Functional visualization of NK cell-mediated killing of metastatic single tumor cells
- Research Center for Dynamic Living Systems, Graduate School of Biostudies, Kyoto University, Japan
- Department of Pathology and Biology of Diseases, Graduate School of Medicine, Kyoto University, Japan
- Department of Nanopharmaceutical Sciences, Nagoya Institute of Technology, Gokiso-cho, Japan
- Brain Science Institute, Center for Brain Science, RIKEN, Hirosawa, Japan
- Laboratory for Mouse Genetic Engineering, RIKEN Center for Biosystems Dynamics Research, Japan
- Institute for Integrated Cell-Material Sciences, Kyoto University, Japan
Figures
Figure 1 with 4 supplements
NK cells eliminate a subset of metastatic tumor cells from the lung within 24 hr of arrival.
(A, B) B6 mice were pretreated with either control antibody or αAGM1. Representative merged images of the bright field and the bioluminescence images of mice intravenously injected with 5×105 B16-Akaluc cells are shown. (A) The substrate was i.p. administered immediately after injection of tumor cells. Image acquisition was started at 5 min after tumor injection. See also Figure 1—video 1. Bioluminescence intensity (BLI) is normalized to that at 5 min and plotted over time. Data are representative of two independent experiments with three mice per group and are shown as means ± SD. A dotted line represents 20 min. (B) Substrate was administered i.p. before each round of image acquisition. BLI is normalized to that at 1 hr. Data are representative of three independent experiments with 4–6 mice per group and shown as means ± SD. (C) B16-Akaluc cells were injected into the tail vein of mice that had been inoculated with B16F10 cells in the footpad 14 days before. BLI was quantified at the indicated time and normalized to that at 1 hr after injection of the B16-Akaluc cells. Data are representative of two independent experiments with three mice per group and are represented as means ± SD. (D) Nude mice were pretreated with either control antibody or αAGM1. 4T1-Akaluc cells were injected into the tail vein. BLI was quantified at the indicated time and normalized to that at 1 hr after injection of the tumor cells. Data are representative of two independent experiments with four mice per group and are represented as means ± SD. (E) BALB/c mice were pretreated with either control antibody or αAGM1. Shown are representative merged images of the bright field and the bioluminescence images of mice subcutaneously injected with 5×105 4T1-Akaluc cells into footpad. An arrow and asterisks depict a lung metastasis and primary tumors, respectively. Lung metastasis incidence of control antibody- (n=9) or αAGM1- (n=8) treated mice. (F, G) Identical to (E) except that mice are nude mice and the implanted cell number is 1×104. (F) BLI of primary tumor. Data are representative of two independent experiments with three mice per group and are represented as means ± SD. (G) Lung metastasis incidence of control antibody- (n=11) or αAGM1- (n=18) treated mice. NK, natural killer.
Figure 1—figure supplement 1
Depletion of NK cells by αAGM1.
Flow cytometric analysis of the spleen (A) or lung (B) of mice treated with isotype control antibody or αAGM1. The numbers over the boxes indicate the percentage of NK1.1-positive cells among live single cells. Data are representative of three mice each.
Figure 1—figure supplement 2
NK cells eliminate metastatic tumor cells from the lung within 24 hr.
BrafV600E melanoma (A), MC-38 cells (B), and 4T1 cells (C) expressing Akaluc were injected into the tail vein of C57BL/6 (A, B) or BALB/c (C) mice treated with either control antibody or αAGM1. BLI was quantified at the indicated time and normalized to that at 1 hr after injection of the tumor cells. Data are representative of three (A, B) or two (C) independent experiments with 3–6 mice per group and are shown as means ± SD. BLI, bioluminescence intensity; NK, natural killer.
Figure 1—figure supplement 3
Basophils, macrophages, and neutrophils do not contribute to elimination of metastatic tumor cells.
Flow cytometric analysis of the spleens and bioluminescence imaging in mice depleted of basophils (A, B), macrophages (C, D), and neutrophils (E, F), by αCD200R3 antibody, clodronate liposome, and αLy-6G antibody, respectively. Data are representatives of three mice each for flow cytometric analysis and two independent experiments with three mice per group and are shown as means ± SD for bioluminescence imaging.
Figure 1—video 1
Acute rejection of metastatic tumor cells by NK cells.
Related to Figure 1A. Merged images of the bright field and the bioluminescence images of mice intravenously injected with 5×105 B16-Akaluc cells. Substrate was i.p. administered immediately after injection of tumor cells. Image acquisition was started at 5 min after tumor injection. Bioluminescence intensity is displayed in pseudo-color. NK, natural killer.
Figure 2 with 2 supplements
NK cells patrol pulmonary capillaries in a stall-crawl-jump manner.
