In normal settings, the lymphatic vasculature is inherently antithrombotic due to the high level of fibrinolytic activity in the LECs and the relatively low levels of prothrombotic factors in lymph (Lippi et al., 2012). Lymphatic thrombosis is, therefore, a rare event and implies that there has been a shift in the balance of prothrombotic factors and fibrinolytic activity. We determined whether cigarette smoke extract (CSE) alters the fibrinolytic activity of LECs using an in vitro culture system with human lung microvascular endothelial cells to determine the effects of cigarette smoke extract on LECs. Staining for the human lymphatic markers PROX1 and podoplanin verified that these cells maintained their lymphatic identity in culture (Figure 1A). We performed fibrin zymography as previously described (Kim et al., 1998; Choi and Kim, 2000) using supernatants from LECs treated with the CSE compared to those from controls. We found that supernatants from control LECs demonstrated robust fibrinolytic activity, as expected, as demonstrated by the bands on the zymogram gel formed by protease activity (Figure 1B). However, supernatants from CSE-treated LECs showed decreased protease activity in the zymogram (Figure 1B). The fibrinolytic state of LECs is driven, in part, by relatively low levels of plasminogen activator inhibitor-1 (PAI-1), which inhibits the activity of plasminogen activators tPA and uPA to create a prothrombotic state (Lippi et al., 2012; Sillen and Declerck, 2020). We found increased expression of PAI-1 in LECs treated with CSE compared to those of the controls by both Western blot and quantitative PCR (Figures 1C, D). CSE treatment also led to a small but not statistically significant decrease in the expression of the fibrinolytic protease tPA (Figure 1E), but it did not alter the expression of other endothelial homeostatic factors such as endothelial protein C receptor (EPCR) or thrombomodulin (THBD) (Supplemental Figure S1). We then determined whether the increase in expression of PAI-1 in LECs after CSE exposure is due to oxidative stress. We found that treatment of LECs with the antioxidant N-acetylcysteine partially inhibited the upregulation of PAI-1 after CSE exposure (Figure 1F), suggesting that oxidative stress is indeed a cause of decreased fibrinolytic activity in LECs.

Our previous work has shown that CS exposure results in increased thrombin in lymph draining in the lung lymphatics, which may be due to the overall increase in systemic thrombin that is seen with CS exposure in COPD (Jankowski et al., 2011; Undas et al., 2011). Given the effects of thrombin on endothelial cell function (Rabiet et al., 1994; Rabiet et al., 1996; O'Brien et al., 2000; Minami et al., 2004), we hypothesized that the lung lymphatics may be affected not only by CS itself but also by the increased presence of thrombin that LECs are exposed to after CS exposure. We treated LECs with CSE and thrombin and assessed the expression of thrombotic markers and leukocyte adhesion molecules. We found that CSE alone led to an increased expression of the tissue factor pathway inhibitor (TFPI) and the thrombin receptor protease activated receptor-1 (PAR-1) (Figures 2A–C). While CSE increased the expression of TFPI, LECs exposed to thrombin at decreased expression of TFPI (Figures 2A, B). In addition, LECs treated with thrombin had an increased expression of VCAM-1 and ICAM-1 compared to those treated with CSE alone (Figures 2A, D, and E).

We next determined whether the changes we have observed in vitro are also present in LECs in mice after CS exposure. To determine the short- and long-term effects of CS on the lymphatic endothelium in vivo, we used a novel technique of lung LEC isolation for downstream analysis. Lymphatic reporter mice, in which all LECs are labeled by GFP (Prox1-EGFP) (Choi et al., 2011), were exposed to CS using a whole-body system for acute (8 weeks) or chronic (24 weeks or more) time points, and their LECs were compared to lung LECs from age-matched control mice. Immediately prior to sacrifice, the mice were injected with fluorescently labeled isolectin for intravital labeling of all endothelial cells, and single-cell suspensions from lung homogenates were stained for FACS with negative selection for the epithelial marker EpCAM and a hematopoietic lineage marker cocktail, as well as positive selection for GFP and CD31 (Figure 3A). This technique led to a pure population of LECs that was roughly 0.2% of total lung cells, which is in line with previous single-cell data sets (Schupp et al., 2021).

