Whole-Cell Scale Dynamic Organization of Lysosomes Revealed by Spatial Statistical Analysis

Abstract

In eukaryotic cells, lysosomes are distributed in the cytoplasm as individual membrane-bound compartments to degrade macromolecules and to control cellular metabolism. A fundamental yet unanswered question is whether and, if so, how individual lysosomes are organized spatially to coordinate and integrate their functions. To address this question, we analyzed their collective behavior in cultured cells using spatial statistical techniques. We found that in single cells, lysosomes maintain non-random, stable, yet distinct spatial distributions mediated by the cytoskeleton, the endoplasmic reticulum (ER), and lysosomal biogenesis. Throughout the intracellular space, lysosomes form dynamic clusters that significantly increase their interactions with endosomes. Cluster formation is associated with local increases in ER spatial density but does not depend on fusion with endosomes or spatial exclusion by mitochondria. Taken together, our findings reveal whole-cell scale spatial organization of lysosomes and provide insights into how organelle interactions are mediated and regulated across the entire intracellular space.

Keywords: clustering; endoplasmic reticulum; endosome; lysosomal positioning; lysosome; organelle interaction; spatial distribution; spatial organization; spatial statistical analysis.

Figure 1. Lysosomes Maintain Non-random and Stable Spatial Distributions in a Single Cell

(A) Lysosomes in a BS-C-1 cell at three selected time points. N, nucleus. Scale bars, 10 μm. (B) Maximum intensity projection (MIP; green) of lysosomal movement for 1 min, imaged at 4 frames per second. Each trace corresponds to the trajectory of a lysosome. Scale bar, 10 μm. Scale bars in insets, 5 μm. (C–E) Complete spatial randomness (CSR) test of whole-cell scale lysosomal distribution at the three time points in (A). (C) 0 s; (D) 30 s; (E) 60 s. Adjusted Ripley’s K-function of lysosomes within the cell (solid black line). Adjusted Ripley’s K-function of a random distribution within the same cell boundary (dotted red line). Uncertainty envelope for the random distribution (gray area). The extent of separation between the solid black line and the CSR envelope indicates how close the spatial distribution of lysosomes is to a random distribution. (F–H) Three distance distributions of lysosomes, color coded based on time and plotted every 5 s for 60 s. Their temporal variations were quantified using Sørensen dissimilarity scores. Because the distributions at 0, 5, 10, … 60 s were selected, 13 distributions were compared pairwise, hence $C_{13}^2=78$ pairs. pdf, probability density function. Temporal variations (means ± SDs; n = 78): (F) 3.29% ± 2.98%; (G) 2.65% ± 0.92%; (H) 5.80% ± 1.48%.

Figure 2. Composition of the Lysosomal Population and Roles of Microtubule-Based Active Transport in Maintaining Its Stable Spatial Distribution

