicates.

Lymphatic Network Assay (Tube and Loop Formation).

Cells were starved (1% FBS) and pre-treated before seeding at 30,000 per well in a Matrigel-coated 24-well tray in triplicate and re-treated. Images were captured using Olympus CKX41 at 6 hours from three areas in each well and counted for number of tubes (straight lines) and loops (when multiple tubes form a circle). Each data point represents one biological replicate, determined by an average of three technical replicates within the experiment. Experiments were performed using three independent biological replicates.

RNA Sequencing (RNAseq), RNA Extraction, and Quantitative Real Time PCR (qRT-PCR).

RNA from cultured cells was collected with Trizol (Invitrogen, Cat. 15596018) and isolated using chloroform-isopropanol extraction. The quality of the RNA was evaluated via NanoDrop 260/280 ratio. As recommended, we used RNA samples with values between 1.8 and 2.0, indicating minimal contamination. RNA samples at 0 hours, and 24 hours were sent to GENEWIZ Azenta Life Sciences (Beijing, China) at 2.2μg. The number of reads averaged 76,733,329.5 reads per sample with an average yield of 22,570 Mbases per sample. Sequence reads were trimmed to remove possible adapter sequences and nucleotides with poor quality using Trimmomatic v.0.36. The trimmed reads were mapped to the Homo sapiens GRCh38 reference genome available on ENSEMBL using the STAR aligner v.2.5.2b. Unique gene hit counts were calculated by using feature Counts from the Subread package v.1.5.2. downstream differential expression analysis. Using DESeq2, a comparison of gene expression between the defined groups of samples was performed. The Wald test was used to generate p-values and log2 fold changes. Genes with an adjusted p-value < 0.05 and absolute log2 fold change > 1 were called as differentially expressed genes for each comparison. Gene ontology analysis was performed on the statistically significant set of genes using the software GeneSCF v.1.1-p2. To minimize batch effect, experiments were performed utilizing strict protocols with the consistent use of the same equipment and lots of reagents. GENEWIZ also ensures that samples are processed together through extraction, library prep and sequencing to minimize batch effects due to processing. For qRT-PCR, RNA was subsequently reverse transcribed into cDNA using random primers (Promega), dNTP (BioBasic), DDT, SSIV Buffer, RNase inhibitor, and SuperScript IV reverse transcriptase (Invitrogen). Target genes were then amplified using AppliedBiosystems QuantStudio 3 for qRT-PCR. Multiple mRNAs (Ct values) were quantified simultaneously using QuantStudio 3 software. Primer pairs were purchased from Integrated DNA Technologies, (Coralville, Iowa), and SYBR green (Applied Biosystems) was used for detection. Stability of mRNA was assessed as we described ^51,56. Briefly, samples were treated with 200ng/mL rhIL-19 for 24 hours, then 10μg/mL of the transcription inhibitor Actinomycin D was added, and RNA was extracted at indicated time points after addition of Actinomycin D. Each data point represents one biological replicate, determined by an average of three technical replicates within the experiment. Experiments were performed using three independent biological replicates. The following human primer pairs were used:

Gene	Forward	Reverse
GAPDH	ATCTTCTTTTGCGTCGCCAG	ATACGACCAAATCCGTTGACTC
Prox1	GCTCTCCTTGTCGCTCATAAA	GCATTGCACTTCCCGAATAAG
MAFB	GTGAGAAGGGATCGCAGTTT	CTTGCTGCCACGTTCTCTAT
KLF4	ACCTACACAAAGAGTTCCCATC	ATCTGAGCGGGCGAATTT
FOXO1	GATGGGCCTCATGTCTTGATAA	GGAGATGCAGAATGGAGATTCA
HHEX	GCGCTAAATGGAGGAGACTAAA	TTCTGTTCACTGGGCAAATCT
Angpt2	ATCAGGACACACCACGAATG	CATCCTCACGTCGCTGAATAA
CXADR	AGAAGCTACATCGGCAGTAATC	CTCTGAGGAGTGCGTTCAAA
C2CD4B	CCCATATCTCTTTCCAGACCATT	ACAATGCAGATGGGTGTACTT

Western Blotting and Protein Determination.

Cultured cells were starved (0.5% FBS) and treated with rhIL-19 for 0–48 hours and collected in protein extraction buffer, sonicated, boiled for 5 minutes, and lysates were frozen until use. Samples were run at approximately 15–30μg per well. Membranes were incubated in 5% bovine serum albumin (BSA) in tris buffered saline with tween (TBST) with a dilution of 1:1000 for IL-20Rα (Abclonal, Cat. A10308), IL-20Rβ (Abclonal, Cat. A7980), IL-19 (final concentration 0.2ug/mL, R&D Systems, Cat. AF1035), pSTAT3 (Cell Signaling Technology, Cat. 9131S), total STAT3 (Cell Signaling Technology, 9132), pVEGFR3 (Affinity Biosciences, Cat. AF3676), total VEGFR3 (final concentration 0.113ug/mL, Invitrogen, Cat. PA5–32409), VEGFC (final concentration 0.9ug/mL, Proteintech, Cat. 22601–1-AP), cyclin D1 (final concentration 0.1ug/mL, Santa Cruz Biotechnology, Cat. SC-718), cyclin B1 (NeoMarkers, Cat. MA-868-P0), cyclin E1 (final concentration 0.1ug/mL, Santa Cruz Biotechnology, Cat. SC-481), Prox1 (final concentration 1ug/mL, Invitrogen, Cat. PA5–85552; final concentration 0.2ug/mL, R&D Systems, Cat. AF2727), Angpt2 (final concentration 0.64ug/mL, Invitrogen, Cat. PA5–27297), C2CD4B (final concentration 0.5ug/mL, Invitrogen, Cat. PA5–70721), pFOXO1 (Cell Signaling, Cat. 9461S), total FOXO1 (Cell Signaling Technology, Cat. 2880S), KLF4 (Abclonal, Cat. A13673), pVE-cadherin Y685 (Invitrogen, Cat. 2460291), pVE-cadherin Y658 (final concentration 0.2ug/mL, Phosphosolutions, Cat. CP1981), total VE-cadherin (final concentration 0.2ug/mL, R&D Systems, Cat. AF1002), and a dilution of 1:6000 for GAPDH (Cell Signaling Technology, Cat. 2118L) and HSC70 (Santa Cruz Biotechnology, Cat. Sc-7298). Secondary HRP-labeled IgG antibodies (rabbit, Cell Signaling Technology, Cat. 7074S; mouse, Cell Signaling Technology, Cat. 7076S; goat, R&D Systems, Cat. HAF017) were incubated at a dilution of 1:3000. Reactive proteins were visualized using enhanced chemiluminescence according to manufacturer’s instructions and ThermoFisher Scientific iBright for imaging. Experiments were performed using three independent biological replicates and quantified using ImageJ for densitometry analysis of pixel intensity for each band. All values were normalized to housekeeping proteins (HSC70 or GAPDH) in each individual sample before dividing the values by the control samples in order to obtain the fold change with the control normalized to 1 (time = 0, or in the case of transfections Scram time = 0). Each data point represents one biological replicate.

Electric Cell-substrate Impedance Sensing (ECIS).

