LungVis 1.0: an automatic AI-powered 3D imaging ecosystem unveils spatial profiling of nanoparticle delivery and acinar migration of lung macrophages

Abstract

Targeted (nano-)drug delivery is essential for treating respiratory diseases, which are often confined to distinct lung regions. However, spatio-temporal profiling of drugs or nanoparticles (NPs) and their interactions with lung macrophages remains unresolved. Here, we present LungVis 1.0, an AI-powered imaging ecosystem that integrates light sheet fluorescence microscopy with deep learning-based image analysis pipelines to map NP deposition and dosage holistically and quantitatively across bronchial and alveolar (acinar) regions in murine lungs for widely-used bulk-liquid and aerosol-based delivery methods. We demonstrate that bulk-liquid delivery results in patchy NP distribution with elevated bronchial doses, whereas aerosols achieve uniform deposition reaching distal alveoli. Furthermore, we reveal that lung tissue-resident macrophages (TRMs) are dynamic, actively patrolling and redistributing NPs within alveoli, contesting the conventional paradigm of TRMs as static entities. LungVis 1.0 provides an advanced framework for exploring pulmonary delivery dynamics and deepening insights into TRM-mediated lung immunity.

Fig. 1. Data-centric, active learning, artificial intelligence-driven precise segmentation of complete murine lung airways.

a Schematic illustration of the key features of LungVis 1.0 including 3D LSFM lung imaging and the development of artificial intelligence-driven airway segmentation to resolve the spatial and temporal nanoparticle (NP) deposition profiles in the mouse lung. Illustrations created partially with Biorender.com. b Manual extraction of ground truth (MS_GT, lung annotations) from the raw, non-stained LSFM images. 2D and 3D images show the original LSFM lung structure (autofluorescence, AF in green), manually segmented GT (in blue), and merged images. c LungVis 1.0 AI pipelines overcome multiple imaging challenges for high-performance airway segmentation. Occasionally, poor image quality arises from imaging shadows, blurring effects, out-of-focus central region (poor illumination in lung center), inconsistent slice illumination, poor and variable signal-to-noise ratio, false gray structures, etc. can be observed in some of LSFM lung images. With the data-centric active learning approach and method improvement, LungVis 1.0 achieved high quality and robust segmentations even in the most challenging cases, as demonstrated for label-free AF lung images in the visible (high AF1 - default AF channel) and near-infrared channel (low AF2). d Two exemplary AI segmentations of complete bronchial trees from either a whole lung or a single lung lobe with different imaging errors (i.e., imaging shadow and blurring) are displayed. Representative data from n = 78 biological samples. e Average time investment for complete airway labeling in lungs via manual versus AI segmentation. Data are presented as mean ± SD, n = 4 biological replicates. f–g The Dice Score and centerline Dice Score were evaluated across three GT lungs in three AI iterations from the test datasets. Data are presented as mean ± SD, n = 3 biological replicates. Scale bars:1000 µm. Source data are provided as a Source Data file.

Fig. 2. Qualitative and quantitative profiling of initial NP distribution in lung airways and acini enabled by LungVis 1.0.

