TRAIL signaling promotes entosis in colorectal cancer

Abstract

Entosis is a form of nonphagocytic cell-in-cell (CIC) interaction where a living cell enters into another. Tumors show evidence of entosis; however, factors controlling entosis remain to be elucidated. Here, we find that besides inducing apoptosis, TRAIL signaling is a potent activator of entosis in colon cancer cells. Initiation of both apoptosis and entosis requires TRAIL receptors DR4 and DR5; however, induction of apoptosis and entosis diverges at caspase-8 as its structural presence is sufficient for induction of entosis but not apoptosis. Although apoptosis and entosis are morphologically and biochemically distinct, knockout of Bax and Bak, or inhibition of caspases, also inhibits entotic cell death and promotes survival and release of inner cells. Analysis of colorectal cancer tumors reveals a significant association between TRAIL signaling and CIC structures. Finally, the presence of CIC structures in the invasive front regions of colorectal tumors shows a strong correlation with adverse patient prognosis.

Figure 1.

Single-cell time-lapse microscopy reveals simultaneous induction of entosis and apoptosis upon TRAIL stimulation. (A–D) HCT116 cells exhibit heterogeneity in caspase activation kinetics in response to TRAIL but not TRAIL + CHX. (A) Representative time-lapse images of HCT116 cells stably expressing IETD-FRET probe treated with TRAIL + CHX. Venus, TMRM, and CFP/FRET emission ratio images are shown. Depicted area highlights an apoptotic cell. Scale bar: 20 μm. (B) Schematic representation of the intact (left) and the cleaved (right) forms of IETD-FRET probe. Ex is the excitation wavelength; see Materials and Methods for details. (C) Quantification of single-cell CFP/FRET emission ratio traces in response to TRAIL (n = 57, left) and TRAIL + CHX (n = 60, right). (D) Single-cell IETD substrate cleavage kinetics in response to TRAIL and TRAIL + CHX. Data are shown as individual values for each cell as well as median and quartiles from three experiments. ****, P < 0.0001 by unpaired two-tailed t test. (E–H) Live cell TMRM staining accumulates in inner cells during entotic cell death. Representative time-lapse images of control (E and F) and TRAIL-treated (G and H) HCT116 IETD showing formation of entotic structures. DIC, Venus (green), TMRM (red), and CFP/FRET emission ratio images are shown. Representative entosis events are highlighted in a circle (entotic 1), a square (entotic 2), and a hexagon (entotic 3). Scale bar: 20 μm. Combined Venus (green) and TMRM (red) images highlighting stages of depicted entosis events (H). White-dashed and yellow-dashed lines indicate inner cells and outer cells, respectively. White asterisk indicates another entotic event in an outer cell of interest. Entotic 1, inner cell shows apoptotic features, then undergoes entotic cell death; entotic 2, inner cell shows caspase activation and loss of TMRM signal, then undergoes entotic cell death; entotic 3, inner cell with caspase activation invades into another cell, then is released. Scale bar: 20 μm. (I–K) Outer cells exhibit slow caspase activation kinetics and show less cleaved caspase-3 levels in TRAIL treatment. (I) Quantification of CFP/FRET emission ratio traces of inner (black) and outer (gray) cells in entotic 1, entotic 2, and entotic 3. Dashed line with arrow indicates time of internalization. Red dot on the single-cell trace of interest represents time of TMRM intensity loss. (J) Quantification of maximum mean values of CFP/FRET emission ratio traces in outer cells, apoptotic cells, or neighboring cells in HCT116 treated with or without TRAIL. Single-cell CFP/FRET emission ratio traces of n = 52 cells (13 cells/group) were generated over 20 h of time-lapse microscopy, and the maximum value in each trace was recorded. Data are shown as individual values for each cell as well as median and quartiles from three experiments. ****, P < 0.0001 by one-way ANOVA followed by Tukey’s multiple comparison test. (K) Quantification of immunofluorescence staining of cleaved caspase-3 (Asp175) levels in apoptotic cells (n = 36) and outer cells (n = 30) in HCT116 treated with TRAIL. Entosis events were characterized using 3D confocal microscopy images of DIC and Hoechst by detecting a round-shaped Hoechst-stained cell inside a vacuolar structure within another Hoechst-stained cell showing a crescent-shaped nuclear morphology. Cells showing fragmented nuclei and increased cleaved caspase-3 intensity were considered apoptotic. Data are shown as individual values for each cell as well median and quartiles from three experiments. ****, P < 0.0001 by unpaired two-tailed t test. MFI, mean fluorescence intensity.