Experiments were run as recommended by the manufacturer, AppliedBiophysics. In short, 8W10E+ cultureware were incubated with 10mM L-cysteine (ThermoFisher Scientific, Cat. A10435) at room temperature and 150μg/mL rat tail collagen (Advanced BIOMATRIX, Cat. 5153–100MG) diluted in 0.15M NaCl (ThermoFisher Scientific, Cat. CAS 7647–14-5) at 37°C before adding culture media and evaluating proper connection to the AppliedBiophysics Z Theta Machine using Multiple Frequency/Time (MFT) Setting. Cells were starved (0.5% FBS) for 24 hours then seeded at 180,000 per well and a barrier would form over the course of 50-hours, represented by a plateau in resistance measured over time. Cells were then treated, and resistance was continually measured over time by AppliedBiophysics Z Theta Machine. Experiments were performed individually using three independent biological replicates, and at least three technical replicates used within each experiment.

Immunocytochemistry (ICC).

Cells were seeded equally into chamber slides, starved (0.5% FBS), and treated for 24 hours. Cells were fixed using 4% paraformaldehyde (PFA) (Electron Microscopy Sciences, Cat. 15710), blocked with 5% donkey serum (Jackson Labs, Cat. 017–000-121) and stained with VE-cadherin at 1:100 (final concentration 2ug/mL, R&D Systems, Cat. AF1002). Anti-goat IgG AF568 was used for visualization at 1:300 (final concentration 6.67ug/mL, Invitrogen, Cat. A11057). Cells were counter-stained with DAPI at 1:10000 (Invitrogen, Cat. D3571) and mounted using Fluoro-Gel (Electron Microscopy Sciences, Cat. 17985–10). Cells were imaged using Olympus BX53. Quantification of mean fluorescence intensity of VE-cadherin was performed using ImageJ. Each individual data point represents blindly selected high power field regions of each image in order to minimize differences in cell confluence, staining within slides and across slides. Experiments were performed using three independent biological replicates, and three technical replicates used within each experiment ^57–59.

Mice and Study Design.

LDL receptor knock out (Ldlr^−/−) mice on the C57BL/6J background purchased from Jackson labs (Catalog #002207) were bred with Il19^−/− mice previously described ^50,54,55 to obtain homozygous Ldlr^−/− Il19^−/− double knock out (dKO) mice. Mice were housed in an ALAC-accredited facility and maintained on a standard laboratory diet until study commencement, when appropriate. Mice of both sexes were at 8–12 weeks of age when entered into a study and either maintained on standard laboratory diet (5053 IRR, PicoLab, Rodent Diet 20, 20% protein, 4.5% fat, 6% fiber) or replaced with an atherogenic diet (AIN-76A Western Diet IRR, TestDiet, Cat. 1810061 (5342), 17.4% protein, 20% fat, 5% fiber) for 16 weeks. At the end of the study, lymphatic function and whole mount immunostaining were evaluated as explained below. All animal procedures were performed following Temple University’s Institutional Animal Care and Use Committee approved protocols.

Lymphatic Function with Evans Blue dye.

Mice were anesthetized with isoflurane (1–3%) and 3μL of Evans Blue dye (Fisher Chemical, Cat. E515–25, dissolved in sterile PBS at 1mg/mL) was injected in the subcutaneous tissue of the ear with a specialized Hamilton syringe (Hamilton Company, Reno, NV, USA, Ref. 87908, beveled, gauge 32, needle length 12mm, angle 12) and images were taken with Olympus SZ61. The following day, mice were euthanized, images were taken again, and ears were collected in 400μL formamide (EM Science, Cat. FX0420) for Evans Blue dye extraction and left overnight shaking at 55°C. The Evans Blue-formamide extraction was then loaded into a 96-well tray and colorimetric detection was read at 620nm on VersaMax from Molecular Devices (San Jose, CA, USA), compared to a standard curve of Evans Blue dye dissolved in formamide. Experiments were performed using a total of 48 mice, 12 mice per group. Ldlr^−/− HFD n=12 (5 male, 7 female), dKO HFD n=12 (6 male, 6 female), Ldlr^−/− Control n=12 (6 male, 6 female), dKO Control n=12 (6 male, 6 female). Each data point represents an average of three wells read from Evans Blue-formamide extraction individual mouse.

Whole Mount Immunohistochemistry.

Ears were collected and submerged in cold 1–4% PFA in PBS overnight on a shaker at 4°C. All steps were performed on a shaker at 4°C unless otherwise stated. Samples were washed in PBS for 10 minutes 3 times. If needed, samples were stored in PBS with 0.05% Sodium Azide (Sigma, S8032) for extended storage. The inner and outer layers of the ear were carefully separated with forceps. Samples were washed in PBST (0.3% Triton X-100 Sigma, T8787) for at least 3 hours. Samples were blocked with Immunomix (1% BSA Sigma, A3803, 5% Donkey Serum Jackson, 017–000-121, 0.3% Triton X-100, dissolved in PBS) for at least 3 hours. Samples were incubated overnight with primary antibodies diluted in Immunomix. The following antibodies are used: rat VE-cadherin at 1:200 (final concentration 0.625ug/mL, BD Biosciences, 550548) and rabbit LYVE1 at 1:200 (final concentration 0.5ug/mL, AngioBio, 11–034). Samples were washed in PBST for at least 6 hours. Samples are incubated overnight with secondary antibodies diluted in PBST. The following secondary antibodies were used: anti-rat AF647 at 1:800 (final concentration 1.875ug/mL, Jackson, 712–605-153) and anti-rabbit AF488 at 1:300 (final concentration 6.67ug/mL, Invitrogen, A21206). Samples were washed in PBST for at least 6 hours. Samples were incubated with DAPI at 1:500 in PBS for 7 minutes and washed in PBS for 7 minutes. Ears were mounted in Fluoro-Gel (Electron Microscopy Sciences, 17985–11) with the inner layer containing lymphatic vessels facing the coverslip. Images were taken using Leica DM4 B for 10x images to analyze lymphatic vessel branch number and Zeiss LSM 900 Airyscan 2 for 40x images to analyze VE-cadherin junctional morphology. Experiments were performed using a total of 24 mice, 6 mice per group group (Ldlr^−/− HFD, dKO HFD, Ldlr^−/− Control, dKO Control). Within each group, 3 male and 3 female were used. Data was analyzed by a blinded scorer. For all image analysis, at least 3 images were analyzed per mouse and averaged to yield each data point. Each individual data point represents one mouse.

Lymph Node Area Quantification.

Draining superficial cervical lymph nodes were dissected from each mouse and imaged using Olympus SZ61 next to a ruler. Images were equally scaled and quantified for lymph node area using ImageJ. Experiments were performed using a total of 24 mice, 6 mice per group group (Ldlr^−/− HFD, dKO HFD, Ldlr^−/− Control, dKO Control). Within each group, 3 male and 3 female were used. Data was analyzed by a blinded scorer.

Quantification and Statistical Analysis.

Results are expressed as mean±SD. Differences between groups were evaluated with the use of 1-way ANOVA for comparison of multiple groups, or 2-way ANOVA for comparison of multiple groups with varying treatments using GraphPad Prism (version 10.1.2), where appropriate. Subsequent post-hoc tests included Dunnett’s, Tukey’s, and Šídák’s multiple comparisons tests, when appropriate, as listed in the figure legends. Each experiment was performed using at least three independent biological replicates. When appropriate, at least three technical replicates were performed within each experiment, as described in the protocols above. Differences were considered significant when p < 0.05.

Results

IL-19 is expressed in and activates hdLECs.

