Selective activation and proliferation of a quiescent stem cell population in the neuroepithelial body microenvironment

Lung tissue was obtained both from wild type (WT) C57BL/6 J (WT-Bl6) mice, and from a C57BL/6 J based glutamate decarboxylase 67-GFP (GAD67-GFP) mouse strain that harbors GFP fluorescent pulmonary NEBs [58] (The Jackson Laboratory, Charles River, Saint-Germain-sur-l’Arbresle, France). Both male and female mice, mainly at postnatal day (PD) 21, were used. Three-week-old mice have a much smaller lung volume –but a similar number of NEBs– compared to adults, and hence a higher density of NEBs in lung sections, resulting in more efficient studies. The results were however double-checked for adult mice (8-week-old). The young animals were housed together with their mothers in acrylic cages in an acclimatized room (12/12 h light-dark cycle; 22 ± 3 °C) and were provided with water and food ad libitum. National and international principles of laboratory animal care were followed, and experiments were approved by the local animal ethics committee of the University of Antwerp (ECD 2014–66 and 2017–49). All animals were killed by intraperitoneal (i.p.) injection of an overdose of sodium pentobarbital (Nembutal® 200 mg/kg bodyweight (BW), CEVA Sante Animale, Brussels, Belgium).

As an analog of thymidine, 5-Bromo-2′-deoxyuridine (BrdU) will be incorporated in DNA during the S-phase of cell division. To mark cells that have divided during the experimental window, i.p. injections of BrdU (10 mg/kg; B5002; Sigma, Bornem, Belgium) were given immediately following intratracheal instillation in the treated groups, and at the same time point in untreated matched control mice. BrdU injections were repeated at 24 h and 48 h post-instillation. After recovery, animals were returned to their mothers until the day of euthanasia, 24 h, 48 h or 72 h after LPS challenge or any of the other treatments. Additionally, a few mice received BrdU at time points 0 h and 24 h but were sacrificed 7 days following LPS instillation. The lungs were then processed for cryosectioning and immunostaining (see further on).

Sixteen hours after LPS or sham instillation, WT-Bl6 mice (n = 4 for each group) were euthanized and exsanguinated. Untreated WT-Bl6 mice (n = 4) served as an additional control. Bronchoalveolar lavage fluid (BALF) was collected by instillation and suction of a salt solution (2 × 1 ml; 0.9% NaCl) in the lungs via a tracheal cannula. BALF was stored at 4 °C until further use. In accordance with literature [59], the time point of 16 h after LPS treatment was chosen to allow the animals to recover completely from the anesthetics and still be able to detect effects from a combination of early and late phase mediators in the BALF. The collected BALF was first centrifuged (12 min, 150G), the cell-free supernatant fluid was removed and used as a stimulus solution for the NEB ME in lung slices in live cell imaging (LCI) experiments (see below). The cell pellet was used to prepare cytospin slides for evaluation of the cells that are present in the pulmonary air spaces. The pellet was resuspended to an end concentration of about 10⁶ cells/ml in phosphate-buffered saline (PBS; 0.01 M, pH 7.4), containing 5% bovine serum albumin (BSA; B4287, Sigma). 150 μl of this mixture was centrifuged (5 min, 800G) to generate cytospin preparations (Shandon Cytospin 3 Cytocentrifuge, Fisher Scientific, Erembodegem, Belgium) of BALF cells. After air-drying, the slides were fixed and stained using a fast routine blood cell staining method (Diff-Quick; DQ-ST, MICROPTIC, Barcelona, Spain), and mounted in Entellan (Merck; Overijse, Belgium). Cell types in the BALF cytospins were morphologically characterized under a light microscope and compared between LPS-challenged, sham-treated and untreated controls. Data were used only for general interpretation of the inflammatory effect of instillation, and no further quantification was performed.

Live lung slices were prepared as previously published [60]. In short, WT-Bl6 mice (n = 3) were euthanized and the lung tissue was stabilized by slowly instilling a 2% agarose solution (37 °C, low-melt agarose, A4018, Sigma) via a tracheal cannula. After inflation, lungs were dissected and transferred to an ice-cold physiological solution. Precision-cut lung slices (120 μm thick) were sectioned using a vibratome (Microm HM650 V; Microm International, Walldorf, Germany) and 6–8 slices per animal were subsequently kept in the cold physiological solution until further manipulation within the next few hours (maximally 6 h).

