Human umbilical vein endothelial cells (HUVECs) were isolated from fresh human umbilical cord veins by collagenase I digestion according to the standard technique documented by Jaffe et al. (Takaoka et al., 2009). Isolated HUVECs were cultured with endothelial cell growth medium (ECM) with 5% fetal bovine serum (FBS), 1% penicillin and streptomycin, and 1% endothelial cell growth factor (EGF) under standard cell culture conditions (37°C, 5% CO₂). HUVECs of passages 3–5 were used for experiments. Once HUVECs reached 70–80% confluence, they were transfected with miRNA control or miR-19b mimic (transfection concentration: 30 pmol/ml, GenePharma, China) using Lipofectamine 2000 (Invitrogen) for 24 h. Then the medium was changed to serum-free medium, and the cells were exposed to hypoxic conditions (37°C, 3% O₂) for 12 h to generate modified EMPs. After cell collection, the production of EMP^control or EMP^miR-19b was performed as previously described (Jansen et al., 2013). Briefly, EMPs were isolated from HUVEC culture medium after centrifugation at 800 g for 15 min (to remove cells) and 12,500 g for 5 min (to remove debris). The supernatant was further centrifuged at 4°C at 20,500 g for 2.5 h. EMPs were collected from the sediment, and standard assessment of isolated EMPs was performed as previously described (Li et al., 2018).

All animal experiments were approved by the Medical Ethics Committee of Peking University People’s Hospital and carried out according to the National Institutes of Health guide for the care and use of laboratory animals (NIH Publications No. 8023, revised 1978). Six-to eight-week-old male ApoE^-/- mice were provided by Peking University Experimental Animal Center and were housed in pathogen-free barrier facilities under appropriate conditions with a temperature of 22 ± 2°C, humidity of 55 ± 5%, and 12-h light/12-h dark time cycle. After a 1-week acclimatization period with normal chow, mice were divided into two groups, one fed normal chow and the other fed a high-cholesterol atherogenic western diet (21% fat and 0.15% cholesterol, HFK Bioscience, China). The mice fed the high-cholesterol atherogenic diet were further subdivided into three groups that received injection of phosphate-buffered saline (PBS) as a control or different types of EMPs (1 × 10⁷ EMPs) through the tail vein three times per week throughout the experiment according to previous protocol (Li et al., 2018). In our study, we also monitored the miR-19b level in blood to guarantee the levels in EMP^miR-19b mice were higher than those in the control groups. These three groups included the PBS group (injected with PBS, as a blank control), the EMP^control group (injected with EMP^control, as a negative control), and the EMP^miR-19b group (injected with EMP^miR-19b). EMPs were injected over an 8-week study period, and blood, lymph and aortic arch tissues were collected at 4, 6, and 8 weeks.

For imaging of carotid arteries, the mice were anesthetized, and the hair was shaven from the anterior cervical skin. Ultrasound imaging was performed with mice in the supine position using a high-frequency ultrasound system (Vevo2020, VisualSonics). Carotid aorta pulse wave velocity (PWV) was determined with a 24-MHz transducer following the instructions of the device manufacturer. Briefly, pulse wave Doppler tracing was used to measure aortic flow velocity (V). Immediately thereafter, the aortic diameter (D) in the systolic and diastolic stages was measured on 700 frames-per second B-mode images of the carotid aorta in the EKV imaging mode. The carotid resistive index (RI) was calculated according to the peak systolic velocity (PSV) and end-diastolic velocity (EDV), using the formula: RI = [(PSV-EDV)/PSV].

The aortic arch tissues were fixed in 4% paraformaldehyde for more than 24 h, embedded in paraffin, and cut into 4.5-μm-thick sections. The sections were stained with hematoxylin and eosin (H&E) and Masson’s trichrome staining.

The 4.5-μm-thick paraffin-embedded aortic arch tissue sections were deparaffinized, and nonspecific reactions were blocked with goat serum for 15 min at 37°C. Then, the sections were incubated with primary polyclonal antibodies overnight at 4°C. The primary antibodies used included lymphatic vessel endothelial hyaluronan receptor-1 (LYVE1) mouse monoclonal antibody (Proteintech, United States) and CD68 rabbit polyclonal antibody (Proteintech). The following secondary antibodies were used for detection: goat anti-mouse immunoglobulin G (IgG) H&L (Alexa Fluor^® 488) and donkey anti-rabbit IgG H&L (Alexa Fluor^® 647) (both from Abcam, USA). Sections were counterstained with 4,6-diamidino-2-phenylindole (DAPI) for visualization of cell nuclei and mounted for analysis.