IL-19 expression in hdLECs has not previously been reported ^46,52,53. To determine IL-19 expression and function in these cells, primary hdLECs were serum deprived in 0.5% FBS for 24 hours, then stimulated with 200ng/mL recombinant human IL-19 (rhIL-19) or 100ng/mL recombinant human VEGFC (rhVEGFC), a potent pro-lymphangiogenic factor. Figure 1A is a representative western blot indicating that both IL-19 and its receptor complex heterodimer, composed of IL-20Rα and IL-20Rβ, were expressed at basal levels in hdLECs. Stimulation with either 200ng/mL rhIL-19 or 100ng/mL rhVEGFC upregulated endogenous IL-19 protein and maintained stable receptor complex expression (Figure 1A–B). IL-19 does not induce VEGFR3 phosphorylation (Supplemental Data I A), which is expected as IL-19 is not a known activator of VEGFR3 nor does IL-19 induce expression of VEGFR3-specific activator VEGFC (Figure 1 C–D). The IL-20 heterodimeric receptor complex has been associated with STAT3 activation, but the effects of IL-19 stimulation on hdLEC signaling events have not been reported. Figures 1E and 1F show that IL-19 rapidly activated STAT3 phosphorylation, suggesting important events distal to IL-19 stimulation.

Figure 1.

IL-19 and its receptor subunits are expressed in hdLECs and IL-19 activates STAT3. A. Western blot showing IL-19 protein expression is induced within hdLECs by both rhIL-19 and rhVEGFC while IL-19’s receptor units (IL-20Rα and IL-20Rβ) remain unchanged. B. Densitometric analysis of rhIL-19 and rhVEGFC stimulated induction of endogenous IL-19 expression, n=3, 1-way ANOVA followed by Dunnett’s multiple comparisons test. C. Western blot showing VEGFC expression is unchanged by rhIL-19 treatment. D. Densitometric analysis indicating IL-19 does not induce VEGFC protein expression, n=3, 1-way ANOVA followed by Dunnett’s multiple comparisons test. E. Western blot indicating IL-19 induces STAT3 phosphorylation. F. Densitometric analysis of IL-19 stimulated STAT3 phosphorylation, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. All values are presented as the mean±SD.

IL-19 is pro-lymphangiogenic.

Induction of IL-19 expression in hdLECs suggested an important function for this cytokine in LEC pathophysiological processes, therefore, we assessed IL-19’s lymphangiogenic properties through migration, lymphatic network formation, and proliferation assays. Two assays were performed to evaluate migration: scratch assay and transwell migration. For the transwell migration, hdLECs were seeded into the upper chamber while the stimuli were placed in the lower chamber. Stimuli included PBS, 100ng/mL rhIL-19, 200ng/mL rhIL-19, or 100ng/mL rhVEGFC. After 24 hours, the number of hdLECs that migrated to the bottom of the transwell were removed and counted, revealing both rhIL-19 and rhVEGFC attracted significantly more hdLECs compared to PBS controls (Figure 2A). For the scratch assay, hdLECs grew to confluency in chamber slides, starved with 0.5% FBS, scratched with a 2mm apparatus, and stimulated with the same stimuli listed above. Images were taken immediately at 0 hours and 20 hours after treatment to rule out any potential effects of proliferation, as shown by IL-19’s inability to significantly increase hdLEC number by day 3 (Supplemental Data II A–B). Figure 2B shows that treatment with rhIL-19 significantly increased the percent change in scratch area over 20 hours. We next performed a vascular network assay in which pre-treated hdLECs were seeded equally into growth-factor depleted (1% FBS) Matrigel-coated wells and re-treated for 6 hours with the same stimuli described above in addition to co-treatment with both rhIL-19 and rhVEGFC. Lymphatic network (tube and loop) formation was quantified, showing that rhIL-19 significantly increased each aspect of vascular network formation (Figure 2C–D and Supplemental Data II C). We next determined IL-19’s effects on hdLEC proliferation using multiple methods. First, we determined that IL-19 upregulates cyclin D1, cyclin B1, and cyclin E1 protein expression (Figure 2E–F). Next, hdLECs were seeded equally, starved in 1% FBS and stimulated with the same stimuli described in the experiments above. The cells were re-stimulated on day three and counted for the number of cells on day six. Treatment with rhIL-19 significantly increased the proliferation of hdLECs by day six compared to PBS negative control (Figure 2G). Finally, hdLECs were seeded equally in 0.05% FBS for overnight starvation and stimulated for 24 additional hours. BrdU was added for the final 4 hours and detected through absorbance at 450nm. Treatment with IL-19 induced more BrdU incorporation into hdLECs compared to PBS treatment (Figure 2H). Cell number was not significantly increased with rhIL-19 or rhVEGFC on day 3 (Supplemental Data II B) despite rhIL-19 inducing cyclin expression (Figure 2E–F) and BrdU incorporation (Figure 2H) within 24 hours. This likely occurred because the doubling time for LECs is much greater than 24 hours while maintained in serum-reduced medium. In each of these assays, rhIL-19 effects on migration, lymphatic network formation, and proliferation were comparable to that observed for rhVEGFC, a recognized pro-lymphangiogenic factor. Together, these data suggest that IL-19 is a pro-lymphangiogenic factor in vitro.

Figure 2.

IL-19 is lymphangiogenic for hdLEC. A. Number of hdLECs migrating to the bottom of a transwell, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. B. Migration measured by percent change in scratch area, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. C. Quantification of tube formation from D, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. D. Representative photomicrographs of IL-19-induced hdLEC lymphatic tube and network formation. E. Western blot showing IL-19 induction of cyclin D1, B1, and E1. F. Densiometric analysis of rhIL-19 induction of cyclin D1, B1, and E1, n=3 each, all analyzed 1-way ANOVA followed by Dunnett’s multiple comparisons test. G. IL-19 induces hdLEC proliferation by cell number, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. H. IL-19 increases BrdU incorporation in hdLEC, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. All values are presented as the mean±SD.

IL-19 promotes a lymphangiogenic and functional transcriptional profile.

The potent effects of IL-19 on hdLEC physiological processes suggested the necessity of IL-19-induced lymphangiogenic transcriptional programs. The effect of IL-19 on hdLEC gene expression was determined by non-biased bulk RNAseq which showed robust induction of many genes after 24 hours of 200ng/mL rhIL-19 stimulation compared with the unstimulated control (Figure 3A). Gene ontology analysis showed that five of the top 10 genetic processes participated in transcription and transcriptional regulation (Figure 3B). Several mRNA transcripts associated with cellular processes relevant to lymphangiogenesis and lymphatic function through junctional morphology were also upregulated (summarized in Table 1). The expression of a select few genes from these groups were validated with qRT-PCR after IL-19 treatment and compared to the expression profile after VEGFC treatment (Figures 3C–D). Most notable pro-lymphangiogenic genes included: Prospero homeodomain protein 1 (Prox1), Krϋppel-like transcription factor 4 (KLF4), Forkhead box protein O1 (FOXO1), MAF BZIP Transcription Factor B (MAFB), and Hematopoietically-expressed homeobox protein (HHEX). Importantly, Prox1 is the master transcription factor recognized to control lymphatic development and lymphangiogenesis ^60–62 and its IL-19-driven expression was confirmed by western blot (Figure 3E–F). KLF4, FOXO1, MAFB, and HHEX, have also been implicated as upstream regulators of lymphatic gene expression while also individually promoting lymphatic vessel formation ^60,63,64.

Figure 3.