We analyzed LECs after acute and chronic CS exposure using RNA sequencing and compared it to lung LECs from age-matched controls. We found significant alterations in the transcriptome of LECs after both acute and chronic CS exposure (Figures 3B–E). Interestingly, we did not observe many transcriptional changes in control room air LECs isolated at 8 weeks compared to those isolated at greater than 24 weeks (Figure 3B), suggesting that CS confers more changes in LECs than age alone. In addition, while there were many transcriptional changes that were unique to acute and chronic CS exposure, there were also many genes that were altered at both time points (Figures 3C–E). Notably, pathways associated with cell adhesion and leukocyte migration as well as those that are predicted to play a role in inflammation and immune cell recruitment were upregulated in CS-exposed LECs at both acute and chronic exposure time points (Figures 4A, B). In addition, we observed the upregulation of pathways associated with interactions with the extracellular matrix, particularly after chronic CS exposure (Figures 4A, B). Key genes crucial for leukocyte migration and adhesion, including Vcam1, Icam1, endothelial cell-specific adhesion molecule (Esam), and endothelial selectin (Sele), were also significantly upregulated in LECs from CS-exposed mice compared to control LECs (Figure 4C). Additionally, we observed CS-induced upregulation of chemokines and markers of inflammation, including platelet-derived growth factor B (Pdgfb), matrix metalloproteinase 16 (Mmp16), and interleukin receptor-17 (IL-17r), in CS-exposed LECs compared to control LECs (Figure 4C). Interestingly, we observed a decreased expression of Tfpi in LECs from CS-exposed mice, similar to what was seen in LECs exposed to thrombin in vitro (Figure 4C). Immunohistochemistry of lung tissue from mice exposed to CS for 8 weeks showed increased staining of VCAM-1 on the lymphatic vessels compared to room air control mice (Figures 5A–C). Thus, in vivo transcriptomic analysis supports our in vitro data and suggests that LECs in CS-exposed mice undergo transcriptional changes that may result from the effects of CS-induced thrombin in the lung lymphatic vasculature. We studied publicly available single-cell RNA sequencing data of lung tissue obtained from the explanted lungs of patients with advanced COPD requiring lung transplant and control donor lungs (Sauler et al., 2022). Interestingly, we also found a relatively high level of expression of Tfpi in LECs, which was then decreased in LECs from COPD patients, identical to what we observed in LECs from CS-exposed mice (Supplementary Figure S2). The expression of many of the genes that we observed using RNA sequencing of isolated LECs from mice was not detected in LECs from this dataset of whole lung tissue, likely due to scarcity of LECs in the lungs making detection of these transcripts difficult without specific isolation of these cells. However, we did find an increased expression of Mmp16 and Pdgfb in LECs from COPD patients compared to those from controls, though the expression of these genes was low.

We next sought to determine whether some of the changes we observed in lung LECs from mice were also present in the lymphatics from patients with COPD using immunohistochemical analysis on lung tissue samples from patients with end-stage COPD compared to lung tissue from control nonsmokers without lung disease. Staining for the human lymphatic marker podoplanin and VCAM-1 revealed a robust expression of VCAM-1 in the lymphatics in end-stage COPD patients that was not present in control lymphatics (Figures 6A–D). Interestingly, we also observed that VCAM-1 was more highly expressed throughout the lung tissue from COPD patients compared to controls, likely reflecting the known upregulation of VCAM-1 in the blood endothelium in COPD (Borek et al., 2023). We also investigated whether the expression of PDGFB was increased in lung lymphatics in COPD patients compared to controls, as this was seen in CS-exposed mice at acute and chronic exposure time points by RNA sequencing. Using immunohistochemistry, we found evidence of increased staining for PDGFB around the lymphatic vessels in the lungs of COPD patients (Figures 7E–H) compared to control patients (Figures 7A–D). This PDGFB staining was associated with the basement membrane and extracellular matrix of the lymphatic vessels, similar to previous reports (Wang et al., 2017) where PDGFB retention is necessary for its role in accessory cell recruitment to these vessels. Normal lung lymphatics have minimal smooth muscle cell (SMC) coverage, as opposed to lymphatics in other organs, and ectopic recruitment of SMC lymphatics is a marker of dysfunction in these vessels (Meinecke et al., 2012; Sweet et al., 2015; Reed et al., 2019). Interestingly, we found increased staining for smooth muscle actin-positive SMCs associated to the lung lymphatics in lung tissue from patients with COPD compared to controls (Figures 7I–P). Single-cell RNA sequencing data suggested that PDGFB is increased in the lungs in COPD, though LECs are likely not the major source (Supplementary Figure S2). These data suggest that increased PDGFB driven by LECs and other cell types may drive SMC recruitment to these vessels in COPD, though this has not been definitely shown.