a Schematic illustration delineates pulmonary NP delivery methodologies, longitudinal sampling, and quantitative NP dose in tissue homogenates and spatial NP dosimetry (created with Biorender.com). b Representative TEM and epi-fluorescence images of MF NPs. n = 3, independent replicates with similar results. Scale bars: 500 nm (left) and 15 µm (right). c Optical transparency of a mouse lung prior to and after tissue clearing. Scale bars: 1 cm. d–o AI-powered holistic mapping and cellular resolution views of NP lung distribution for four delivery routes stratified for the airway (e, h, k, n) and acinar region (f, i, l, o). d, g, j, m Schematic depiction of NP airway and acinar deposition (created with Biorender.com). Orange and white arrowheads indicate central/upper and lower/peripheral bronchioles, respectively. Blue arrowheads show NP deposition in upper peripheral regions (e, h). Red arrowheads show tracheal-deposited NPs (k). An acinus is delineated into the proximal and distal acinar region (PAR and DAR) artificially separated by black (d, g, j, m) or yellow (f, i, l, o) dashed lines. TB: terminal bronchioles. Scale bars: 1000 µm (overview), 100 µm (cellular resolution). p Quantitative analysis of NP-lung delivery efficiency. n = 5 INLA, n = 3 ITLI, n = 7 VAAD, n = 5 NOAI, biological replicates. q Central/peripheral NP deposition ratio. n = 3 INLA, n = 7 ITLI, n = 5 VAAD, n = 5 NOAI, biological replicates. r Bronchial/acinar NP deposition ratio. Representative data (d–o) from n = 3 INLA, n = 8 ITLI, n = 7 VAAD, n = 4 NOAI, biological replicates. s Linear correlation between C/P ratio and B/A deposition ratio with mean and 95% confidence intervals indicated by solid and dashed lines, respectively. n = 3 INLA, n = 5 ITLI, n = 6 VAAD, n = 4 NOAI, biological replicates. Data are presented as mean ± SD, calculated using one-way ANOVA with Holm-Šídák’s multiple comparison test (q, r). Source data are provided as a Source Data file.

Fig. 3. Time-resolved 3D qualitative and quantitative profiling of NP biokinetics in intact ITLI lungs enabled by LungVis 1.0.

a–p Longitudinal 3D view of NP distribution pattern in the left lung after intratracheal liquid instillation (ITLI) at time points of 0 h, 2 h, 24 h, and 14 d. a, e, i, m AI-powered visualization of NP airway deposition pattern. b, f, j, n Visualization of NP lung distribution (raw LSFM images). AI-powered global (c, g, k, o) and cellular resolution views (d, h, l, p, within white dashed lines) of NP acinar deposition pattern. White arrowheads indicate no/low NP deposition in ITLI lungs and blue arrowheads indicate different NP (re-)distribution features in the proximal acinar region (PAR). PB: Primary bronchus, TB: Terminal bronchioles. Scale bars: 1000 µm (a–c, e–g, i–k, m–o), 400 µm (ROIs), 200 µm (d, h, l, p). Representative data from n = 8 ITLI_0 h, n = 3 ITLI_2 h, n = 8 ITLI_24 h, n = 5 ITLI_14 days, independent biological replicates. NP intensity (brightness) optimized for spatial visualization, was scaled to match longitudinal lung homogenate dose trends but may not reflect the actual dose. q Schematic depiction of NP inter-acinar and intra-acinar transfer in the ITLI lungs (created partially with Biorender.com).

Fig. 4. Time-resolved 3D qualitative and quantitative profiling of NP biokinetics in intact VAAD lungs enabled by LungVis 1.0.

a–p Longitudinal 3D view of NP distribution pattern in the left lung after ventilator-assisted aerosol delivery (VAAD) at time points of 0 h, 2 h, 24 h, and 14 d. a, e, i, m AI-powered visualization of NP airway deposition pattern. b, f, j, n Visualization of NP lung distribution (raw LSFM images). AI-powered global (c, g, k, o) and cellular resolution views (d, h, l, p, within white dashed lines) of NP acinar deposition pattern. Purple arrowheads refer to low NP deposition in parts of the upper lung after VAAD. Blue arrowheads indicate different NP (re-)distribution features in PAR. PB: Primary bronchus, TB: Terminal bronchioles. Scale bars: 1000 µm (a–c, e–g, i–k, m–o), 400 µm (ROIs), 200 µm (d, h, l, p). Representative data from n = 7 VAAD_0 h, n = 3 VAAD_2 h, n = 6 VAAD_24 h, n = 5 VAAD_14 days, independent biological replicates. NP intensity (brightness) optimized for spatial visualization, was scaled to match temporal lung homogenate dose trends but may not reflect the actual dose. q Schematic depiction of NP inter-acinar and intra-acinar transfer in the VAAD lungs (created partially with Biorender.com). r, s Longitudinal bronchial and acinar NP deposition fractions in the lung. n = 8 ITLI_0 h, n = 3 ITLI_2 h, n = 8 ITLI_24 h, n = 5 ITLI_14 days, n = 7 VAAD_0 h, n = 3 VAAD_2 h, n = 6 VAAD_24 h, n = 5 VAAD_14 days, independent biological replicates. t, u Fraction of NP retention in different compartments of the lung. The retention fractions are normalized to the mean dose at 0 h. n = 3, independent biological replicates. Data are presented as mean ± SD, calculated using one-way ANOVA with Holm-Šídák’s multiple comparison test (r–u). p values are indicated separately for ITLI (r) and VAAD (s) groups. Additional p values for bronchial dose fractions between ITLI_0 h and VAAD_0 h determined using unpaired two-tailed t-test (r). Source data are provided as a Source Data file.