(A) Maximum intensity projection of movement of lysosomes in a COS-7 cell imaged at 10 frames per second for 20 s. Scale bar, 15 μm. (B) MSD of randomly selected 10% of all lysosomal trajectories. (C) Percentage of each subpopulation (means ± SDs; n = 9 cells): constrained: 49.41% ± 6.24%; directed: 30.94% ± 5.40%; free: 19.64% ± 2.19%. Error bars indicate SD. (D) Color-coded trajectories of lysosomes at three time points in a cell treated with 2.5 μM nocodazole (NCD). (E) Changes in lysosomal subpopulations under NCD treatment over time (means ± SDs; n = 9 cells). Constrained (0, 15, and 30 min): 49.41% ± 6.24%, 67.74% ± 4.17%, and 77.20% ± 3.32%, respectively. Directed (0, 15, and 30 min): 30.94% ± 5.40%, 16.77% ± 2.05%, and 11.71% ± 2.48%, respectively. Free (0, 15, and 30 min): 19.65% ± 2.19%, 15.49% ± 3.32%, and 11.05% ± 3.80%, respectively. Comparison of pooled data from the same cells before and after NCD treatment. p values (0 min versus 15 min, 0 min versus 30 min, and 15 min versus 30 min): constrained: 1.8 × 10⁻⁴, 9.6 × 10⁻⁶, and 8.1 × 10⁻⁴, respectively; directed: 2.7 × 10⁻⁵, 3.2 × 10⁻⁵, and 2.0 × 10⁻³, respectively; free: 2.6 × 10⁻², 7.3 × 10⁻⁵, and 1.3 × 10⁻², respectively. *p < 0.05; **p < 0.01; ***p < 0.001. Error bars indicate SD. (F) The three distance distributions before and after 30 min of NCD treatment in 9 different cells. (G) Comparison of different distance distributions of lysosomes before versus after NCD treatment. Cutoff p value for statistical significance: 0.05. Inter-organelle (same, different): 0%, 100%, respectively; to-nucleus: 55.5%, 44.4%, respectively; and nearest-neighbor: 77.8%, 22.2%, respectively. (H) Comparison of median distances before versus after NCD treatment. Inter-organelle (smaller, same, and larger): 22.2%, 11.1%, and 66.7%, respectively; To-nucleus: 33.3%, 44.4%, and 22.2%, respectively; and nearest-neighbor: 33.3%, 66.7%, and 0%, respectively. (I) The three distance distributions before and after 1 hr of ciliobrevin D (80 μM) treatment in 13 different cells. (J) Comparison of three distance distributions before versus after ciliobrevin D treatment of the same cells: Inter-organelle (same, different): 0%, 100%, respectively; to-nucleus: 15.4%, 84.6%, respectively; and nearest-neighbor: 23.1%, 76.9%, respectively. (K) Comparison of median distances before versus after ciliobrevin D treatment. Inter-organelle (smaller, same, larger) 7.7%, 0%, 92.3%, respectively; To-nucleus: 0%, 15.4%, 84.6%, respectively; Nearest-neighbor: 69.2%, 30.8%, 0%, respectively.

Figure 3. Roles of Interaction with the Actin Cytoskeleton and Lysosomal Biogenesis in Maintaining Subpopulations and Spatial Distributions of Lysosomes

