Pulmonary lymphangiogenesis in influenza pneumonia.
To investigate the pulmonary lymphatic vessel responses to influenza infection, we intratracheally infected the left lobe of mice with PR8 influenza and harvested left lungs at 3-, 7- and 21-days post-infection (dpi) for paraffin-embedding and sectioning. Mice were weighed daily during the time-course of infection to assess morbidity (Supp. Figure 1). We then stained for VEGFR3 in lung sections of influenza-infected and control mice to identify lymphatic vessels. We observed a significant and sustained enlargement in the diameter of VEGFR3-positive vessels in the lung during influenza infection as soon as 3 dpi and until at least 21 dpi (Fig. 1a and b). This indicated that the lymphatic vessels are dilated in response to influenza infection.
To determine whether lymphangiogenesis accompanied vessel dilation, we quantified the lung LEC population by labeling LEC nuclei during a time-course of influenza infection. In order to do this, we stained for PROX1 at 3, 7 and 21 dpi and enumerated LECs in each histological section comparing influenza-infected lungs with controls. This experiment revealed a doubling of LECs by 7 dpi that continued to increase through 21 dpi (Fig. 2a and b). Our findings indicate that influenza infection not only leads to the enlargement of existing lymphatic vessels but also stimulates the expansion of LEC number.
Pulmonary lymphangiogenesis in influenza is driven by LEC proliferation.
To identify the source of new LECs in the lymphatic network in influenza-infected lungs, we employed two independent approaches. Firstly, to label nascent DNA incorporation, 5-Ethynyl-2-deoxyuridine (EdU) was administered to mice during influenza infection. Lungs were harvested at 7 dpi and immunofluorescently co-stained for PROX1 and EdU (Fig. 3a and b). While proliferating LECs were very rarely observed in control tissues, approximately 20% of LECs in influenza-infected lungs had incorporated EdU by 7 dpi (3b). Secondly, we investigated the possibility that an exogenous progenitor cell might play a role in influenza-induced lymphangiogenesis. To address this question, tamoxifen-induced LEC lineage labeling was performed in PROX1-CreERT2/tdTomato mice prior to influenza infection. The proportion of tdTomato positive LECs observed during influenza infection was then compared to control lungs. Despite the increased number of LECs observed during influenza (data not shown), there was a similar proportion of lineage labeled cells at 7 dpi (Fig. 3c and d). These findings demonstrate that PROX1-negative progenitor cells are not contributing to lymphangiogenesis during influenza. However, we cannot exclude the possibility of a PROX1-positive progenitor. Notably, EdU uptake in LECs was not observed in the liver, heart, or esophageal tissues, indicating that LEC proliferation during influenza was specific to the lungs (data not shown). Taken together, these results suggest endogenous lung LEC proliferation in response to influenza infection contributes to LEC expansion.
Lymphatic transport in influenza pneumonia.
Consistent with published literature, influenza infection leads to pronounced inflammation, and pulmonary edema. In this regard, we observed significantly higher lung wet-to-dry weight ratios during influenza infection as compared to controls, indicative of pulmonary edema (Fig. 4a). To investigate the functional properties of the expanded lymphatic network, we instilled a 10 kDa fluorescent dextran molecule into the left lung airspaces and measured fluorescence in the lung-draining mediastinal lymph node (mLN) (Fig. 4b and c). We observed a significantly higher fluorescent signal in the mLNs of influenza-infected mice as compared to control at 15 minutes. There was minimal fluorescent signal detected in non-draining inguinal LNs indicating local lymphatic drainage as opposed to blood vessel drainage was primarily responsible for transport of dextran (data not shown). Collectively, these findings indicate that the expanded lymphatic network in the influenza-infected lung is associated with enhanced lymphatic transport.
Role of Hippo signaling during influenza-induced pulmonary lymphangiogenesis.
The Hippo pathway is a fundamental mediator of cell proliferation during development and is activated in response to injury and mechanical cues. Notably, Hippo signaling in LECs is critical for lymphangiogenesis and lymphatic patterning during embryonic development (23, 24). To determine the role for Hippo signaling in LECs during influenza-induced lymphangiogenesis, we utilized a previously published model of Hippo pathway deletion in LECs (Prox1-CreERT2/Yap1(YAP)-fl/Wwtr1(TAZ)-fl, hereafter YAP/TAZ△LEC). In all experiments, Cre(-) littermates given tamoxifen were used as controls.
First, we validated YAP and TAZ depletion in LECs using immunofluorescent staining of lung histologic sections and by validating the presence of the floxed YAP and TAZ alleles after tamoxifen administration (Supp. Figures 2 and 3). To confirm efficient PROX1-Cre-mediated recombination, we analyzed tdTomato expression in LECs in mice at least 2 weeks after tamoxifen administration to Prox1-CreERT2 / TdTomato mice (Supp. Figure 4). Next, we analyzed influenza-infected lungs for differences in histologic lymphatic phenotype, including lymphatic vessel diameter measurement and LEC enumeration as previously described. We identified no significant differences in these two parameters between Cre(+) and Cre(-) YAP/TAZ△LEC littermates, either at baseline, 7 or 16 dpi (Fig. 5a-f). Similarly, no differences in Prox1 or Flt4 mRNA were observed in whole lung homogenates obtained from either Cre(+) or Cre(-) YAP/TAZ△LEC littermates at 7 dpi (Fig. 5g).
