LPS intranasal instillations induce pulmonary TLS formation
First, bronchus-associated TLS were induced in the pulmonary tissue of mice receiving LPS. Adult female mice received an intranasal injection daily for 5 days. Lung removal was performed on days 0, 3, 5, 8, 10, 15, 18 and 25 following the first injection of LPS. The experimental design is shown Fig. 1A. We observed the presence of a few small lymphoid aggregates in the close vicinity of the bronchus and blood vessels at day 3 by hematoxylin and eosin (H&E) counterstaining (Fig. 1B). The density and absolute number of aggregates increased thereafter and peaked at day 10 (Fig. 1C and D), with many aggregates evidenced at that time (Fig. 1B). Aggregates were no longer observed at day 25 (Fig. 1B-D).
At days 3 and 5 (i.e., during treatment and the day after the last LPS instillation), aggregates were mostly composed of B220+ B cells, with the presence of a few CD3+ T cells and CD11c+ myeloid cells in close proximity. No PNAd+ HEV were detected at these time points (Fig. 1F). In parallel to the increase in size of lymphoid aggregates over time, PNAd+ HEV became detectable at day 8 (Fig. 1F). At day 10, these aggregates were composed of B220+ B cells, with the presence of CD3+ T cells and PNAd+ HEV, as detected by multiplexed immunofluorescence labeling of mouse lung sections. A representative lymphoid aggregate stained by H&E (left panel) and immunofluorescence (right panel) is shown Fig. 1E. B-cell areas contained germinal centers displaying FDC-M1+ follicular dendritic cells (FDC). Figure 1G shows B220 and FDC-M1 staining at day 10 of a tissue section adjacent to the one in Fig. 1F. Thus, most of these aggregates were mature TLS. Control mice receiving NaCl intranasal instillations did not show any lymphoid aggregates in the lungs (Supplementary Figure S1). B220+ B cells and CD3+ T cells were scattered in the lung parenchyma. PNAd+ HEV and lymphoid aggregates were detected only in LPS-treated animals.
Selective depletion of sympathetic peripheral fibers by 6-OHDA provokes a significant reduction in pulmonary alveolar space
We explored the impact of the depletion of sympathetic peripheral fibers on LPS-induced lymphoid aggregates and TLS. For that purpose, we used 6-OHDA, a neurotoxin that selectively destroys sympathetic peripheral fibers [14]. First, the impact of two subsequent intraperitoneal (i.p.) injections of 6-OHDA (day 0 and day 2) on nerve fibers of naive C57Bl/6 mice was examined. Figure 2A shows the experimental design. Tyrosine hydroxylase (TH), a specific marker of sympathetic fibers, was detected neither at day 30 nor at day 40 in 6-OHDA-treated mouse ears in comparison to control mice receiving NaCl (Fig. 2B).
To assess the drug efficacy in suppressing sympathetic nerve fibers in various organs of treated mice, western blots revealing the presence of TH were performed on lung, kidney and brain extracts. TH could not be detected in the lungs or kidneys of 6-OHDA-treated mice compared to NaCl-treated mice (Fig. 2C). TH levels in the brain were not modified following 6-OHDA injections, confirming that the drug did not cross the blood‒brain barrier [14]. A considerable drop in catecholamine levels in the plasma of 6-OHDA-treated mice was also observed by ELISA, suggesting a massive destruction of sympathetic fibers in these animals (Fig. 2D).
We then assessed the pulmonary structure of treated mice on lung tissue sections at day 30. A reduction in the alveolar space was detected in the pulmonary parenchyma of 6-OHDA-treated mice compared to control mice (Fig. 3A). No respiratory signs (difficulty in breathing, sniffling, sneezing) were recorded in treated animals. TH+ sympathetic innervation of PFA-fixed whole lungs in 6-OHDA- and NaCl-treated animals was next visualized at day 40 by 2D and 3D imaging. The vascular α-SMA+ network was detected in both groups (Fig. 3B and C). This vasculature was surrounded by TH+ sympathetic fibers in control mice (Fig. 3B, upper panel, and C, left panel). In contrast, no fibers but only green scattered dots could be detected in the lungs of 6-OHDA-treated mice (Fig. 3B, lower panel, and C, right panel) (see also Supplementary videos S1 and S2).