(A) A schematic of the intravital imaging system for the lung. The left lobe of the lung was exposed by 5th or 6th intercostal thoracotomy using custom-made retractors and fixed to the objective by a vacuum‐stabilized imaging window. (B) (Left) A micrograph of the lung of an NK-tdTomato mouse, in which NK cells express tdTomato (magenta). Lectin (green) was injected intravenously to stain endothelial cells. (Right) A magnified image of the boxed region in the left panel. (C) A representative time-lapse image of NK cells (magenta). The track of an NK cell is shown with a cyan dotted line and white arrowheads at both ends. White and yellow dotted circles show the positions of stall and jump, respectively. See also Figure 2—video 1. (D) Distribution of the crawling duration times in a 0.25-mm2 field of view (FOV). Data are pooled from two independent experiments (n=583). (E, F) B6 mice expressing tdTomato in NK cells (magenta) were observed by 2P microscopy. During time-lapse imaging with a 30-s interval, 100 µg of αLFA-1α, αMac-1, or isotype control antibody was intravenously injected. The number of NK cells in the 0.25-mm2 FOV was counted 0–10 min before and 30 min after antibody injection. The percentage of NK cells after versus before antibody injection is shown in (F). Data were pooled from three independent experiments and represented as means ± SD. n=3 mice for each group. 2P, two-photon; NK, natural killer.
Figure 2—figure supplement 1
Intravascular staining of NK cells.
Flow cytometric analysis of the lung (top) and bone marrow (bottom) of mice intravenously injected with αCD45 BV510 antibody and counterstained ex vivo with αCD45 FITC antibody. Left, Histogram of ex vivo CD45 expression on a live single-cell gate. Center, Counter plots of CD3 and NCR1 expression on ex vivo CD45+ cells. Right, Histogram of intravenously injected CD45 antibody on CD3− NCR1+ cells. Data are representative of three mice each. NK, natural killer.
Figure 2—video 1
Stall-crawl-jump movement of NK cells in the pulmonary capillary.
Related to Figure 2C. An example of an NK cell that moves within a pulmonary capillary in a stall-crawl-jump manner. A white arrow points to an NK cell showing a stall-crawl-jump movement. NK, natural killer.
Figure 3
NK cells patrol capillaries deliberately in the presence of melanoma.
NK-tdTomato mice were injected with 1.5×106 B16-SCAT3 cells and observed under a 2P microscope for 2 hr from 6 hr after injection. NK-tdTomato mice without any treatment were used as the control. SHG stands for second-harmonic generation. (A) A micrograph of the lung of an NK-tdTomato mouse. B16-SCAT3 cells, green; NK cells, magenta. (B) The average number of NK cells and tumor cells in each FOV at 10–35 µm from the pleura 10 min after intravenous injection of B16F10 cells (n=9). The red lines represent the mean. (C) Trajectories of crawling NK cells in the presence (left) or absence (right) of B16F10 cells. For 3D tracking, images of a 0.25-mm2 FOV and 25 µm thickness at 10–35 µm from the pleura were acquired every 30 s for 120 min. Shown here are the trajectories of NK cells projected onto the XY plane. Each track is shown in pseudo-color based on track duration. Data are at least from two independent experiments for each condition. n=1127 cells in the absence and n=718 cells in the presence of tumor cells. (D–F) Shown are mean squared displacement (MSD) (D), instantaneous speed (E), and track duration (F). The statistical differences between the two experimental groups were assessed by Mann–Whitney U-test. 2P, two-photon; FOV, field of view; NK, natural killer.
Figure 4 with 5 supplements
Intravital 2P imaging with biosensors visualizes apoptosis and calcium influx of tumor cells induced by crawling, but not flowing, NK cells.
(A) A representative time-lapse image of a lung of an NK-tdTomato mouse after B16-SCAT3 cell injection. An NK cell and B16-SCAT3 cells are depicted in magenta and green, respectively (top). Bottom, the CFP/FRET ratio in B16-SCAT3 cells is shown in the intensity-modulated display (IMD) mode and an NK cell is shown in white. See also Figure 4—video 1. (B) Quantification of the CFP/FRET ratio in a B16-SCAT3 cell in (A). (C) The percentage of NK cell-tumor cell contacts with or without caspase activation. Data were pooled from three independent experiments. (D) Time intervals between NK cell contact and caspase three activation in B16-SCAT3 cells. Data are pooled from three independent experiments. (E) A representative time-lapse image of a lung of NK-tdTomato mice after B16-GCaMP cell injection. An NK cell and B16-GCaMP cell are depicted in magenta and green, respectively (top). Bottom, GCaMP6s intensity in a B16-GCaMP cell is displayed in pseudo-color and an NK cell is shown in white. See also Figure 4—video 3. (F) Quantification of GCaMP6s intensity shown in (E). (G) Time intervals between NK cell contact and Ca2+ influx in B16-GCaMP cells. Data were pooled from four independent experiments. Red lines represent the median. (H) Comparison of the number of NK cell contacts that were followed by Ca2+ influx between the WT and Necl5−/− Nectin2−/−. Data were pooled from four (WT) and two (Necl5−/− Nectin2−/−) independent experiments. (I) In seven independent experiments, 40 contact events with calcium influx were observed and classified into those caused by crawling or flowing NK cells. (J) αLFA-1α or isotype control antibody was intravenously administered 2 hr before injection of 5×105 B16-Akaluc cells. The bioluminescence signals are normalized to those of 1 hr. Data are representative of two independent experiments with 3–4 mice per group and presented as means ± SD. (K) Representative macroscopic images of the metastasis to the lung and number of metastatic nodules per lung are shown. Red lines represent the median. Data were pooled from two independent experiments. Control, n=7; αLFA-1α, n=8. 2P, two-photon; NK, natural killer.