As previously reported in detail [8], different cell types in the NEB ME and in the surrounding airway epithelium, such as PNECs, CLCs, CCs and ciliated cells, can be identified in live lung vibratome slices after staining with the fluorescent dye 4-Di-2-ASP (Molecular Probes D-289, Fisher Scientific) and loading with the Ca²⁺ indicator Fluo-4 AM (Molecular Probes F-14202; Fisher Scientific). In short, lung slices were consecutively incubated for 4 min in 4 μM 4-Di-2-ASP in Dulbecco’s modified Eagle’s medium/F-12 (DMEM-F-12; Gibco, Fisher Scientific) at 37 °C, rinsed, and incubated for 1 h at room temperature in 10 μM Fluo-4 AM. For experimental LCI purposes, the lung slices were submerged in a tissue bath (2 ml) mounted on the microscope stage, continuously perfused with physiological solution by a gravity-fed system (flow rate of 0.5 ml/min). Perfusion was paused when BALF was manually pipetted onto the lung slice in the tissue bath. Physiological solution containing a high extracellular potassium concentration ([K⁺]_o) was prepared by equimolar substitution of KCl for NaCl [8].

High-resolution LCI was performed as extensively described before [8, 13, 60]. In short, a microlens-enhanced dual spinning disk confocal system (UltraVIEW ERS; PerkinElmer, Zaventem, Belgium), equipped with an argon-krypton laser was used. Time-lapse images of changes in Fluo-4 fluorescence (excitation max. 494 nm; emission max. 516 nm) were recorded (2 images/s; 488-nm laser excitation; bandpass 500–560 emission filter) and analyzed off-line by Volocity 2 software (Improvision, Coventry, UK). Regions of interest were manually drawn around identified cell groups of interest, and the fluorescence intensity was plotted against time. Changes in Fluo-4 fluorescence should be interpreted as relative changes in the intracellular Ca²⁺ concentration ([Ca²⁺]_i). All graphs presented are representative of multiple experiments performed under the respective conditions.

Immunohistochemical incubations were performed at room temperature in a closed humidified container. All primary and secondary antisera were diluted in PBS containing 10% normal horse serum and 0.1% BSA (PBS*). Before incubation with the primary antisera, cryostat sections were permeabilized for 1 h with PBS* containing 1% Triton X-100. Sections were then incubated overnight with a combination of the primary antibodies listed in Table 1. For visualization of the immunolabeling, sections were further incubated for 4 h with secondary antibodies (Table 2).

After a final wash in PBS, the sections were mounted in Citifluor (19,470; Ted Pella, Redding, CA).

For each animal, a few cryosections were routinely stained with hematoxylin and eosin (H&E), dehydrated in a graded series of ethanol and xylene, and mounted in Entellan. Lung sections were evaluated under a light microscope for the histological evaluation of epithelial damage, interstitial inflammation, edema and leukocyte recruitment, as characteristics of pulmonary inflammation and lung injury [61].

Quantification of the BrdU-positive cells was performed by manually counting the fluorescent nuclei in the areas of interest. Lung cryosections (20 μm-thick) were collected and selected in a reproducible manner. Per slide, two sections were mounted in such a way that the distance between both sections is 200 μm. In short, ten consecutive sections were mounted on different slides, after which the following 10 sections were mounted in the same order on these slides. The next 20 consecutive sections were collected on 10 new slides, and so on until the lung tissue was completely cut. Then, a first slide for staining was selected based on the presence of airway branches and presumably NEBs. Starting from this slide, six more were taken every 10 slides, thereby avoiding the possibility that more than one section of the same NEB ME could be found and/or counted in the selected slides when immunostained for BrdU and some reference markers. As such, for every mouse in the different treatment groups (LPS-treated, sham treated and untreated control), between 60 and 100 NEBs were visualized under the microscope, by their GFP fluorescence in GAD67-GFP mice or CGRP immunostaining in WT-Bl6 mice, and the PNECs and BrdU-positive cells in the NEB ME were counted.

For each animal in all of the experimental groups, the mean number of BrdU-positive cells per NEB ME was calculated and the data were statistically compared between the different treatment groups, using a nonparametric Kruskal-Wallis test followed by Dunn’s multiple comparisons test. Data are represented as (mean ± SEM).