Figure S1.

Long-term time-lapse microscopy verifies induction of entotic structures by TRAIL treatment. (A) Representative time-lapse images of Hoechst and LysoTracker staining in HCT116 cells showing formation of entotic structures in TRAIL treatment. Representative entosis events of an inner cell undergoing entotic cell death (1) and an inner cell being released (2) are highlighted. Formation and stages of depicted entosis events are shown. Inner cells and outer cells are highlighted with white arrows. (B) Quantification of relative cell movement of inner and outer cells in TRAIL-treated HCT116 cells. Thirteen entotic structures per group from three experiments were monitored by time-lapse microscopy. Data are shown as individual values for each cell as well as median and quartiles. ****, P < 0.0001 by unpaired two-tailed t test. (C) Representative time-lapse images of DIC, Venus, LysoTracker, and TMRM showing an inner cell being released (left) and inner cell undergoing entotic cell death (right). Accumulation of TMRM and LysoTracker coincides with the loss of Venus signal in the entotic inner cell (right). Solid and dashed lines indicate inner and outer cells, respectively. Scale bar: 20 μm. (D) Representative time-lapse images of DIC, Hoechst, TMRM, and LysoTracker showing the effects of FCCP treatment on inner cell TMRM signal during entotic cell death. Quantification of TMRM mean fluorescence intensity in the corresponding inner, outer, and neighboring cells. Yellow arrows indicate inner cell. Scale bar: 20 μm. (E) Representative 3D confocal microscopy images of Hoechst (blue) and cleaved caspase-3 (Cl-Casp3; red) staining showing Cl-Casp3–positive and Cl-Casp3–negative inner cells in control and TRAIL-treated cells. Quantification of inner cell Cl-Casp3 status in control and TRAIL-treated cells. White and black arrowheads indicate inner and outer cells, respectively. Scale bar: 20 μm. Data are shown as mean ± SEM from three experiments. *, P < 0.05 by unpaired two-tailed t test. IC, inner cell; Norm., normalized; OC, outer cell.

Figure 2.

Large-scale HCS-based quantification of simultaneous induction of entosis and apoptosis upon TRAIL and TRAIL + CHX stimulation. (A–D) Overview of steps involved in HCS-based analysis of entotic structures. After seeding (7,500 cells/well), staining (Hoechst, LysoTracker, PI) and treatments (A), 36 random fields of view per treatment per experiment were recorded for 72 h (B). CellProfiler defined cell nuclei, lysosomes, and PI-positive cells, then overlaid images were used to detect entosis events (C). Entosis events detected by CellProfiler were manually verified using overlaid images of brightfield, Hoechst (blue), LysoTracker (magenta), PI (cyan), and Venus (yellow) by analyzing Hoechst-stained and PI-negative inner cells within another Hoechst-stained and PI-negative cell showing a crescent-shaped nuclear morphology (D). Both inner cells that do (arrows) and do not (dashed arrows) show LysoTracker accumulation were quantified as entotic. Scale bars: 25 μm. (E and F) TRAIL, but not TRAIL + CHX, promotes entosis in HCT116 cells. Quantification of PI-positive cells (E) and entotic structures (F) in HCT116 cells treated with or without CHX, TRAIL, and TRAIL + CHX for 72 h. Data are shown as individual values for each experiment as well as mean ± SEM (n = 3). Asterisks on top of individual bars indicate comparisons with control. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****P < 0.0001, by one-way ANOVA followed by Tukey’s multiple comparison test.

Figure S2.