Fig. 5. Lung airway and acinar deposited particles are effectively phagocytosed by tissue-resident macrophages.

a Complementary analytical technique for deciphering intra-acinar NP fate (created with Biorender.com). b, c Representative cellular-resolution 3D views of NP distribution in PCLS at different time points. White and yellow arrowheads indicate NP aggregates engulfed by F4/80⁺ tissue-resident macrophages (TRMs, anti-F4/80) and free NPs in alveolar epithelium, respectively. Podoplanin (PDPN): alveolar epithelial cell Type 1, Phalloidin: cell actin filaments, DAPI: cell nuclei, Scale bars: 50 µm. d Fraction of pulmonary retained NP dose in lung PCLS F4/80⁺ macrophages. b–d n = 5 all groups, except n = 7 ITLI_24 h and VAAD_24 h, n = 4 VAAD_2 h, independent biological replicates. e The fraction of NP⁺F4/80⁺ to total F4/80⁺ macrophages in 24 h ITLI (n = 5) and VAAD (n = 4) lungs. f Four typical localizations of F4/80⁺ macrophages (inset/type I, II, III and IV) and the fraction of tissue-crossing (III) and inner-tissue macrophages (IV) to total NP⁺ macrophages in both 24 h ITLI (n = 6) and VAAD lungs (n = 5). Scale bars: 10 µm. g, h Typical localization of tissue-crossing or inner-tissue NP⁺F4/80⁺ macrophages in 3D-reconstructed PCLS from 2 h ITLI (n = 5) and VAAD (n = 4) lungs. Scale bars: 10 µm. i Exemplary cell population analysis of NP⁺ macrophages stained with SiglecF or CD11b and quantifications on SiglecF⁺ (AM) and CD11b⁺ (IM) cells to all NP⁺ cell (j, k), and the NP fraction attributed to each type (l) in 24 h ITLI normal and lavaged lungs. j n = 7 normal, n = 6 lavaged lungs. k, l n = 7 CD11b⁺_normal, n = 5 SiglecF⁺_normal, n = 6 CD11b⁺_lavaged, n = 3 SiglecF⁺_lavaged, independent biological replicates. Data are presented as mean ± SD, calculated using two-tailed (e, f) or one-tailed (j) unpaired t test and multiple two-tailed unpaired t test with Holm-Šídák correction (d, k–l). Source data are provided as a Source Data file.