Potential differences in the number of BrdU-positive cells between the two mouse strains were statistically evaluated using the unpaired t-test for each treatment group, after checking normal distribution of the counts.

Although the recorded plethysmographic data did not qualify for quantification, due to individual variation inherent to the use of unrestrained young mice, some of the observations were of importance for the presented study. Apart from clear but variable differences in the measurements of TE, RT, EIP and TV between untreated controls and LPS-challenged (and to a lesser extent also sham-treated) mice during the first 2 to 6 h, plethysmography could no longer distinguish LPS-challenged from untreated animals 8 h or longer after treatment (data not shown).

To assess possible inflammatory changes in the airway environment, BALF was collected from the same animals that had been monitored by plethysmography (16 h after instillation of LPS or saline and untreated), and processed for the generation of cytospin preparations. While BALF of healthy control animals showed macrophage-like cells only (Fig. 1a), the majority of leukocytes in the LPS-treated mice appeared to be neutrophils (Fig. 1c, d). Neutrophils were also seen in BALF of the sham-treated mice (Fig. 1b), but clearly to a lesser extent than after LPS challenge.

The applied single low-dose intratracheal LPS challenge did not cause obvious histological changes in the airway epithelium or alveolar areas in H&E-stained lung sections (Fig. 2).

An important element for the interpretation of the following experimental data is that this combined approach (plethysmography, evaluation of the collected BALF; lung histology), supports the idea that the applied single low dose of intratracheal LPS induces a mild and transient inflammation.

Apart from detecting/recording [Ca²⁺]_i fluctuations in the NEB ME, freshly cut live lung slices co-loaded with 4-Di-2-Asp and Fluo-4, also allowed to differentiate between all cell types in the NEB ME and the surrounding airway epithelium (for detailed explanation see [8]). In short, very small and moderately fluorescent grouped NEB cells are encircled by a virtually non-fluorescent rim that harbors much larger CLCs, and are further surrounded by intermingled strongly fluorescent polygonal ciliated cells and large rounded unstained CCs (Fig. 3b).

To allow fast evaluation of the effect of LPS-induced inflammatory mediators on the NEB ME, BALF that had been collected from mice 16 h after intratracheal instillation of LPS, was applied to lung slices of healthy control mice in the course of the LCI experiments.

BALF of LPS-treated mice was seen to induce a reversible and reproducible [Ca²⁺]_i rise, selectively in CLCs of the NEB ME (Fig. 3). CLCs appeared to react within about 10 s after administration of BALF and revealed [Ca²⁺]_i oscillations for about a minute. On the other hand, NEB cells, CCs and ciliated cells did not show a [Ca²⁺]_i rise. Comparable challenge of lung slices with LPS (15 μg/ml in physiological solution) or with BALF of sham and untreated control mice did not induce a [Ca²⁺]_i rise in any cell type in the NEB ME or airway epithelium, while the physiological responsiveness of the lung slices could be confirmed by the typical fast and reversible [Ca²⁺]_i increase in NEB cells and a slightly delayed [Ca²⁺]_i rise in CLCs, following a 5-s application of 50 mM K⁺ (not shown) [8].

The above observation that CLCs ̶̶ which are presumed quiescent airway epithelial stem cells in healthy control mice ̶̶ show a calcium-mediated activation upon short-term application of BALF that contains soluble mediators from the airways of LPS-challenged mice, raised the question as to what the effects might be on CLCs that experience long-term exposure to these mediators in airways of LPS-treated mice. To follow up on the hypothesis that activation of quiescent (non-dividing) stem cells might be linked to proliferation, the potential effect of intratracheal administration of a single low dose of LPS on airway epithelial cell proliferation in general, and on the NEB ME in particular, was monitored using simultaneous BrdU incorporation as a marker for cells that divide during the experimental window. Three experimental groups were included: untreated control mice, mice receiving an intratracheal instillation with LPS, and mice with an intratracheal instillation of a 0.9% NaCl solution (= sham control).

In cryosections of the lungs of untreated healthy 3-week-old WT-Bl6 mice, hardly any divided cells were discerned in control airway epithelium (CAE) within the investigated time windows (24, 48 (Fig. 4a) or 72 h) as confirmed by BrdU labeling. The very rare divided epithelial cells appeared to be randomly distributed. The majority of cells with BrdU-labeled nuclei in the airways of these untreated controls were found subepithelially (Fig. 4a). As a positive control, BrdU incorporation was evaluated in sections of a piece of small intestine that was additionally collected from each mouse. A large number of divided (BrdU-labeled) epithelial cells was invariably observed in crypts of the intestinal villi (Fig. 4b).