Representative images and manual quantification of entotic structures in HCT116, MCF-7, and LS180 cells as well as HCT116 spheroids with or without TRAIL treatment. (A) Representative confocal microscopy images of Hoechst (blue) and β-catenin (red) staining and quantification of entotic structures in HCT116 cells treated with or without TRAIL in the absence or presence of z-VAD-fmk or Y-27632. More than 500 cells were manually quantified from at least two experiments. White and yellow arrowheads indicate inner and outer cells, respectively. Scale bar: 20 μm. (B) Quantification of entotic events in MCF-7 and LS180 cells treated with or without TRAIL. Representative confocal microscopy images of DIC, Hoechst (blue), and β-catenin (red) staining in MCF-7 cells treated with control or TRAIL are shown. Yellow arrows indicate entosis events. More than 500 cells were quantified from at least two experiments. Scale bar: 20 μm. (C) Quantification of entotic events in HCT116 spheroids treated with control or TRAIL (left).3D projections of light sheet fluorescence imaging of Venus (green), Hoechst (blue), and LysoTracker (red) in control and TRAIL-treated HCT116 spheroids. Three spheroids per treatment were quantified. Yellow arrows indicate LysoTracker-positive inner cells. Scale bar: 100 μm. Data are shown as mean ± SD. **, P < 0.01; ****, P < 0.0001 by unpaired two-tailed t test.

Figure 3.

3D confocal microscopy and CLEM analysis of structural features of entosis in TRAIL-stimulated HCT116 cells. (A and B) 3D confocal microscopy imaging of β-catenin (red), Hoechst (blue), and LysoTracker (green) confirms complete cell internalization in control (A) and TRAIL-treated (B) cells. Representative confocal images of early-stage (top) and late-stage (bottom) entotic structures are shown. Orthogonal views through z-stacks of depicted areas are shown next to each image. Scale bars: 10 μm. (C–E) CLEM analysis of entotic ultrastructures in TRAIL-treated cells. Representative 3D confocal microscopy images of Hoechst (blue) and LysoTracker (green) staining of early (C), middle (D), and late (E) stages of entosis events are shown (left). Scale bar: 10 μm. TEM images of corresponding entosis events are shown (right). Two depicted areas (white and red squares) imaged in higher magnifications are shown. Green, magenta, and yellow pseudocolored areas indicate lysosomes, autophagosomes, and mitochondria, respectively. Blue pseudocolored areas indicate nuclei. Scale bars in TEM images from left to right: 10 μm, 2 μm, and 1 μm (C2) or 500 nm (D2 and E2). (F and G) Late-stage inner cells undergo cathepsin B–mediated lysosomal cell death. (F) Representative time-lapse images of LysoTracker staining (green) in HCT116 cells stably expressing Lamp1-mScarlet plasmid (red) showing an inner cell undergoing entotic cell death. Scale bar: 10 μm. (G) Representative 3D confocal microscopy images of fluorogenic cathepsin B substrate (red), DRAQ5 (blue), and LysoTracker (green) of early- and late-stage entotic cells are shown. Scale bars: 10 μm.

Figure S3.

Ultrastructural localization of LC3 in negative control and middle-stage inner cell shown in Fig. 3. (A) Representative TEM images showing ultrastructural localization of LC3 in negative control and middle-stage inner cell shown in Fig. 3 D. Yellow circles indicate gold particles. Quantification of gold particles in early-, middle-, and late-stage inner and outer cells shown in Fig. 3. Statistical significance was tested using one-way ANOVA followed by Tukey’s multiple comparison test. (B) Quantification of LAMP1 and LysoTracker mean fluorescence intensity (MFI) in depicted inner cell undergoing entotic cell death shown in Fig. 3 F. (C) Quantification of cathepsin B activity in early- and late-stage entotic cells corresponding to Fig. 3 G. n = 37 (early stage) and n = 28 (late stage) inner cells were quantified from three experiments. Data are shown as individual values for each cell as well as median and quartiles. ****, P < 0.0001 by unpaired two-tailed t test. IC, inner cell; OC, outer cell.

Figure 4.