IL-19 increases lymphangiogenic and permeability transcriptional profile in hdLECs. A. Volcano plot indicating the number of transcripts that increase (red) and decrease (blue) with 24-hour rhIL-19 treatment. B. Gene ontology of bulk RNAseq of hdLEC following 24 hours of rhIL-19 stimulation. C., D. Validation of IL-19-driven mRNA induction by qRT-PCR compared to that of VEGFC-driven mRNA induction, n=3 per gene analyzed and per stimulant, each analyzed with 2-way ANOVA followed by Turkey’s multiple comparisons test. E. Western blots of IL-19-driven Prox1, Angpt2, and C2CD4B protein expression. F. Densitometric analysis of Prox1, Angpt2, and C2D4B protein expression, n=3 per protein analyzed, each analyzed by 1-way ANOVA followed by Dunnett’s multiple comparisons test. All values are presented as the mean±SD.

Table 1.

Transcripts upregulated by IL-19 in hdLECs.

Gene Name	Acronym	Fold-Increase
Transcription Factors and DNA Binding
Prospero Homeobox 1	Prox1	6.18
Tigger Transposable Element Derived 3	TIGD3	5.14
MYCL Proto-Oncogene, BHLH Transcription Factor	MYCL	4.98
Y-Box Binding Protein 2	YBX2	4.44
Tigger Transposable Element Derived 4	TIGD4	4.29
Zinc Finger Protein 442	ZNF442	4.21
Nuclear Receptor Subfamily 4 Group A Member 3	NR4A3	4.02
T-Complex 10 Like	TCP10L	3.7
MAF BZIP Transcription Factor B	MAFB	3.6
Zinc Finger Protein 121	ZNF121	3.21
Early Growth Response 1	EGR1	3.18
E2F Transcription Factor 7	E2F7	2.98
Forkhead Box O1	FOXO1	2.95
Zinc Finger Protein 844	ZNF844	2.88
Early Growth Response 2	EGR2	2.68
POU Class 3 Homeobox 1	POU3F1	2.57
KLF Transcription Factor 4	KLF4	1.94
Junctional Regulation and Permeability
C2 Calcium Dependent Domain Containing 4B	C2CD4B	7.51
Coxsackie Virus And Adenovirus Receptor	CXADR	3.55
Cingulin-Like Protein 1	CGNL1	3.06
Chemokines
C-X-C Motif Chemokine Ligand 3	CXCL3	3.68
C-X-C Motif Chemokine Ligand 2	CXCL2	3.25
C-C Motif Chemokine Ligand 25	CCL25	3.08
C-C Motif Chemokine Ligand 7	CCL7	2.72
C-X3-C Motif Chemokine Ligand 1	CX3CL1	2.34
C-C Motif Chemokine Ligand 2	CCL2	2.12
C-X-C Motif Chemokine Ligand 8	CXCL8	1.92

hdLECs were starved for 24 hours, then stimulated with 200ng/mL rhIL-19 or saline for an additional 24 hours. The effect of IL-19 on hdLECs gene expression was determined by non-biased bulk RNA sequencing. Transcripts associated with cellular processes relevant to lymphangiogenesis and lymphatic function including transcriptional regulatory proteins, junctional regulation and permeability, and chemokines were significantly increased by IL-19.

Lymphatic vessel function largely relies on the proper expression and structure of junctional proteins which regulate lymphatic permeability ^4–6,65. Several genes which regulate LEC junctions and button/zipper morphology were also increased, including Angiopoeitin-2 (Angpt2), Coxsackie virus and adenovirus receptor (CXADR), and C2 calcium dependent domain containing 4B (C2CD4B). While C2CD4B has never been studied in the context of LECs, it has been implicated in regulating blood endothelial cell architecture and permeability during inflammation ⁶⁶. IL-19-driven expression of both Angpt2 and C2CD4B expression were confirmed by western blot (Figure 3E–F).

Lymphatic vessels also function as conduits for inflammatory cell egression, which is guided via chemokine gradients ^67,68. Interestingly, several chemokines were increased in hdLECs upon rhIL-19 treatment included C-X-C motif chemokine ligand 2 (CXCL2), C-X-C motif chemokine ligand 8 (CXCL8), C-C motif chemokine ligand 7 (CCL7), and C-X3-C motif chemokine ligand 1 (CX3CL1) (summarized in Table 1).

VEGFC is considered the canonical LEC growth factor, but remarkably, IL-19 did not increase VEGFC expression or VEGFR3 phosphorylation (Figure 1C–D, Supplemental Data I A). Therefore, we investigated if genes that were known to be upregulated by IL-19 on RNAseq were differentially regulated by VEGFC. Figure 3C shows that the magnitude and kinetics of the expression of several mRNA transcripts, including Prox1, KLF4, Angpt2, and C2CD4B, were similar when comparing IL-19 and VEGFC treatments. However, several mRNA transcripts of interest that were increased by IL-19 treatment, including FOXO1, MAFB, HHEX, and CXADR, were not upregulated as robustly with VEGFC treatment (Figure 3D). This suggests that IL-19 and VEGFC have similarities but also distinct differences in affecting the hdLEC transcriptome.

STAT3 does not regulate IL-19-driven hdLEC proliferation or Prox1 expression.

Because STAT3 is rapidly phosphorylated in response to rhIL-19 stimulation (Figure 1E–F), we investigated if STAT3 was necessary for IL-19-driven Prox1 expression. STAT3 specific siRNA was used to knock STAT3 down in hdLECs, which were subsequently stimulated with rhIL-19. Figure 4A shows that STAT3 protein was successfully knocked down by siRNA. However, the abundance of IL-19-driven Prox1, cyclin D1, and cyclin E1 were not reduced in STAT3 siRNA treatment groups, indicating that IL-19 induced proliferation and expression of Prox1 was not dependent on STAT3 activation (Figure 4A, C). In a second approach to determine the role of STAT3 in IL-19-induced hdLEC activation, a BrdU assay was performed in the presence of the STAT3-specific inhibitor WP1066 ^47,54. Cells were equally seeded in 0.05% FBS overnight then stimulated for 24 hours with the addition of either 1μM WP1066 or an equal volume of DMSO for control. BrdU was added for the final 4 hours of treatment and detected through absorbance at 450nm. Figure 4B indicates pharmacological inhibition of STAT3 did not reduce IL-19-driven BrdU incorporation, further supporting the notion that IL-19-driven hdLEC activation is STAT3-independent.

Figure 4.

IL-19 induction of Prox1 is independent of STAT3 but downstream of KLF4. A. Representative western blot showing that IL-19-driven protein expression is not dependent on STAT3. B. IL-19-driven proliferation was unchanged in the presence of STAT3 inhibitor WP1066, measured by BrdU incorporation, n=3, 2-way ANOVA followed by Tukey’s multiple comparisons test. C. Densitometric analysis showing STAT3 knockdown does not affect Prox1, cyclin D1, or cyclin E1 expression, normalized to representative housekeeping protein and Scram siRNA 0-hour timepoint, n=3, 2-way ANOVA followed by Tukey’s multiple comparisons test. D. Representative western blot and densitometric quantification of KLF4 upregulation with IL-19 treatment, n=3, 1-way ANOVA followed by Dunnett’s multiple comparisons test. E. Representative western blot showing KLF4 is upstream of Prox1 and Angpt2 upregulation. F. Densitometric analysis showing a decrease in Prox1 and Angpt2 upregulation with the knockdown of KLF4, n=3, 2-way ANOVA followed by Tukey’s multiple comparisons test. All values are presented as the mean±SD.