Total RNA was isolated from the aortic tissues, lymph, and lymphatic endothelial cells (LECs) using TRIzol Reagent (Thermo Fisher Scientific, United States). Briefly, add 750 µl TRIzol Reagent in each sample, all of them are equal in volume or weight, and exogenous cel-miRNA 39 was added in lymph samples as an internal control. Completely dissociated tissues and cells were incubated with trichloromethane with one-fifth of the volume of TRIzol after 5 min. The samples were mixed thoroughly and then centrifuged at 12000 g for 15 min at 4°C. An equal volume of isopropanol was added to the supernatant, and after centrifugation, the RNA was purified with anhydrous ethanol. Finally, the extracted RNA was diluted in diethylpyrocarbonate (DEPC)-treated water, and the concentration was determined. Reverse transcription quantitative real-time PCR (RT-PCR) for determination of mRNA and miRNA expression was performed on a LightCycler Run 5.32 Real-Time PCR System (Roche, Swiss) using SYBR Green detection chemistry and Taqman Universal Mix II. All samples were quantitated by the comparative CT method for relative quantitation of gene expressions. The expression of gapdh was used as the mRNA internal control, whereas U6 and miNA-39 were used as the miRNA internal controls for tissues and lymph, respectively. The following primer sequences were used: homo-tgfβRII: forward 5′-GCT TTG CTG AGG TCT ATA AGG C-3′, reverse 5′-GGT ACT CCT GTA GGT TGC CCT-3'; homo-gapdh: forward 5′-GAG TCA ACG GAT TTG GTC GT-3′, reverse 5′-GAC AAG CTT CCC GTT CTC AG -3'. The samples were analyzed in triplicate.

For lymphatic fluid collection, the mice were fasted for 12–14 h. At 1 h before collection, 200 μl soybean oil was administered by gavage to mice to make the cisterna chyli more visually obvious. The mice were anesthetized, and the abdomen was opened horizontally. Lymphatic flow was confirmed by the white lucent color in the intestinal trunk and the cisterna chyli under stereomicroscopy. The intestinal trunk is formed by the confluence of the efferent vessels of the cranial mesenteric and celiac nodes, and the single trunk enters into the cisterna chyli located along the abdominal vena cava and aorta on the cranial side of the renal veins. The cisterna chyli was cut with the bevel of a fine needle, and the lymphatic fluid was collected with a 24G catheter needle connected to a glass tube moistened with heparin. The lymphatic fluid was finally collected in an Eppendorf tube moistened with heparin for anticoagulation. The collected lymph was centrifuged at 1000 rpm for 5 min to generate a pellet of the cells present in the lymph. Then the supernatant was removed and saved for further analysis.

Evans blue dye was used to assess lymphatic vessel permeability through tracing the path of lymph through popliteal lymphatic vessels. Mice were anesthetized by isoflurane inhalation anesthesia, and skin was carefully removed from the legs. Following Evans blue intradermal injection in the footpad, popliteal collecting lymphatic vessels were visualized under a Leica M320 operating microscope (Leica). The effusion of Evans blue around the vessels marked the area of the leakage and was analyzed using ImageJ software (National Institutes of Health, United States).

Lymphatic vessel transport of molecular substances was assessed following injection of fluorescein isothiocyanate (FITC)-labeled dextran in the dermis of the footpad of the mouse (Fukasawa et al., 2021). After injection, we continuously detected the fluorescence intensity change in the popliteal lymph node by confocal microscopy (Leica SP5). Briefly, the mouse was placed in prone position on a heating pad at 40°C, and lower limb skin was removed, ensuring no blood vessels were severed as the skin was pulled away. Warm saline was administered to keep the limb tissue hydrated at all times. We first identified the location of lymph nodes under brightfield microscopy, and then 50 μl FITC-dextran (5 mg/ml) was injected in the footpad (Fukasawa et al., 2021). Once perfusion of the tracer appeared, the lymph node was recorded for 45 min to continuously observe the fluorescence intensity change in the popliteal lymph node from the first appearance of fluorescence until the fluorescence intensity declined. To quantify the fluorescence intensity, we traced a line of equivalent length at three different regions of interest (ROIs) along the lymph node, and the average value was used for analysis.

Primary human dermal lymphatic endothelial cells (HDLECs, Promocell) were cultured according to the manufacturer’s protocol (Lonza) in EBM-2 medium containing EGM-2 MV SingleQuots. Cells of passage 5–6 were seeded in culture dishes at 80% confluence and transfected with miRNA control, miR-19b mimic, miRNA inhibitor control, or miR-19b inhibitor (GenePharma, China) using Lipofectamine 2000 (Invitrogen) for 24 h.

The cell migration assay was performed using 24-well Transwell chambers. HDLEC suspension (100 μl ECMV2 without FBS) from each group was added to the upper chamber (1 × 10⁵ cells/well), whiles 600 μl ECMV2 with 5% FBS was added to the bottom chamber. After 12 h of incubation at 37°C, cells on the upper membrane surfaces were removed, and migrated cells were washed three times with PBS at 37°C. The migrated cells were fixed with 4% paraformaldehyde for 15 min and stained with crystal violet for 20 min. The number of migrated cells was determined by counting in five fields with a uniform cell distribution under an inverted microscope. The assay was repeated three times, and the average migration rate was calculated.