Forty-eight hours after a single intratracheal instillation with low-dose LPS, a considerable number of divided (BrdU-positive) epithelial cells was detected, typically clustered in distinct areas of the airway epithelium of WT-Bl6 mice (Fig. 5a).

In sham-treated WT-Bl6 mice, BrdU-positive cells also appeared to be more numerous at specific locations in the airway epithelium (Fig. 5c) than in untreated controls, although the increase was less pronounced than that seen in LPS-treated mice.

Double immunostaining of the BrdU-labeled sections for CGRP, as a marker for NEBs, revealed that the great majority of divided airway epithelial cells were located in the NEB ME following LPS challenge (Fig. 5b, d).

All of the above data were obtained using 3-week-old mice, but a similar selective cell proliferation (BrdU positive nuclei) in the NEB ME was observed in LPS-challenged adult mice (not shown).

Given the apparent close link between NEBs and divided (BrdU-labeled) epithelial cells in the airways of LPS-challenged mice, the use of GAD67-GFP mice that harbor intrinsically GFP-fluorescent NEBs [58] would offer considerable advantages for further quantification. Therefore, we verified whether similar results could be obtained in WT-Bl6 and GAD67-GFP mice. Comparable to what was seen using WT-Bl6 mice with CGRP as a NEB marker (Figs. 5a, b and 6a, b), LPS-treated GAD67-GFP mice also showed BrdU-labeled (divided) airway epithelial cells that were selectively grouped around GFP-fluorescent NEBs (Fig. 6c, d).

Both in healthy controls, sham-treated and LPS-challenged mice, the numbers of CGRP- or GFP-positive PNECs with BrdU-positive nuclei were very low after 48 h exposure.

Since the majority of proliferating (BrdU-labeled) airway epithelial cells in LPS-treated mice was located in the NEB ME, closely surrounding NEB cells, and because BALF of LPS-challenged mice was seen to selectively activate CLCs in the NEB ME, we further investigated whether BrdU-positive cells co-label with markers for CCs/CLCs.

To further confirm the identity of BrdU-labeled cells in the NEB ME as CLCs, lung sections (48 h after LPS instillation) were additionally immunostained for CCSP, a marker of both CCs and CLCs surrounding NEB cells. CCSP was used due to the lack of a selective marker for CLCs in postnatal lungs and revealed that cells with BrdU-labeled nuclei in the NEB ME invariably co-stained for CCSP (Fig. 7c), confirming that mainly CLCs were concerned.

Because single intratracheal instillation of LPS appeared to specifically induce cell division (BrdU incorporation) in the NEB ME in a 48 h time window, we next evaluated whether this was the most relevant time point for visualization/quantification of cell proliferation in the NEB ME. To this end, three time windows were included in a pilot experiment; lungs were collected 24 h, 48 h and 72 h after LPS challenge. Our preliminary data revealed BrdU-labeled airway epithelial cells in the NEB ME already after 24 h, although the effect appeared to be much more prominent 48 h after LPS challenge (Figs. 5a,b, 6 and 7). After 72 h, some of the BrdU-labeled epithelial cells were located a few cells away from NEBs, making this time window somewhat less selective for the NEB ME.

Interestingly, 7 days after LPS challenge, BrdU-positive nuclei could still be detected in CCSP-positive CLCs in the NEB ME, similar to what was seen after 48 h, but also in CCSP-labeled CCs in the surrounding airway epithelium (Fig. 8).

Because the NEB ME is the main area of interest for the present investigation, and considering the observed calcium-mediated activation of CLCs, the time window of 48 h BrdU incorporation after LPS challenge was chosen for further quantification of cell proliferation in the NEB ME.

For a limited number of experiments (n = 3 GAD67-GFP mice per group), all BrdU-labeled cells of each NEB ME were individually identified ̶̶ as CLCs or PNECs ̶̶ and counted (data shown in Fig. 11). Clearly, the number of divided PNECs was very limited compared to that of CLCs in the NEB ME. Moreover, the mean number of BrdU-labeled PNECs was not significantly different between the three experimental groups (Kruskal-Wallis; p = 0.25).