Prox1-EGFP mice have been previously described (Choi et al., 2011). C57Bl/6 or Prox1-EGFP mice were exposed to cigarette smoke for 8 weeks (acute) or ≥24 weeks [using a whole-body exposure system (TE-10) (supplied by Teague Enterprises)] with 3R4F composition cigarettes (University of Kentucky Center for Tobacco Reference Products). Mice were exposed to CS (∼150 mg/m³) 5 days a week for abovementioned time points. Control mice exposed to room air alone were age-matched and identically housed in the same room and on the same rack. Mice were housed in the Weill Cornell animal facility in 12/12 h light/dark cycles with ad libitum access to water and food. At the respective time points, CS-exposed C57Bl/6 mice or Prox1-EGFP mice and age-matched room air control mice were identically housed in the animal facility. We used both male and female mice in our experimental and control groups.

Human lung microvascular lymphatic endothelial cells (Lonza) were cultured in EGM-2 media with a EGM-2 Endothelial Cell Growth Medium-2 BulletKit (Lonza, CC-3162) at 37°C and 5% CO₂ to ensure confluence and then plated at a density of 187,000 cells per well. The cells were then starved in EGM-2 without added supplemental growth factors or serum. For cells treated with cigarette smoke extract (CSE) alone, 2% CSE in DMEM was then added to EGM-2 media without supplemental growth factors overnight. For cells treated with CSE and thrombin, 2% CSE was added to cells initially for 6 h. After 6 h, 1U/μL thrombin (Sigma, T4393) was added to the wells for 10 min. All the wells were then washed with serum-free media, and media with 2% CSE was again added to cells and incubated overnight. Cells treated with thrombin alone were treated for 10 min, washed two times in serum-free media, and incubated overnight. After overnight incubation for all wells, cell lysates and supernatants were collected using RIPA and protease inhibitors the next morning for protein and RNA isolation. For co-incubation with N-acetylcysteine (NAC), LECs were cultured and plated as discussed above. All the wells were washed with EGM-2 media without added supplemental growth factors or serum. The control cells were left in serum-free media overnight. For cells treated with CSE alone, 2% CSE in serum-free media was then added and incubated overnight. For cells treated with CSE and NAC, a freshly prepared 5 µM NAC solution (Sigma, A9165) in serum-free media was added to the wells for 2 h. CSE (2%) in serum-free media was then added to these wells and incubated overnight. After overnight incubation for all wells, cell lysates and supernatants were collected using RIPA and protease inhibitors the next morning for protein and RNA isolation.

Zymography was performed using modifications of previously described protocols (Kim et al., 1998; Choi and Kim, 2000). The supernatants obtained from both untreated (control) and 2% CSE-treated LECs were run on a 12% SDS-PAGE gel containing thrombin and fibrinogen, with tissue plasminogen activator (tPA) run as a positive control for fibrinolysis within the gel. The gel was washed in Triton-X and incubated overnight at 37° in a CaCl₂-containing Tris-HCL buffer. Staining was performed with Coomassie Blue, and the gel was de-stained in methanol and acetic acid to reveal areas of fibrin cleavage.

Western blotting was performed according to the standard protocols and probed with PAI-1 (R&D Systems, MAB1786), TFPI (Abcam, ab180619), PAR-1 (Sigma, sab4500823), VCAM1 (Novus, NBP2-29413), ICAM1 (Novus, NBP2-67518), and GAPDH (Santa Cruz Biotechnology, SC365062). Blots were imaged using a Konica Minolta SRX-101A Medical film processor and quantified with ImageJ software. The scanned Western blot images were imported into the ImageJ software. Regions of interest (ROIs) were defined around the bands corresponding to the protein of interest and the reference protein, GAPDH, by selecting the “rectangle tool.” Densitometric analysis was then conducted with the ImageJ using the “measure” option to quantify the intensities of these bands. The obtained target protein intensities were then normalized to the corresponding GAPDH bands, which served as an internal control.

Cell lysates from HMVLECs were harvested and homogenized for RNA isolation. Total RNA from cell lysates was isolated using an RNEasy Kit (Qiagen). cDNA was made according to the manufacturer’s instructions using the Superscript III First-Strand Synthesis System (Invitrogen). For qPCR reactions, cDNA was diluted at a ratio of 1:5. Each qPCR reaction was performed in triplicate. Analysis of gene expression was performed using SYBR Green PCR Master Mix (Applied Biosystems) and Quant Studio 6 Real-Time PCR System. qPCR gene expression analysis was carried out using the comparative Ct method (ΔCT) using GAPDH as the reference housekeeping gene.