For use in the tube formation assay, Matrigel was thawed at 4°C overnight. Then each well of a prechilled 48-well plate was coated with 150 μl Matrigel and incubated for 1 h at 37°C in a 5% CO₂ incubator. Then, 200 μl cell suspension was added to each well (2 × 10⁵ cells/well). After incubation of the different groups of cells under specified culture conditions for 4–6 h, tube-like structures and their arrangement, amount, and degree were evaluated in photographs taken of five randomly selected five fields under a microscope. The tube number was quantified using ImageJ software.

Data are expressed as the mean ± standard error of the mean (SEM). Statistical significance was evaluated by unpaired t-test or, for multiple comparisons, one-way analysis of variance (ANOVA) using appropriate corrections when data were not normally distributed or for unequal variances. All calculations were done with GraphPad Prism software (GraphPad Software, La Jolla, CA, United States), and p values < 0.05 were considered statistically significant.

We first observe pathological changes in the lymphatic system of atherosclerotic mice after 4 weeks of high-fat diet feeding. The transport function of lymphatic vessels was measured via confocal microscopy, and the permeability of lymphatic vessels was assessed by Evans blue staining in each group. Fluorescence images and videos revealed that, upon injection, FITC-labeled dextran was taken up by peripheral lymphatic vessels of the legs and then drained into the popliteal fossa lymph nodes. Comparison of the different groups showed that the time to reach maximum intensity in high-fat diet-fed mice was significantly extended and the maximum intensity significantly reduced compared with those in mice fed the control diet (p < 0.05, Figure 1A). The results of Evans blue staining revealed that the lymphatic vessels in the atherosclerotic mice were thinner, and exudation of blue dye from these vessels was increased compared with the control (Figure 1C). Next, the residual amount of Evans blue dye and the amount of Evans blue dye in the opposite leg lymph node were measured at 10 min after injection. The results showed that the amount of residual dye in atherosclerotic mice was greater than that in mice fed a normal diet (Figure 1D), whereas the amount of Evans blue dye in the opposite leg lymph node was less than that in the normal diet-fed mice (Figure 1E). Both confocal microscopy and Evans blue staining indicated that the transport function and permeability of the lymphatic system were disrupted in atherosclerotic mice after 4 weeks of high-fat diet feeding.

We also detected the level of miRNA-19b in aortic tissue and blood. Quantitative PCR analysis revealed that the level of miRNA-19b was increased in the aortic tissue and blood of high-fat diet-fed mice compared with normal diet-fed mice (Figures 1F,G).

To investigate the role of miR-19b-containing EMPs in the development of atherosclerosis, 6-week-old ApoE^-/- mice fed a high-fat diet were treated with PBS, EMP^control, or EMP^miR-19b three times per week by tail vein injection (Figure 2A). After the 4, 6, or 8 weeks of treatment, miR-19b levels in lymph and aortic tissue samples were measured. The RT-PCR results show that in each sampling point, the miR-19b levels in both aortic tissue samples (Figures 2B–D) and lymph fluid (Figures 2E–G) were elevated in EMP^miR-19b-treated mice compared with PBS- and EMP^control-treated controls, indicating that systemic EMP^miR-19b treatment could increase miR-19b expression in lymphatic circulation and arterial tissue at the same time.

We then further assessed the effect of increased miR-19b expression on the pathological process of atherosclerosis. Firstly, carotid ultrasound showed that after 4, 6, and 8 weeks of treatment, the RI of the carotid artery in the EMP^miR-19b group mice was significantly elevated compared with that in the other control groups (Figure 3A). Meanwhile, the diameter of the carotid artery in the systolic stage showed a decreasing trend in the EMP^miR-19b group after 4 and 6 weeks of treatment. Then after 8 weeks of treatment, compared with the PBS- and EMP^control-treated groups, the EMP^miR-19b group had a significantly decreased carotid artery diameter in the systolic stage (Figure 3B). All of the ultrasonography results strongly suggested that EMP^miR-19b accelerated the development of atherosclerosis.

The results of histopathological analyses supported the findings from imaging examinations. After 6 weeks of high-fat diet feeding, atherosclerotic plaque began to form in the aorta, especially in the aortic arch. After 6 weeks of PBS or EMP treatment, atherosclerotic plaque had formed in the aortic arch in all of three treatment groups; however, in the EMP^miR-19b group, the size of the plaque was larger and the distribution of plaque was much more extensive compared with results for the other groups, and these differences were even clearer after 8 weeks of treatment (Figure 3C). Cross-sectional analysis revealed that compared with PBS and EMP^control treatment, EMP^miR-19b treatment significantly increased the size of H&E-stained atherosclerotic plaque and the composition of plaque was more complex with a thinner fibrous cap and profoundly increased lipid accumulation (Figures 3D,E, p < 0.05). Together, the results of histopathological and ultrasound examinations suggested that treatment with EMP^miR-19b promoted the development of aortic atherosclerosis.