To determine whether there were differences in pulmonary edema in the context of Hippo deletions, we measured lung wet-to-dry weight ratios during influenza as well as control conditions, and found no difference between Cre(+) and Cre(-) YAP/TAZ△LEC littermates (Fig. 5h).
The dextran transport assay was utilized to interrogate the functionality of lymphatics for passive drainage in uninfected lungs. In these experiments, no significant difference was observed in fluorescent signal measured in the lung-draining mLNs of Cre(+) vs Cre(-) YAP/TAZ△LEC littermates at baseline or at 7dpi (Fig. 5i and j). Collectively, the targeted deletion of YAP and TAZ in LECs did not affect lymphangiogenesis or lung lymphatic drainage at baseline or during influenza pneumonia.
Infection severity and inflammatory response after deletion of YAP and TAZ in LECs.
Infection severity (28) and inflammatory responses (9, 29) are known to be affected by aberrant lymphatic function. To address this issue, we first compared survival and weight loss curves after influenza pneumonia in Cre(+) and Cre(-) YAP/TAZ△LEC littermates. No mice reached the humane endpoint in either group, and we found no differences in weight change during or after influenza infection between the two Cre genotypes (Fig. 6a). We then assessed infection severity by plaque assay and qRT-PCR for influenza nucleoprotein (NP) mRNA in murine lungs at 7 dpi and found no difference in these readouts between Cre genotypes (Fig. 6b and c). We also characterized the inflammatory response to influenza in the lung at 7 and 16 dpi in both Cre genotypes. These timepoints mark two critical phases: the lowest point of weight loss and the subsequent return to baseline weight after influenza infection. Detailed inflammation scoring and histopathologic characterization was provided by a veterinary pathologist who was blinded to genotypes. In summary, there were no apparent differences due to YAP/TAZ-deletion in morphological histopathology as assessed by the percent area of consolidation or semiquantitative pathology score (Fig. 6d and e, Supp. Tables 1 and 2).
Table 1
List of antibodies used for histologic staining.
Antibody | Application | Source | Catalog # | Dilution |
Rabbit ⍺-PROX1 (EPR19273) | Immunostaining | Abcam | ab199359 | 1:500 for immunostaining |
Goat ⍺-VEGFR3 | Immunostaining | R&D Systems | AF743 | 1:500 |
Rabbit ⍺-TAZ (E8E9G) | Immunostaining | Cell Signaling Technologies | 83669 | 1:100 |
Rabbit ⍺-YAP (D8H1X, XP) | Immunostaining | Cell Signaling Technologies | 14074 | 1:100 |
Goat ⍺-PROX1 | Immunostaining | R&D Systems | AF2727 | 1:300 in conjunction with YAP/TAZ staining |
Table 2
List of antibodies used for flow cytometry.
Cell types | Target | Fluorophore | Concentration / Dilution | Source and Catalog # |
Live / dead | 7-AAD | 7-AAD | 1:60 | BD Biosci. 51-68981e |
Leukocytes | CD45 (clone 30-F11) | Per-CP 5.5 | 0.2 mg/mL | Biolegend 103131 |
| CD45.2 (IV) (clone 104) | BUV737 | 0.2 mg/mL | BD Biosci. 612778 |
T cells | CD4 | APC-Cy7 | 0.2 mg/mL | Biolegend 100413 |
| CD8a | BV510 | 0.05 mg/mL | Biolegend 100751 |
B cells | CD19 | BUV395 | 0.2 mg/mL | BD Biosci. 563557 |
Dendritic cells | CD11b | PE | 0.2 mg/mL | Biolegend 101207 |
| CD11c | APC | 0.2 mg/mL | Invitrogen 17-0114-81 |
| CD103 (clone 2E7) | PE-Cy7 | 0.2 mg/mL | Biolegend 121425 |
| Ly6c | Efluor450 | 0.2 mg/mL | Invitrogen 48-5932-80/2 |
Dump gate | CD64 Ly6g Siglec F | FITC FITC FITC | 0.5 mg/mL 0.5 mg/mL 0.5 mg/mL | Biolegend 161007 Biolegend 127606 Biolegend 155503 |
Lastly, to evaluate differences in immune cell transit between the lung and the lung-draining mLN, we designed an 11-color flow cytometry panel to determine the immunophenotype of the lymph node at 2 and 7 dpi. Using markers for T- cells (CD4, CD8) and B-cells (CD19) as well as markers for dendritic cell subtypes (Ly6C, CD11c, CD11b and CD103), we were able to define cell populations known to be important in the developing adaptive immune response, for which a functional lymphatic vasculature is known to be crucial (28, 30). While we did observe predictable differences in the mLN immunophenotype between 2 and 7 dpi, we identified no differences in any of the defined immune cell populations between Cre(+) or Cre(-) YAP/TAZ△LEC littermates (Fig. 7a and b).