Depletion of sympathetic fibers leads to a decrease in LPS-induced lymphoid aggregates and TLS formation in lung parenchyma
We assessed LPS-induced TLS formation in the lungs of mice lacking sympathetic nerve fibers. 6-OHDA- and NaCl-treated mice were given LPS intranasal instillations (Fig. 4A). Day 30 was selected as the starting time for LPS instillations, and lungs were retrieved at day 40, i.e., 10 days later, when TLS density and number were maximal (Fig. 1C and D) and TH+ fibers were still undetectable (Fig. 2B and C). Sympathetic denervation was verified by whole-mount TH staining of 6-OHDA-treated mouse ears. H&E counterstaining of FFPE lung sections showed the presence of lymphoid aggregates in both 6-OHDA- and NaCl-treated mouse lungs (Fig. 4B). The number of lymphoid aggregates observed per lung section, most of which were TLS (indicated as LA-TLS), was significantly lower in 6-OHDA-treated mice than in NaCl-treated controls (P = 0.0006, Fig. 4C, left panel). A decrease in the density of LA-TLS was also observed in 6-OHDA-treated mice compared to control animals (P = 0.0140, Fig. 4C, right panel). As previously observed (Fig. 3A), the lung parenchyma of 6-OHDA-treated mice contained fewer and smaller alveoli than that of control mice (Fig. 4B). Thus, we examined whether this change in lung parenchyma architecture impacts LA-TLS density assessment. Quantification of the ratio of alveolar surface to total lung surface also showed a decrease following 6-OHDA treatment (P < 0.0001) (Supplementary Figure S2). Quantification of LA-TLS density in the lung tissue excluding the alveolar surface confirmed the difference between the two groups (P < 0.0001, Fig. 4D).
Multiplexed immunofluorescence staining of adjacent sections confirmed that most of the lymphoid aggregates were TLS (containing B220+ B cells, CD3+ T cells, CD11c+ myeloid cells and PNAd+ high endothelial venules) (Fig. 5A). The maturation stage of induced TLS was then assessed 10 days after the beginning of LPS instillations (day 40). FDC-M1+ FDC were observed within the B-cell areas in the lungs of both 6-OHDA- and NaCl-treated mice, indicating the presence of mature TLS (Fig. 5B). TLS infiltration by FoxP3+ Tregs in 6-OHDA- vs NaCl-treated mice was similar at day 40 (Fig. 5C).
Pulmonary naive B-cell levels after LPS treatment and systemic primary antibody response are decreased following sympathetic fiber depletion
We assessed whether sympathetic fiber depletion impacts immune cell subsets in the lungs and spleen prior to LPS treatment, when no LA-TLS are detected in the lungs, and after treatment. Animals were i.p. injected with 6-OHDA or NaCl at day 0 and day 2 and intranasally treated with LPS starting at day 30. Spleens and lungs were collected at days 30 and 40 (Fig. 6A), and recovered immune cells were analyzed by flow cytometry. Gating strategies are shown in Supplementary Figures S3 and S4. 6-OHDA treatment did not affect CD45+ leukocyte proportions among live cells in the spleen and in the lungs at day 40 after receiving LPS. Among them, the proportions of CD19+B220+ B cells (Fig. 6B and Supplementary Table S3), CD3+ T cells and IA/IE+CD11b+ plasmacytoid and conventional dendritic cells (Supplementary Table S3) were not altered in 6-OHDA-treated mice compared to NaCl-treated animals. However, at day 40 following LPS treatment, a large difference in the proportion of CD23+ naive B cells was observed between 6-OHDA- and NaCl-treated mice (Fig. 6C and Supplementary Table S3), while no difference was observed at day 30 (40.95% vs 33.48%, P = 0.3086). No drop in naive B cells was observed in the spleen (Fig. 6C and Supplementary Table S3). Finally, no difference in Treg proportions among lung CD4+ helper T cells was observed between the two groups following LPS instillations (Supplementary Table S3), as already shown by immunofluorescence on tissue sections (Fig. 5C). 6-OHDA treatment also had no impact on the expression of immune checkpoints by helper T cells in the spleen or lungs of treated mice (Supplementary Table S3). Although significant differences were reported between the proportions of pulmonary plasma cells between groups, the low values of these proportions, along with the small number of cells it referred to, prompted us to ignore this change.
We then examined the impact of sympathetic fiber depletion on the systemic antibody response of ovalbumin-immunized mice. Mice were treated with 6-OHDA or NaCl and then immunized with ovalbumin, and plasma was collected at different time points (Fig. 7A). Day 30 was chosen for the first boost of ovalbumin, since no TH+ fibers were detected at that time in 6-OHDA-treated animals (Fig. 2B and C). Anti-ovalbumin IgM and IgG responses were assessed by ELISA (Fig. 7B and C). Five days after the first injection (day 35), a lower IgM response was detected in the plasma of 6-OHDA-treated mice compared to 3 out of 4 control animals (Fig. 7B), whereas no IgG response was observed in either group (Fig. 7B, inset). In contrast, there was no difference in the anti-ovalbumin IgG response of hyperimmunized mice at day 56 (Fig. 7C) and later on (data not shown). Denervated mice (3 out of 4) still exhibited a lower IgM anti-ovalbumin response (Fig. 7C, inset).