For isolation of lung LECs, Prox1-EGFP reporter mice, in which all LECs are labeled with GFP, were injected with Alexa Fluor 647-conjugated isolectin (Invitrogen) for 12 min just prior to sacrifice. The lungs were then harvested and digested using collagenase and dispase in HBSS to generate a single-cell suspension for FACS, with additional staining using antibodies for CD31 (BioLegend 102418), lineage cocktail (BioLegend 133311), and EpCAM (BioLegend 118225). LECs were sorted using positive selection for GFP and CD31. Negative selection for EpCAM and a hematopoietic lineage marker cocktail was also performed. Total RNA from the LECs was isolated using a Qiagen RNeasy Mini Kit. For bulk RNA sequencing of LECs, at least 20 ng of total RNA from freshly harvested LECs was used for each replicate. For the 8-week CS time point, LECs were pooled from mice of the same experimental condition (n = 2–5) to have sufficient cell numbers and RNA quantity for analysis. Samples from the chronic CS exposure time points were not pooled. Libraries were prepared using the SMART-Seq v4 kit (Takara Bio), and bulk RNA sequencing was performed using Illumina NovaSeq 6000. Sample files were checked for sequence quality (FastQC v 0.11.9) and processed using the Digital Expression Explorer 2 (DEE2) (Ziemann et al., 2019) workflow. Adapter trimming was performed with Skewer (v0.2.2) (Jiang et al., 2014). Further quality control was carried out with Minion, part of the Kraken package (Davis et al., 2013). The resultant filtered reads were mapped to the mouse reference genome GRCm38/mm10 using the STAR aligner (Dobin et al., 2013) and gene-wise expression counts generated using the “-quantMode GeneCounts” parameter. After further filtering and quality control, the R package edgeR (Robinson et al., 2010) was used to calculate TMM normalization factors. FPKM and Log2 counts per million (CPM) matrices were quantified using these factors to normalize for library size. Principal component analysis and Pearson correlation values were quantified with the resulting Log2 CPM values. Gene enrichment analysis was run using g:profiler (Liao et al., 2019). Volcano plots were generated using the EnhancedVolcano R library (Blighe, K., Rana, S., and Lewis, M., 2018, https://github.com/kevinblighe/EnhancedVolcano).

Slides of de-identified human lung tissue from patients with end-stage COPD or nonsmokers were collected and banked at Washington University School of Medicine under approved protocols, prior to use for immunohistochemical staining. COPD lung tissues (n = 10) were collected at the time of lung transplantation (WU IRB 201103213), and nonsmoker control tissues were collected from decedents that were not usable for lung transplantation at the Mid-America Transplant Organ Procurement Organization (n = 5) in St. Louis, Missouri. Lung tissues were formalin-fixed and paraffin-embedded in blocks. Fresh slides were cut prior to staining. Data mining of human single-cell RNA sequencing was performed by studying the publicly available data set generated by Sauler et al. (2022) (https://copdcellatlas.com) using the gene explorer and multi-gene query tools.

Mice were sacrificed with CO₂ exposure, and prior to harvest, the lungs were inflated with 4% PFA at a constant pressure of 25 cm H₂O. The lungs were fixed in 4% PFA overnight at 4°. The lung tissue was then dehydrated by immersing it in a series of ethanol solutions of increasing concentrations and embedded in paraffin for sectioning. Six-micrometer sections were immunostained with antibodies for VEGFR3 (R&D Systems, AF743) and VCAM (Abcam, ab134047). Human lung tissue was stained with antibodies for podoplanin (D240, BioLegend, 75782-960) and VCAM (Abcam, ab134047) overnight at 4°. After washing, the slides were incubated with Alexa Fluor-conjugated secondary antibodies for 1 h at room temperature. Slides were treated with DAPI-containing Vectashield that was used on the tissue sections, and a cover slip was applied. Secondary antibodies alone were used on the negative control slides to check for autofluorescence of lung tissue. Immunofluorescence was performed either using a Nikon Eclipse microscope and NIS Elements software or a Zeiss SP8 confocal microscope and Leica software. Image analysis was performed using ImageJ.

Data are expressed as the mean ± SD. Statistical significance was determined by unpaired, 2-tailed Student’s t-test or ANOVA using GraphPad Prism software. p values of less than 0.05 were considered statistically significant.

All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of Weill Cornell Medicine and performed in accordance with relevant guidelines and regulations.