In search of alternative mechanisms for IL-19 regulation of Prox1 expression, we turned to prior reports indicating that IL-19 affects post-transcriptional processing of inflammatory mRNA ^50,56. To determine if IL-19 affected Prox1 mRNA stability, we performed qRT-PCR in the presence of the transcriptional inhibitor Actinomycin D to determine RNA degradation normalized to GAPDH. In contrast to the literature indicating that IL-19 decreases mRNA stability, IL-19 stimulation of hdLECs significantly increased Prox1 mRNA stability in hdLECs, suggesting additional mechanisms of IL-19-mediated gene expression (Supplemental Data III A).

FOXO1 is a transcription factor that plays an important role in lymphatic vessel function and Angpt2 upregulation ^69,70, and our RNAseq determined that IL-19 induced FOXO1 expression 2.95-fold, therefore, we wanted to determine if it plays a role in the IL-19 pathway. We validated that IL-19 increased FOXO1 protein expression over 48 hours (Supplemental Data III B–C). FOXO1 was also regulated via phosphorylation, which is associated with nuclear exclusion and degradation in the cytoplasm ^71–73. Interestingly, IL-19 rapidly phosphorylated FOXO1 by 90 minutes without affecting total FOXO1 levels (Supplemental Data III D–E). This indicates a possible self-regulatory pathway where IL-19 may induce FOXO1 phosphorylation for potential degradation but promote long-term upregulation of total FOXO1 protein levels. This suggests multiple tightly synchronized pathways that are activated by IL-19. We then depleted FOXO1 from hdLECs by siRNA knockdown, and determined that Angpt2, but not Prox1 expression, was significantly lower in hdLECs depleted of FOXO1 (Supplemental Data III F–G). Of note, Angpt2 expression was decreased at basal levels and after 24 hour IL-19 stimulation with FOXO1 knockdown, indicating that FOXO1 is an important upstream regulator of Angpt2 expression even outside of IL-19’s pathway, as shown in the literature ^74,75.

To uncover potential initial regulators of the IL-19 molecular pathway, we evaluated KLF4, as IL-19 stimulation upregulated KLF4 mRNA (Figure 3C) and protein expression (Figure 4D). Subsequent knockdown of KLF4 with siRNA revealed IL-19 was no longer able to increase Prox1 or Angpt2 protein expression (Figure 4E–F). This is noteworthy because KLF4 is described in the literature to regulate expression of Prox1 but not Angpt2 ^60,76. Taken together, IL-19 may activate multiple and distinct pathways responsible for Prox1 and Angpt2 upregulation with subsequent ramifications for lymphangiogenesis and LEC function.

IL-19-associated lymphangiogenesis and Angpt2 upregulation are Prox1-dependent.

Prox1 is central to lymphatic function and is considered the master transcriptional regulator of lymphangiogenesis. Since Prox1 was significantly induced with IL-19 treatment, we performed experiments to confirm the role of Prox1 in IL-19-stimulated lymphangiogenesis. hdLECs were transfected with Prox1 siRNA, and Figure 5A and 5B confirm siRNA knockdown of Prox1 protein abundance. Next, Prox1 siRNA treated hdLECs were challenged with 200ng/mL rhIL-19 and hdLEC proliferation and migration were performed, as described previously. Figure 5C shows that Prox1 knockdown significantly reduced IL-19-driven hdLEC proliferation. Figure 5D and 5E indicate that Prox1 knockdown significantly reduced IL-19-driven hdLEC migration. Interestingly, Prox1 knockdown also mitigated IL-19-driven Angpt2 expression, suggesting Prox1 mediates IL-19-driven Angpt2 expression (Figure 5A–B), which is novel as Prox1 has not been reported as an upstream regulator of Angpt2 in hdLECs. Because both IL-19 and VEGFC induced endogenous IL-19 expression (Figure 1A–B), it was important to determine a potential autocrine pathway for IL-19 activity in hdLECs. IL-19 was knocked down with specific siRNA and treated with exogenous IL-19, validating that, in the absence of endogenous IL-19, the hdLECs were no longer able to upregulate Prox1 protein (Figure 5F, H). Separately, hdLECs were stimulated with exogenous VEGFC after endogenous IL-19 was knocked down with siRNA, and hdLECs were again unable to significantly upregulate Prox1 protein (Figure 5G, I). Importantly, knockdown of endogenous IL-19 did not affect abundance of its receptor complex subunits IL-20Rα or IL-20Rβ (Supplemental Data I B–C). Together, this indicates that endogenous IL-19 produced by hdLECs is a previously unrecognized mediator of VEGFC, the canonical regulator of lymphatic development.

Figure 5.

IL-19-stimulated lymphangiogenesis is Prox1-dependent and IL-19 induces a vital autocrine pathway. A. Representative western blot showing siRNA knockdown of Prox1 protein and mitigation of IL-19-stimulated Angpt2 expression. B. Densitometric analysis of Prox1 and Angpt2 protein expression with Prox1 siRNA knockdown, normalized to representative housekeeping protein and Scram siRNA 0-hour timepoint, n=3, 2-way ANOVA followed by Tukey’s multiple comparisons test. C. Prox1 knockdown mitigates IL-19-driven hdLEC proliferation, n=4, 2-way ANOVA followed by Tukey’s multiple comparisons test. D. Representative photomicrograph showing Prox1 knockdown mitigates IL-19-driven hdLEC migration. E. Quantification of reduction in IL-19-driven hdLEC migration with Prox1 siRNA, n=3, 2-way ANOVA followed by Tukey’s multiple comparisons test. F., G. Representative western blot showing siRNA knockdown of IL-19 mitigates rhIL-19-stimulated (F) and rhVEGFC-stimulated (G) Prox1 expression. H., I. Densitometric analysis of Prox1 protein expression with IL-19 siRNA knockdown and rhIL-19 (H) or rhVEGFC (I) treatment, n=3 each, all analyzed by 2-way ANOVA followed by Tukey’s multiple comparisons test. All values are presented as the mean±SD.

IL-19 regulates lymphatic barrier function.

RNAseq indicated that IL-19 modified permeability-related gene expression (Figure 3A–D), suggesting effects on junction morphology and permeability which are closely related to lymphatic function. Lymphatic vessels have two distinct segments with differing junction morphology, which dictates each segment’s functionality ^4,6,65. Of all the known junctional proteins, VE-cadherin morphology is the most studied and well-defined through molecular mechanisms regulated by Angpt2 ^3,8,77,78. Angpt2 is known to differentially regulate both discontinuous VE-cadherin buttons in initial lymphatics and continuous VE-cadherin zippers in collecting lymphatics throughout development by altering VE-cadherin phosphorylation ^7–9,11.