(A) Color-coded trajectories of lysosomes in COS-7 cells treated with 0.8 μM latA. Videos were collected at 10 frames per second for 20 s. (B) Changes in the displacement for 5 s of the three subpopulations under latA treatment. Comparison of pooled data from the same 11 cells before and after treatment. Constrained diffusion: 0.32 ± 0.0058 μm (mean ± SEM; 0 min, n = 2,052 trajectories), 0.39 ± 0.0076 μm (7.5 min, n = 1,821), and 0.41 ± 0.0075 μm (15 min, n = 1,747). Free diffusion: 0.57 ± 0.019 μm (0 min, n = 688 trajectories), 0.66 ± 0.020 μm (7.5 min, n = 633), and 0.68 ± 0.019 μm (15 min, n = 675). Directed movement: 1.61 ± 0.031 μm (0 min, n = 1,757 trajectories), 1.67 ± 0.031 μm (7.5 min, n = 1,833), and 1.78 ± 0.032 μm (15 min, n = 1,713). p values (0 min versus 7.5 min, 0 min versus 15 min, and 7.5 min versus 15 min): constrained diffusion: 1.2 × 10⁻¹⁴, 1.9 × 10⁻²², and 0.083, respectively; free diffusion: 2.1 × 10⁻³, 3.6 × 10⁻⁵, and 0.32, respectively; directed movement: 0.17, 1.6 × 10⁻⁴, and 0.014, respectively. Error bars indicate SEM. (C) Changes in the three subpopulations under latA treatment over time (means ± SDs; n = 11 cells). Constrained (0, 7.5, and 15 min): 42.14% ± 6.90%, 39.34% ± 5.09%, and 38.96% ± 5.60%, respectively. Free: 20.34% ± 1.31%, 20.41% ± 1.68%, and 21.74% ± 2.51%, respectively. Directed: 37.52% ± 7.02%, 40.25% ± 6.00%, and 39.30% ± 6.20%, respectively. p values (0 min versus 7.5 min, 0 min versus 15 min, and 7.5 min versus 15 min): constrained: 0.0058, 0.0053, and 0.49, respectively; free: 0.98, 0.15, and 0.22, respectively; directed: 0.0043, 0.10, and 0.61, respectively. Error bars indicate SD. (D) The three distance distributions before and 15 min after latA treatment in 11 cells. (E) Comparison of distributions before and after latA treatment (n = 55 pairs). Inter-organelle (same, different): 0%, 100%, respectively; to-nucleus: 45.5%, 54.5%, respectively; nearest-neighbor: 72.7%, 27.3%, respectively. (F) Comparison of median distances before and after latA treatment (n = 55 pairs). Inter-organelle (smaller, same, and larger): 54.5%, 9.1%, and 36.4%, respectively; to-nucleus: 27.3%, 36.4%, and 36.4%, respectively; nearest-neighbor: 9.1%, 81.8%, and 9.1%, respectively. (G) Comparison of lysosomal spatial distributions in a control cell (left) versus a cell treated with trehalose (right; 50 mM, 12 hr). Lysosomes (red); nuclei (blue). Scale bars, 15 μm. (H) The three distance distributions in control cells (n = 7) versus trehalose-treated cells (12 hr, 7 cells). (I) Comparison of distributions in control cells versus cells treated with trehalose (n = 49 pairs). Inter-organelle (same, different): 0%, 100%, respectively; to-nucleus: 14.3%, 85.7%, respectively; nearest-neighbor: 2.1%, 97.9%, respectively. (J) Comparison of median distances in control cells versus cells treated with trehalose (n = 49 pairs). Inter-organelle (smaller, same, and larger): 91.8%, 0%, and 8.2%, respectively; to-nucleus: 69.4%, 16.3%, and 14.3%, respectively; nearest-neighbor: 91.8%, 8.2%, and 0%, respectively.

Figure 4. Lysosomes Form Dynamic Clusters throughout the Intracellular Space

(A) Selected frames from a time-lapse video of lysosomes in a COS-7 cell. Scale bar, 20 μm. (B) Color-coded spatial density plots of lysosomes calculated for the frames shown in (A). Scale bar, 20 μm. (C) Clusters of lysosomes identified computationally by DBSCAN. Arrowheads point to cluster-splitting sites. Arrows point to cluster-merging sites. (D) Composition of clusters, calculated for each cluster and then pooled for analysis. Constrained diffusion 47.9% ± 19.7%; directed movement, 30.6% ± 17.7%; free diffusion, 21.5% ± 14.9% (means ± SDs; n = 376 clusters from 9 cells, within 5 frames randomly selected from each cell). Error bars indicate SD. (E) Size distribution of clusters, measured by their numbers of lysosomes. Same clusters from the same cells as those in (D). The average number of lysosomes was 15.4 ± 1.3 per cluster (means ± SEMs; n = 376 clusters from 9 cells). The median number of lysosomes was 8. (F) Lifetime of large clusters with >10 lysosomes in the same cells before and after NCD treatment, as shown in Figure 2D. Before NCD treatment, 6.14 ± 1.22 s (mean ± SEM; n = 34 clusters from 8 cells); after NCD treatment, 10.69 ± 1.76 s (mean ± SEM; n = 24 clusters from 8 cells); p = 0.013. Error bars indicate SEM.