Entosis induced by TRAIL requires death receptors and structural presence of caspase-8. (A–C) Internalization process during entosis is independent of caspase activity. (A) Inhibition of entosis does not affect the rate of TRAIL-induced apoptosis. Quantification of PI-positive cells in HCT116 cells treated with or without TRAIL in the absence or presence of z-VAD-fmk, z-IETD-fmk, or Y-27632. (B and C) Inhibition of caspase activity does not affect the rate of TRAIL-induced entosis. (B) Quantification of entotic structures in HCT116 cells treated with or without TRAIL in the absence or presence of z-VAD-fmk, z-IETD-fmk, or Y-27632. (C) Representative field of view from HCS-based entosis analysis is shown. Arrows indicate entotic events, four representative events are depicted and shown. Scale bar: 50 μm. (D–H) TRAIL death receptors are required for cell internalization during TRAIL-induced entosis. (D) Quantification of PI-positive cells in HCT116 WT and DR4^−/− DR5^−/− cells treated with or without TRAIL. (E) Western blot images of DR4 and DR5 expressions in HCT116 WT and DR4^−/− DR5^−/− cells. β-Actin was used as loading control. (F) Quantification of entotic structures in HCT116 WT and DR4^−/− DR5^−/− cells treated with or without TRAIL in the absence or presence of Y-27632. (G and H) DR4 and DR5 are required for inner cells during cell internalization. (G) HCT116 WT and DR4^−/− DR5^−/− cells were labeled with CellTracker Green and CellTracker Red, mixed in 1:1 ratio, and treated with or without TRAIL in the absence or presence of z-VAD-fmk. Images were recorded by HCS and analyzed using CellProfiler. Scale bar: 25 μm. Distribution of entosis events divided into four subcategories as indicated; absolute numbers are shown in parenthesis. Data are shown as mean from three experiments. n indicates number of events analyzed per group. (H) Representative images showing entotic structures formed between “wt in wt” (green) and “wt in −/−” (green and red, respectively) cells. Scale bar: 10 μm. (I–K) Induction of entosis by TRAIL requires structural presence of caspase-8. (I) Quantification of PI-positive cells in HCT116 WT and Casp8^−/− cells treated with or without TRAIL. Data are shown as mean ± SEM from three experiments. (J) Western blot images of CASP8 expression in HCT116 WT, CASP8 C360A mutant, and CASP8^−/− cells. β-Actin was used as loading control. (K) Quantification of entotic structures in CASP8^−/− and CASP8 C360A mutant cells treated with or without TRAIL in the absence or presence of Y-27632. Data are shown as individual values for each experiment as well as mean ± SEM from three experiments except as noted in G and I. ****, P < 0.0001 by one-way ANOVA followed by Tukey’s multiple comparison test (A, B, D, F, G, I, and K).

Figure S4.

Representative fields of view and quantification of cells with cleaved IETD substrate in HCT116 cells treated with or without TRAIL in the absence or presence of Y-27632 or caspase inhibitors. (A) Representative images and quantification of cells with cleaved IETD substrate in HCT116 cells treated with or without TRAIL in the absence or presence of z-VAD-fmk, z-IETD-fmk, or Y-27632. Scale bar: 25 μm. Data are shown as individual values for each experiment as well as mean ± SEM from three experiments. ****, P < 0.0001 by one-way ANOVA followed by Tukey’s multiple comparison test. (B) Representative images of Hoechst and LysoTracker staining in HCT116 cells treated with or without TRAIL in the absence or presence of z-VAD-fmk or Y-27632. Arrows indicate entosis events. Scale bar: 25 μm. (C) Representative images of Hoechst-stained (blue) HCT116 CASP8^−/− cells (pX330-Cas9-sgCASP8; green). Quantification of entotic events in control, TRAIL-treated, or glucose-starved HCT116 WT and CASP8^−/− cells. More than 500 cells were manually quantified from two experiments. Scale bar: 20 μm. Data are shown as mean ± SD. *, P < 0.05 by one-way ANOVA followed by Tukey’s multiple comparison test.

Figure 5.