We initially determined that IL-19 induced rapid VE-cadherin phosphorylation at both Y658 and Y685, an independent molecular readout of LEC button formation and permeability ^8,78,79 (Figure 6A–C). Both phosphorylation sites are thought to be important for inducing a discontinuous and button-like morphology of VE-cadherin ^8,80,81. Since IL-19 induced expression of Angpt2 and C2CD4B (Figure 3E–F), we next investigated hdLECs paracellular permeability utilizing ECIS to measure the resistance across an hdLEC barrier in vitro. This allowed us to determine hdLECs permeability under long-term conditions which might be affected by gene expression. In this experiment, hdLECs were starved (0.5% FBS) for 24 hours, seeded at 180k per well, and allowed to form a stable barrier for an additional 50 hours in starvation media ^82–84. VE-cadherin is a major component of lymphatic junctional complexes and is often described in two structural forms: buttons or zippers ^3,4. Buttons are discontinuous structures of VE-cadherin found in blunt-ended initial lymphatic vessels, promoting paracellular permeability for the uptake of fluid, macromolecules, and immune cells ^13,85. Zippers are continuous structures of VE-cadherin found in collecting lymphatic vessels which are less permeable and function to maintain lymph within the system ^78,86. Therefore, VE-cadherin ICC staining was used to visualize IL-19 effects on hdLEC monolayer junctional morphology as it relates to the functional output measured by ECIS. hdLECs successfully formed a stable barrier that impeded resistance, indicated by a plateau in resistance readout (Figure 6E, H, K). After treatment with 20μg/mL or 100μg/mL human oxLDL, there is a dose-dependent increase in resistance across the hdLEC barrier (Figure 6F), indicative of decreased permeability and visualized by a continuous staining pattern of VE-cadherin zippers in ICC (Figure 6G). This was expected, as oxLDL is a known pro-inflammatory molecule in atherosclerotic plaque formation and has previously been associated with pro-inflammatory lymphatic zippering during liver pathology ^58,87. With rhIL-19 or rhVEGFC treatment, there was a significant decrease in resistance (Figure 6I), suggesting increased permeability or buttoning of lymphatic junctions as visualized by discontinuous VE-cadherin in ICC (Figure 6J). When hdLECs are co-treated with both rhIL-19 and oxLDL, IL-19 mitigated the oxLDL-associated increase in resistance by maintaining the barrier function closer to that of control cells (Figure 6L). This is suggested visually by the appearance of discontinuous and fragmented VE-cadherin immunohistochemical staining pattern similar to that of PBS alone (Figure 6M). Additionally, quantification of the mean fluorescence intensity of VE-cadherin staining at the borders of hdLECs exhibited a higher intensity with oxLDL treatment compared to PBS control, indicating a higher concentration of VE-cadherin at cell-cell borders and suggestive of a zipper-like morphology (Figure 6D) ^57–59. This intensity is decreased when hdLECs are co-treated with both IL-19 and oxLDL, suggesting a lesser concentration of VE-cadherin at the cell borders and a more fragmented button-like morphology (Figure 6D) ^57–59. This indicates that IL-19 may regulate hdLEC junctional morphology and subsequently hdLEC barrier function, potentially playing a protective role by reducing the zippering of lymphatics induced by pro-inflammatory stimuli such as oxLDL. Collectively, these data indicate that IL-19 can regulate hdLEC monolayer junctional properties and mitigate the effects of oxLDL on the permeability of cultured hdLECs.

Figure 6.

IL-19 regulates hdLEC permeability. A. Representative western blots of IL-19-driven VE-cadherin phosphorylation. B., C. Quantification of rapid phosphorylation of VE-cadherin at Y658 (A) and Y685 (B) with rhIL-19 treatment, n=3 each, 1-way ANOVA followed by Tukey’s multiple comparisons test. D. Quantification of mean fluorescence intensity of VE-cadherin staining at the borders of hdLECs within images G., J., and M, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. E., H., K. Successful ECIS barrier formation is confirmed prior to each treatment, n=3 each. F. Treatment with oxLDL increases hdLEC barrier resistance suggestive of decreased permeability n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. G. VE-cadherin immunohistochemical staining of oxLDL-treated hdLEC suggestive of continuous (zipper) formation. I. Treatment with rhIL-19 decreases hdLEC barrier resistance similar to rhVEGFC, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. J. VE-cadherin immunohistochemical staining suggestive of discontinuous (button-like) pattern. L. Co-treatment with both oxLDL and rhIL-19 mitigates oxLDL-induced increase in resistance and maintains stable hdLEC barrier resistance, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. M. VE-cadherin immunohistochemical staining suggestive of discontinuous (button-like) pattern. All values are presented as the mean±SD.

IL-19 regulates LEC Function in vivo.

It was important to determine the effects of IL-19 on lymphatic function in vivo. We previously described an Il19^−/− Ldlr^−/− double knockout (dKO) mouse which develops exacerbated plaque formation ⁵⁰. We placed dKO and Ldlr^−/− control mice on 16 weeks western high fat diet (HFD) to induce pro-atherogenic conditions and used these mice as a model to determine IL-19’s effects on lymphatic function using the Evans Blue dye removal assay. Evans Blue dye (3μL of 1mg/mL concentration) was injected into the subcutaneous tissue of the ear. After 16 hours, dye was extracted from the ear and quantified. Lower values indicate minimal amounts of exogenous dye remaining in the tissue, suggesting proper lymphatic clearing of extraneous fluid and macromolecules. Higher values indicate dye presence in the tissue, suggesting poor lymphatic clearance. Figure 7A and 7B show that dKO mice had significantly more dye remaining in ear tissue compared with Ldlr^−/− control mice on HFD (9.526*10⁻⁴ M +/− 5.12*10⁻⁴ M versus 3.625*10⁻⁴ M +/− 2.04*10⁻⁴ M for dKO versus Ldlr^−/−, p=0.0002, respectively). These mice had no overt signs of lymphatic dysfunction such as changes in weight (Figure 7C), fluid retention (edema), chylous ascites, chylothorax, or lymph node size (Figure 7D–E). However, age-matched dKO and Ldlr^−/− mice on standard laboratory diet did not show statistically significant differences in the amount of dye remaining in their ear tissue (Figure 7A–B). Similar to the ECIS experiments (Figure 6E–L), this suggests IL-19 plays an important role in preventing lymphatic malfunction during pro-inflammatory conditions such as the presence of HFD or oxLDL. As expected, mice on HFD weighed more at the end of the study compared to age-matched mice on standard laboratory diet (Figure 7C).

Figure 7.

IL-19 regulates lymphatic function in vivo. A. Reduced lymphatic uptake of Evans Blue in ears of dKO mice compared to Ldlr^−/− after 16-week high fat diet (HFD). B. Quantification of amount of dye remaining in the ear, Ldlr^−/− HFD n=12 (5 male, 7 female), dKO HFD n=12 (6 male, 6 female), Ldlr^−/− Control n=12 (6 male, 6 female), dKO Control n=12 (6 male, 6 female), 2-way ANOVA followed by Tukey’s multiple comparisons test. C. Weight of dKO and Ldlr^−/− mice at the end of the study, Ldlr^−/− HFD n=12 (5 male, 7 female), dKO HFD n=12 (6 male, 6 female), Ldlr^−/− Control n=12 (6 male, 6 female), dKO Control n=12 (6 male, 6 female), 2-way ANOVA followed by Tukey’s multiple comparisons test. D. Representative gross images of superficial cervical lymph nodes. E. Quantification of lymph node size by area, Ldlr^−/− HFD n=6 (3 male, 3 female), dKO HFD n=6 (3 male, 3 female), Ldlr^−/− Control n=6 (3 male, 3 female), dKO Control n=6 (3 male, 3 female), 2-way ANOVA followed by Tukey’s multiple comparisons test. F. Whole mount immunostaining of ears for LYVE1-positive (green) initial lymphatic vessels. G. Quantification of number of lymphatic vessel branches from F, Ldlr^−/− HFD n=6 (3 male, 3 female), dKO HFD n=6 (3 male, 3 female), Ldlr^−/− Control n=6 (3 male, 3 female), dKO Control n=6 (3 male, 3 female), 2-way ANOVA followed by Tukey’s multiple comparisons test. H. High magnification whole mount immunostaining of LYVE1-positive (green) initial lymphatic vessels in ears with VE-cadherin (red) junction staining. I. Evaluation of VE-cadherin junctional morphology as percent of continuous junctions, Ldlr^−/− HFD n=6 (3 male, 3 female), dKO HFD n=6 (3 male, 3 female), Ldlr^−/− Control n=6 (3 male, 3 female), dKO Control n=6 (3 male, 3 female), 2-way ANOVA followed by Tukey’s multiple comparisons test. Control = standard laboratory diet. All values are presented as the mean±SD.