Figure 5. Formation of Lysosomal Clusters Is Associated with Local Increases in ER Spatial Density but Does Not Depend on Spatial Exclusion by Mitochondria

(A) Lysosomes remain tethered to the ER network. Left: a selected frame from a time-lapse video of a COS-7 cell in which ER and lysosomes were labeled. Right: zoomed-in views of selected frames of lysosomes in the rectangles. 1′: a lysosome undergoing directed movement. 2′: a lysosome undergoing constrained diffusion. 3′: a lysosome undergoing directed movement at the tip of a growing ER tubule. Scale bars, 5 μm (left), 1 μm (right). (B) MSD analysis of the lysosome in panel 1′ of (A). (C) MSD analysis of the lysosome in panel 2′ of (A). (D) MSD analysis of a lysosome undergoing free diffusion (image not shown). (E) Comparing ER spatial densities in clustered regions versus non-clustered regions. Segmented ER shown in white and background in gray. Lysosomes in clusters (red dots).Lysosomes not in any cluster (brown dots). Selected clusters for analysis (green contours). Randomly selected non-cluster regions with the same shape as a cluster (blue dash contours). Scale bar, 10 μm. (F) Comparison of ER densities in cluster regions versus non-cluster regions. Cluster regions: 47.5% ± 7.6% (mean ± SD; n = 31 regions from 3 cells); non-cluster regions: 36.9% ± 9.0% (mean ± SD; n = 116 regions from 3 cells); p = 1.2 × 10⁻⁸. Error bars indicate SD. (G–I) Selected frames at 0s (G), 28.5s (H), and 60.4s (I), from a time-lapse video of a COS-7 cell in which mitochondria and lysosomes were labeled. Scale bar, 10 μm. (J–L) Lysosomal clusters and their spatial relations with mitochondria (shown in grayscale) 0s (J), 28.5s (K), and 60.4s (L). Lysosomes in clusters (red dots). Lysosomes not in any cluster (blue dots). Lysosome cluster boundaries (green dashed contours). Scale bar, 10 μm.

Figure 6. Clustering of Lysosomes Increases Their Interaction with Late Endosomes

(A) Selected frames from a time-lapse video of a COS-7 cell in which lysosomes (green) and late endosomes (red) were labeled. Scale bars, 15 μm. (B) An example of a computationally detected interacting lysosome-endosome pair that remained together for >4 min. Top three rows: the actual fluorescence signals. Bottom row: computational detection result. Scale bars, 1 μm. (C) Clusters of late endosomes and lysosomes detected from the video in (A). Interacting lysosome-endosome pairs (magenta). Lysosome (green). Late endosome (blue). Lysosome cluster boundaries (light blue lines). Endosome cluster boundaries (orange lines). (D) Same as (E) but showing only the clusters and detected lysosome-endosome pairs. (E) Number of detected interacting lysosome-endosome pairs in 10 cells within 5 min: 89, 26, 101, 54, 64, 61, 63, 32, 117, and 58. (F) Histogram of the duration of the detected interacting pairs staying together. (G) Relation between the location of interacting pairs and the clusters of endosomes and lysosomes.

Figure 7. Tracking of Fusion between Lysosomes and Endosomes and a Model of Lysosomal Clustering

(A) An example of lysosome-endosome fusion in a COS-7 cell. At 0–9 s, two lysosomes (arrowheads) moved close to an endosome. At 36 s, the three organelles overlapped with one another in their fluorescence signals, presumably undergoing fusion or partial content exchange. This was followed by the formation of a new lysosome at 69–157 s, indicated by arrows. At 600 s, separation and content exchange were evident, given that the lysosome content (green fluorescence) was present in the endosome (hollow arrows) and vice versa. Lysosome (arrowheads). A newly formed lysosome (solid arrows). An endosome that gained lysosomal content (hollow arrows). Scale bar, 2.5 μm. (B) Computational detection of lysosome-endosome interaction in the example shown in (A). (Upper row) The actual lysosome-endosome pair. (Bottom row) Computational detection results. Frame rate during imaging: 4.2 s per frame. Scale bars, 1 μm. (C) Our active transport-mediated clustering model.