Deletion of Bax and Bak or inhibition of caspase activation alters inner cell fate toward release. (A) Schematic representation of inner cell fates and representative time-lapse images showing that the inner cell undergoes entotic cell death, the inner cell is released, and the inner cell completes cell division. Inner and outer cells are indicated as orange and white dashed lines, respectively. Scale bar: 10 μm. (B) Western blot images of Bax and Bak expressions in HCT116 WT and Bax^−/− Bak^−/− cells. β-Actin was used as loading control. (C) Quantification of inner cell fates in HCT116 WT and Bax^−/− Bak^−/− cells treated with or without TRAIL in the absence or presence of z-VAD-fmk. Data are shown as individual values for each experiment as well as mean ± SEM. n is the total number of cells analyzed over 48 h from three experiments. Asterisks inside individual bars indicate comparisons with control. ****, P < 0.0001 by one-way ANOVA followed by Tukey’s multiple comparison test. (D) Representative 3D confocal microscopy images of DIC, Hoechst (blue), LysoTracker, and fluorogenic cathepsin B substrate of late-stage entotic cells in WT and Bax^−/− Bak^−/− cells treated with or without TRAIL in the presence or absence of z-VAD-fmk. 3D colocalization of LysoTracker and cathepsin B substrate is visualized. Yellow arrows indicate inner cells. Scale bars: 10 μm. (E) Quantification of cathepsin B activity in late-stage entotic inner cells in WT and Bax^−/− Bak^−/− cells treated with or without TRAIL in the presence or absence of z-VAD-fmk. Data are shown as individual values for each cell as well as median and quartiles. Statistical significance was tested using unpaired two-tailed t test. n is the total number of cells analyzed. CatB, cathepsin B; IC, inner cell; MFI, mean fluorescence intensity; OC, outer cell.

Figure 6.

Clinical association among TRAIL signaling, CIC structures, and CRC. (A–C) CIC structures are increased in CRC tumors compared with matched normal tissues. (A) Consecutive H&E (top) and hematoxylin/c-Met (H&c-MET; bottom) sections showing representative CIC structures in a CRC tumor. Plain arrows indicate inner cells, and dashed arrows indicate outer cells. Scale bar: 10 μm. (B) Inter- and intrapatient heterogeneity in CIC events detected in tumor tissue and computed for each patient of the NI240 cohort (n = 223). Patients (x-axis) are sorted in decreasing order of median CIC events (y-axis) detected in individual TMA cores prepared from tumor tissue and stained with either H&E or c-MET. Line and shaded area indicate the median and the minimum and maximum number of CIC structures across the cores for each patient, respectively. (C) Quantification of CIC structures in TMA sections prepared from tumor center, invasive front, and matched normal tissues. n_c and n_pt indicate number of tumor cores and patients, respectively. ****, P < 0.0001. (D) Schematic representation of the absence or presence of CIC structures and the center and the invasive front regions of a tumor. (E) Overexpression of proteins involved in TRAIL signaling is associated with presence of CIC structures in CRC tumors. Association among expression levels of TRAIL, DR4, DR5, caspase-8, c-FLIP, and CIC structures. Relative protein expression between CIC absent (median CIC = 0) and CIC present (median CIC >0) tumors is shown. Statistically significant differences in protein expression between CIC absent and CIC present patients are indicated by solid line (P < 0.05) and dashed line (0.05 ≤ P < 0.1). Individual P values are shown under each protein. *, P < 0.05. (F–H) Presence of CICs in the invasive front region of tumors is an independent prognostic marker for DFS and DSS in CRC. Kaplan–Meier estimates of DFS and DSS of the NI240 cohort (F) comparing stage 2 (G) and stage 3 (H) CRC patients grouped by the presence or absence of CIC structures in the invasive front region. *, P < 0.05. IC, inner cell; OC, outer cell.

Figure S5.

Overview of clinical, demographic, pathological, and molecular information for the CRC patients of the NI240 phase III clinical trial. (A) Each column represents a patient and each row color codes a feature. Missing data are shown in white. CIC-derived features include estimates of absence (CIC = 0) or presence (CIC > 0) of CIC events from TMA sections prepared from tumor, invasive front, and normal tissue. Visualization was generated with the MATLAB package HCP. (B) Comparison between CIC events (median aggregated by patient, tissue, and staining marker) observed in TMA sections stained for H&E and c-MET. Marker color indicates tissue type. Marker size and transparency encode the number of cores. Solid black line and gray shaded area indicate the regression line and CI, respectively. Agreement between CIC estimates from H&E-stained and c-MET–stained TMA cores was computed using the Kendall τ correlation (Python package scipy). (C) CIC events (median and 95% CI) detected in TMA sections prepared from tumor center, invasive front, and matched normal tissue of colon or rectum. n_c and n_pt indicate number of tumor cores and patients, respectively. LVI, lymphovascular invasion.