Fig. 6. Redistribution of acinar deposited NPs can be attributed to phagocytosis and migration of TRMs.

a Lung intravital microscopy (IVM) revealed the dynamic movement of PKH labeled macrophages toward epithelial-deposited NPs by patrolling the alveolar epithelial surface. n = 3, VAAD_0 h lungs, independent biological replicates. Scale bar: 20 µm. b Relative frequencies of average tracking velocities of PKH-labeled macrophages in vivo imaged by IVM. n = 3 VAAD_0 h, n = 4 ITLI_24 h, independent biological replicates. c Trajectory plot outlines relative patrolling behaviors of individual PKH-labeled macrophages from 3 VAAD_0 h lung imaged by IVM (Start points of the migration tracks set to 0,0). d, e Fractions of NP⁺ cells in PKH⁺ cells and fractions of CD11b⁺ or SiglecF⁺ in NP⁺PKH⁺ cells in normal and lavaged 24 h ITLI lungs. n = 3 (lavaged) or 5 (normal), independent biological replicates. f Ex vivo lung living microscopy showed the migration of GFP⁺NP⁺ macrophages in the PCLS (n = 4). Scale bar: 20 µm and 10 µm (ROI). g Exemplary cell population analysis of GFP⁺NP⁺ macrophages and (h, i) Fractions of NP⁺ cells in GFP⁺ cells and fractions of CD11b⁺ or SiglecF⁺ in NP⁺GFP⁺ cells in normal and lavaged lungs. n = 3 (lavaged) or 4 (normal), independent biological replicates. Data are presented as mean ± SD, calculated using one-tailed unpaired t test (d, h) and multiple two-tailed unpaired t test with Holm-Šídák correction (e, i). j Macroscale and microscale views of GFP⁺ macrophages/monocytes and NP deposition in an 4 h ITLI lung lobe (n = 3). White arrowheads indicate merged signals of GFP and NPs. Scale bars: 1000 µm (overview), 200 µm (ROI1), and 100 µm (ROI2). k Co-expression/staining of GFP, PKH dye, and NP⁺ BAL cells obtained from 24 h ITLI mac-green mice (n = 3). White arrowheads and arrows, yellow, and blue arrows indicate triple positive cells, double positive cells (PKH⁺ and GFP⁺), and GFP⁺ cells, respectively. Scale bar: 20 µm. Source data are provided as a Source Data file.

Fig. 7. Cellular activity governs phagocytosis-mediated relocation and wide dispersion of NPs throughout different lung networks.

a Schematic illustration of an ex vivo whole-lung ventilation model to investigate the role of cellular activity in NP-lung relocation. Image illustrations created partially with Biorender.com. b, c Holistic NP distribution in whole lung following 4% PFA fixation or PBS treatment, respectively, succeeded by 24 h of ex vivo ventilation. n = 4, independent biological replicates. d NP (re-)distribution of an in vivo 24 h ITLI lung is characterized by more uniform distribution (less patchy than at 0 h) and its dot-like pattern. White arrowheads: NP-cluster distribution, and blue arrowheads: single NP-dot distribution. n = 8, independent biological replicates. Scale bars (b–d): 1000 µm (overview), 400 µm (ROI1), and 100 µm (ROI2). e, f microscopical 3D and cellular-resolution 2D views of NP distribution in the lung and its association with the endothelial systems including the lymphatics and vasculature networks (LYVE1: lymphatics and blood vessels, PDPN: strong lymphatics and weak alveolar epithelium cell Type I staining). This reveals NPs or NP-laden macrophages in close proximity to the blood vessels (yellow arrows) throughout the entire lung lobe scanned by LSFM. The blood vessels, lymphatics, and airways indicated with arrowheads in blue, purple, and cyan, and the NPs in alveolar epithelium or septum marked with yellow arrowheads. Scale bars: 1000 µm (overview) and 100 µm (ROIs). Representative data from n = 3, independent biological replicates. g LSFM imaging of tracheobronchial (TB) lymph nodes at 24 h and 14 days after ITLI exposure and untreated control. Scale bars: 100 µm. h 2D immunostaining of TB lymph nodes with several TRM markers at 24 h and 14 days post-VAAD exposure, as well as in untreated controls. Arrowheads indicate the NP-laden macrophages in TB lymph nodes. Scale bars: 10 µm. Representative data from n = 3, independent biological replicates (g, h).