Whole mount immunostaining of ear lymphatic vessels indicated significantly decreased number of lymphatic branch points in dKO mice on both standard laboratory diet and HFD when compared to Ldlr^−/− mice on HFD (Figure 7F–G). The number of lymphatic branch points in Ldlr^−/− mice on HFD was significantly higher than that of Ldlr^−/− mice on standard laboratory diet, suggesting possible inflammation-associated lymphangiogenesis with HFD, which was not present in dKO mice (Figure 7F–G). Blunt-ended initial lymphatic vessels in dKO mice on HFD had a significantly higher percentage of continuous zipper-like morphology of VE-cadherin staining compared to all other mice (Figure 7H–I). Together, less lymphatic branch points and increased zippered morphology in dKO mice on HFD could explain the poor lymphatic clearance of Evans Blue dye in dKO mice (Figure 7A–B).

These data, taken together with the ECIS paracellular permeability data, strongly suggest that IL-19 can maintain lymphatic function by regulating the morphology of VE-cadherin junctions between LECs in vitro and in vivo, particularly in the context of hyperlipidemic and pro-inflammatory conditions of oxLDL and HFD.

Discussion

In adults, lymphangiogenesis is stimulated in pathological conditions such as inflammation, which requires expansion and remodeling of lymphatics for removal of tissue fluid, cytokines, chemokines, and immune cells ^31,33,37. However, lymphangiogenesis that occurs during inflammation leads to the formation of malfunctional lymphatic vessels, resulting in poor lymphatic drainage and chronically increased inflammation ^5,88,89. Identification of modalities that could limit vascular inflammation yet increase lymphatic vessel formation and function should be considered a therapeutic goal.

With the exception of VEGFC, the literature reflects a need for investigation into the effects of cytokines on lymphangiogenesis and lymphatic function in adult organisms. Despite IL-19 being understudied in lymphatics, a few cytokines have been previously investigated and have been associated with lymphangiogenesis ⁸⁹. IL-17, a pro-inflammatory cytokine, induces lymphatic microvessel formation in mouse cornea in vivo and promotes proliferation and lymphatic network formation in vitro ⁹⁰. IL-8, another pro-inflammatory cytokine, also upregulates mouse cornea lymphatic vessel formation and in vitro lymphangiogenesis ⁹¹. IL-3 is a cytokine mainly studied for its role in hematopoiesis and has been shown to induce Prox1 expression in LECs, but subsequent effects on lymphangiogenesis were not evaluated ^92,93. Interestingly, anti-inflammatory IL-4 and IL-13 have been shown to be anti-lymphangiogenic in vitro and in vivo ⁹⁴. IL-20 also induced proliferation and migration of a human microvascular endothelial cell line which was transformed to expressed LEC markers ^95,96. These studies support the Janus phenomenon which states that pro-inflammatory cytokines are pro-lymphangiogenic while anti-inflammatory cytokines are anti-lymphangiogenic ³⁹, emphasizing the uniqueness of IL-19 breaking this phenomenon as an anti-inflammatory yet pro-angiogenic and immuno-modulatory factor. Although shown to be expressed at low levels in coronary artery endothelial cells, the expression and functionality of IL-19 in lymphatic physiological processes has not been described. In this study, we show that IL-19 and its receptor complex were expressed in hdLECs. Importantly, IL-19 induced its own endogenous expression, suggesting a feed-forward autocrine mechanism. Further, we show that IL-19 had lymphangiogenic effects on primary hdLECs, increasing proliferation, migration, and lymphatic network formation.

Recognized lymphangiogenic growth factors such as VEGFC, VEGFD, PDGF-BB, FGF2, and S1P regulate lymphangiogenic gene expression ⁹⁷. Numerous genes have been studied as inducers of LEC development and maturation. The most-studied pathway in lymphangiogenesis is VEGFC activation of VEGFR3 leading to Prox1 upregulation ⁶⁰. IL-19 treatment did not increase endogenous VEGFC or phosphorylation of its receptor VEGFR3, but it did promote robust induction of many genes, particularly those associated with gene transcription, lymphatic junction regulation, and chemokines. The magnitude and kinetics of expression of several mRNA transcripts observed to be increased by IL-19 were similarly increased by VEGFC. However, several mRNA transcripts increased by IL-19, particularly transcription factors including MAFB, HHEX, and FOXO1 had a differential response to VEGFC treatment compared to that of IL-19 treatment, suggesting important distinctions between the effects of these cytokines on the hdLEC transcriptome that require further exploration. Among the several transcriptional regulatory proteins upregulated by IL-19, perhaps the most important is Prox1, considered the master transcription factor for lymphatic fate and development ^60,62,98,99. It has been shown that inflammation induces Prox1 expression and subsequent lymphangiogenesis through upregulation of VEGFR3, but very little has been reported on immuno-modulatory cytokine induction of Prox1 expression ¹⁰⁰. STAT3 is a transcription factor activated through phosphorylation in response to various cytokines and growth factors, including IL-19 ^48,54,101. STAT3 is known to mediate the expression of a variety of genes, including cell cycle regulators, in response to stimuli ^102,103. Here, we show IL-19 canonically activated STAT3 through phosphorylation.

In search of mechanisms responsible for IL-19 induction of Prox1 expression, we knocked down STAT3 with siRNA, which had no effect on IL-19-driven proliferation or Prox1 abundance. Since IL-19 has been shown to decrease the stability of mRNA which contains AU-rich elements (AREs) in the 3-prime untranslated region (3’ UTR) ^51,56, we determined if IL-19 regulated post-transcriptional mRNA processing in LECs. In contrast to prior studies in other cell types suggesting IL-19 destabilizes mRNA, IL-19 appeared to increase Prox1 mRNA stability in hdLECs, which may account for its increased mRNA and protein abundance upon IL-19 stimulation. Curiously, Prox1 contains multiple AREs in its 3’ UTR, which points to alternative mechanisms responsible for IL-19 induction of Prox1 abundance ¹⁰⁴. The transcription factor FOXO1 regulates lymphatic valve formation and is generally recognized as a transcriptional repressor ^71–73,105. However, under inflammatory and other conditions, FOXO1 activity can stimulate endothelial Angpt2 expression ^70,106–109. Interestingly, in our hands, IL-19 increased rapid phosphorylation and long-term expression of FOXO1. FOXO1 knockdown reduced IL-19 stimulated Angpt2 expression, but it had no effect on Prox1 expression. Further investigation into the precise mechanism by which IL-19 regulates FOXO1 activation is warranted. To further characterize how IL-19 regulates Prox1 abundance, we next turned to transcription factor KLF4 which was upregulated by IL-19 and is known to play a role in lymphangiogenesis ^60,76. Upon KLF4 knockdown, IL-19 was no longer able to upregulate Prox1 or Angpt2 protein expression, indicating that KLF4 is an upstream regulator in IL-19’s molecular pathway in hdLECs.