To understand whether EMP^miR-19b treatment could impact the function of lymphatic system, we further evaluated the transport function and permeability of lymphatic vessels. We first tested whether the lymphatic trafficking capacity was damaged by EMP^miR-19b treatment. Based on the current literature, we decided to inject FITC-labeled dextran into the foot pad of the mice and continually monitor the dynamic change in fluorescence intensity in popliteal fossa lymph nodes by confocal microscopy (Figure 4A). Fluorescence imaging revealed that, upon injection, FITC-labeled dextran was taken up by peripheral lymphatic vessels of the legs and then drained into the popliteal fossa lymph nodes. The results showed that the fluorescence intensity increased over time in both lymph nodes; however, at each observation time point, the time to reach maximum intensity in the EMP^miR-19b treatment group was progressively extended and the maximum intensity significantly reduced, with significant differences compared with the control groups (p < 0.05). The variation in the differences observed between the EMP^miR-19b group and the two control groups at the 4- and 6-week points indicate that lymphatic system dysfunction occurs in the early stages of disease.

To assess lymphatic vessel permeability, Evans blue dye was also injected in the mouse footpad, and then dye leakage and the size of the lymphatic vessels were quantified. Our results revealed that EMP^miR-19b treatment had no significant effect on dye leakage, but lymphatic vessel thickness was decreased after EMP^miR-19b treatment, compared with the thicknesses in the control groups (Figure 4E).

Both confocal microscopy and Evans blue staining results indicated that EMP^miR-19b treatment damaged the function of the lymphatic system in ApoE^−/− mice.

To understand whether EMP^miR-19b treatment could impact the distribution of lymphatic vessels along atherosclerotic arteries, we applied immunofluorescent staining for LYVE1, a relatively specific marker for lymphatic vessels, to observe lymphatic vessels with arteries in the three groups after 8 weeks of treatment (Figure 5). CD68 staining also was performed, because some CD68⁺ macrophages can also express LYVE1 to interfere with the results. Thus, in the double immunofluorescence staining images, the LYVE1⁺/CD68^- tubular structures are identified as lymphatic vessels (Schlereth et al., 2014). Fluorescence images revealed fewer lymphatic vessels with atherosclerotic arteries in the EMP^miR-19b group than in the two control groups (p < 0.05). This finding suggests that EMP^miR-19b treatment may inhibit the distribution of lymphatic vessels in the pathogenesis of atherosclerosis.

To further understand the effect of miRNA-19b on the lymphatic system, in vitro cell experiments were carried out. Transwell migration and tube formation assays were performed to examine the effects of miR-19b on LEC migration and lymphangiogenesis, respectively. In the Transwell migration assay, comparison of the numbers of migrated cells in each group showed that miR-19b inhibited LEC migration from the upper chamber to the reverse side. However, miR-19b inhibitor promoted LEC migration compared with the negative control (Figures 6A,B, p < 0.05). The results of the tube formation assay showed that the number of total branching points in the miR-19b mimic group was significantly less than that in the negative control group, and the number of loops formed was reduced by 0.8-fold compared to the control group. In contrast, the number of total branching points in the miR-19b inhibitor group was significantly greater than that in the corresponding negative control group, and the number of loops formed was increased by 1.2-fold as compared to the control (Figures 6C–E, p < 0.05).

To identify a potential target gene of miR-19b associated with the distribution and function of lymphatic vessels, we performed an in silico analysis using three different microRNA target prediction algorithms, TargetScan (http://www.targetscan.org/), miRDB (http://mirdb.org/), and miRWalk2.0 (http://zmf.umm.uni-heidelberg.de/apps/zmf/mirwalk2/index.html). The results identified TGFβRII as a potential target commonly predicted by the three algorithms (Figures 7A,B).

To determine whether TGFβRII is a novel target of miR-19b, gain- and loss-of-function experiments were performed by transfecting LECs for 24 h with miR-19b mimic or inhibitor. RT-qPCR revealed that miR-19b was successfully overexpressed by LECs and reached peak expression with a transfection concentration of 70 pmol/ml (Figure 7C) and miR-19b expression was inhibited and reached a minimum value with a transfection concentration of 200 pmol/ml (Figure 7D). Using these conditions, the experiment showed that upregulation of miR-19b inhibited TGFβRII expression by 50% at the mRNA level (Figure 7E), whereas downregulation of miR-19b increased TGFβRII expression by 1.5-fold at the mRNA level (Figure 7E).