To characterize its role in IL-19-stimulated lymphangiogenesis, siRNA knockdown of Prox1 resulted in mitigation of IL-19 driven proliferation and migration, pointing to a central role of Prox1 in IL-19-stimulated lymphangiogenesis. This might be expected as the short-term inability to generate Prox1 transcripts has been shown to decrease cell cycle progression ¹¹⁰. However, since inflammation is associated with malfunctional lymphangiogenesis, it is critical to identify a molecule that promotes functional lymphangiogenesis. Lymphatic vessel functions of absorption and transport largely rely on the proper morphology of junctional proteins such as VE-cadherin ⁵. IL-19 upregulated several junctional-regulating proteins including CXADR, known to be involved in LEC adhesion, migration, tube formation, and importantly lymphatic paracellular permeability ^111,112. While it has never been studied in the context of LECs, C2CD4B has been implicated in regulating blood endothelial cell architecture and permeability during inflammation ⁶⁶. Importantly, IL-19 also significantly upregulated Angpt2 expression, a vital differential regulator of both buttons in initial lymphatics and zippers in collecting lymphatics ⁸. Angpt2 is context-dependent, but it does participate in lymphangiogenesis at the stages of vascular remodeling and vessel maturation, and high doses of Angpt2 are proliferative ^113,114. Mutations in Angpt2 are associated with lymphatic malformations resulting in hydrops fetalis and primary lymphedema ^9,115,116. In our study, in addition to FOXO1, Prox1 knockdown significantly reduced IL-19-driven Angpt2 protein upregulation, a novel finding as Prox1 has not been reported as an upstream regulator of Angpt2. Additionally, both IL-19 and VEGFC induced endogenous production of IL-19 within hdLECs. When endogenous IL-19 is knocked down with siRNA, neither IL-19 nor VEGFC are able to significantly increase Prox1 expression, suggesting that endogenous IL-19 expression might be a previously unrecognized component of the canonical VEGFC lymphangiogenic pathway.

There are a limited number of studies that evaluate the regulation of lymphatic junctional morphology in sterile inflammatory conditions. Many studies use infection-based techniques to induce inflammatory environments and focus solely on changes in button morphology of initial lymphatic vessels ^5,13–15. In general, inflammatory stimuli decrease initial lymphatic vessel permeability and may increase collecting lymphatic vessel permeability, contributing to lymphatic malfunction. Importantly, treatment with anti-inflammatory dexamethasone prevents the zippering seen in Mycoplasma pulmonis inflammation ¹³, suggesting utility for anti-inflammatory factors in promoting proper lymphatic junctional morphology and function during inflammation. In this study, we determine that IL-19 appears to have both short and long-term effects on hdLEC permeability and therefore hdLEC function. IL-19 induced rapid VE-cadherin phosphorylation, an indicator of LEC button formation. Additionally, in longer-term ECIS assays, IL-19 maintained hdLEC permeability for hours, which may be mediated by ongoing gene expression such as Angpt2. Treatment of hdLEC monolayers with oxLDL showed a dose-dependent decrease in paracellular permeability visualized by continuous zipper-like immunocytochemistry for VE-cadherin, consistent with recent studies indicating that oxLDL decreases transwell permeability of hdLECs in vitro and decreases lymphatic function in the liver in vivo ^57,58. Perhaps most importantly, hdLEC co-treatment with both rhIL-19 and oxLDL prevented the decreased permeability induced by oxLDL treatment, maintaining homeostatic barrier function comparable to control, as visualized by discontinuous button-like VE-cadherin junctional morphology. Overall, these data suggest IL-19 reduces oxLDL-associated junctional zippering, possibly promoting LEC functionality. Since removal of cholesterol from atherosclerotic plaque through RCT is a beneficial event in combating atherosclerosis, this suggests IL-19 may have a positive effect on lymphatic function to promote the reduction of atherosclerotic plaque development ^27,30,32,117.

On HFD, dKO mice have increased atherosclerosis compared with Ldlr^−/− control mice ⁵⁰, and is a useful model to determine the effects of loss of IL-19 on lymphatic function. In this study, Evans Blue dye clearance was significantly reduced in dKO mice compared with Ldlr^−/− control on HFD, suggesting impaired lymphatic function. This lymphatic malfunction was not seen in age-matched dKO mice on standard laboratory diet. In addition, HFD induced an increase in lymphatic vessel branch number in Ldlr^−/− mice while dKO mice had significantly less lymphatic branches than Ldlr^−/− mice on HFD. When analyzing VE-cadherin junctional morphology of initial lymphatic vessels, dKO mice on HFD had a significantly higher percentage of continuous zippered morphology when compared to dKO mice on standard laboratory diet or Ldlr^−/− mice on either diet. Together, this suggests that dKO mice have less of the typical discontinuous button-like morphology of VE-cadherin seen in initial lymphatic vessels, which is vital to their function in fluid and macromolecule clearance. This morphological change could be associated with the poor lymphatic clearance of Evans Blue dye seen in the dKO mice. Future studies using lymphatic-specific and IL-20Rα/β receptor subunit knockout mice could add mechanistic insight to these observations.

In summary, this study is the first to investigate the effects of IL-19, an immuno-modulatory cytokine, on lymphatic expansion and function. The observations that IL-19 is both lymphangiogenic and increases lymphatic function, together with the transcriptional program elicited, suggest that IL-19 could represent an endogenous factor that can increase functional lymphatic surface area and mediate important anti-atherogenic functions such as RCT.

Supplementary Material

Final Supplemental Date

Supplemental Data I. A. IL-19 stimulation of hdLEC does not result in VEGFR3 phosphorylation (Y1230). IL-19 treatment from 0–90 minutes shows non-specific bands while VEGFC treatment shows specific bands as expected around 130kDa and 170kDa. Blot shown is representative of 3 western blots performed on individual biological replicates. B. Representative western blot showing IL-19 knockdown with siRNA does not affect its receptor complex expression. C. Densitometric analysis of IL-20α and IL-20β protein expression, n=3 each, all analyzed by 2-way ANOVA followed by Tukey’s multiple comparisons test. All values are presented as the mean±SD.

Supplemental Data II. A. Representative image showing IL-19 increases hdLEC motility in a scratch migration assay after 20 hours of treatment. B. IL-19 does not induce significant hdLEC proliferation by cell number at day 3, indicating migration is independent of proliferation at this timepoint, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. C. IL-19 induces loop formation in hdLEC indicative of lymphatic network formation, n=3 each, all analyzed by 1-way ANOVA followed by Tukey’s multiple comparisons test. D. Percent change in scratch area at 25ng/mL and 50ng/mL rhIL-19, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. All values are presented as the mean±SD.

Supplemental Data III. A. IL-19 increases Prox1 mRNA stability in Actinomycin D-treated hdLEC, n=3, 2-way ANOVA followed by Šídák’s multiple comparisons test. B. Representative western blot showing IL-19 induced total FOXO1 protein expression over 48-hours. C. Densitometric analysis of total FOXO1 protein, n=3, 1-way ANOVA followed by Dunnett’s multiple comparisons test. D. Representative western blot showing IL-19-induced pFOXO1 protein expression. E. Densitometric analysis from showing IL-19 induction of pFOXO1 protein expression, n=3, 1-way ANOVA followed by Tukey’s multiple comparisons test. F. Representative western blot showing FOXO1 knockdown does not affect Prox1 expression but does reduce Angpt2 expression. G. Densitometric analysis of Prox1 and Angpt2 expression in FOXO1 siRNA knockdown hdLEC, n=3 each, all analyzed by 2-way ANOVA followed by Tukey’s multiple comparisons test. All values are presented as the mean±SD.

Highlights:

• IL-19 is expressed in and is pro-lymphangiogenic for hdLECs

• IL-19 increases expression of transcription factors, including Prox1

• IL-19 participates in LEC barrier function and permeability in vitro and in vivo