Figure 4—figure supplement 1
NK cell-induced Ca2+ influx in B16F10 cells in vitro.
(A) A representative time-lapse image of the interaction between an NK cell (arrow) and a B16-R-GECO cell (arrowhead) in vitro. Shown here are merged images of differential interference contrast, YFP fluorescence (NK cell, green) and R-GECO1 fluorescence (B16-R-GECO cell, white). (B) Time course of R-GECO1 intensity. (C) R-GECO1 intensity in each cell is normalized to the intensity at the start of imaging and displayed as a heatmap. n=43 cells from two independent experiments. (D) Percentage of deceased tumor cells that exhibited Ca2+ influx after NK cell engagement. n=103 from six independent experiments. NK, natural killer.
Figure 4—figure supplement 2
Absence of LFA-1 ligands on B16F10 cells.
Flow cytometric analysis of the expression of ICAM-1 (A) and ICAM-2 (B) on the B16F10 cells.
Figure 4—video 1
Induction of caspase three activation by crawling NK cells.
Related to Figure 4A. A time-lapse movie of the lung of an NK-tdTomato mouse after intravenous injection of B16-SCAT3 cells. The CFP/FRET ratio in B16-SCAT3 cells is shown in the IMD mode and an NK cell is shown in white. IMD, intensity-modulated display; NK, natural killer.
Figure 4—video 2
Example of morphological changes of a target cell following Ca2+ influx mediated by NK cell.
Related to Figure 4—figure supplement 1A. A time-lapse movie of the in vitro killing assay. The differential interfering contrast (DIC) and the R-GECO are shown. White arrow indicates the NK cell which induces the Ca2+ spike in the tumor cell. Cell death is ultimately induced in the cell after blebbing. NK, natural killer.
Figure 4—video 3
Induction of calcium influx by crawling NK cells.
Related to Figure 4E. Intravital imaging of the pulmonary capillary of NK-tdTomato mice after intravascular injection of B16-GCaMP cells. GCaMP6s intensity is displayed in pseudo-color. An NK cell is shown in white. A dotted line shows the melanoma cell that exhibits Ca2+ influx after contact with an NK cell. NK, natural killer.
Figure 5 with 2 supplements
Contact-induced ERK activation in NK cells is a necessary event in induction of apoptosis in tumor cells in the first 4 hr, but not after 24 hr.
(A) B6 mice were pretreated with either control antibody or αAGM1 and intravenously injected with 5×105 B16-Akaluc cells. At 1 hr before and 8 hr after injection, a MEK inhibitor (MEKi) or DMSO was administered i.p. Time courses of the signals, which are normalized to those at 1 hr after tumor injection for each mouse. Data are representative of two independent experiments and shown as means ± SD. n.s., not significant. (B) Macroscopic images were acquired at day 14. The number of metastatic colonies is shown. Control, n=5; MEKi, n=7. (C) A time-lapse image of the lung of an NK-ERK mouse expressing the FRET biosensor for ERK. The mouse was intravenously injected with B16-GCaMP cells. Top, FRET/CFP ratio of an NK cell (yellow arrowhead) is shown in IMD mode. A B16-GCaMP cell is shown in white. Bottom, GCaMP6s intensity is displayed in pseudo-color. The NK cell is shown in white. (D) Time course of the FRET/CFP ratio in the NK cell and CaMP6s intensity in the B16-GCaMP cell. (E) Activation probability of ERK in the NK cells upon target cell contact at 0–4 hr or 24 hr after tumor injection. Data were pooled from three independent experiments. (F) The probability of NK cells that exhibited ERK activation with or without induction of Ca2+ influx in the target tumor cells at 0–4 hr or 24 hr after tumor injection. Data were pooled from three independent experiments. (G) B16-Akaluc cells were injected into the tail vein of mice that had been injected PBS or B16F10 cells into tail vein 24 hr before. BLI was quantified at the indicated time and normalized to that at 1 hr after injection of B16-Akaluc cells. Data are representative of two independent experiments with three mice per group and are represented as means ± SD. BLI, bioluminescence intensity; FRET, Förster resonance energy transfer; IMD, intensity-modulated display; NK, natural killer; PBS, phosphate-buffered saline.
Figure 5—figure supplement 1
DNAM-1-mediated ERK activation in the killer NK cells in vitro.
(A) NK cells derived from hyBRET-ERK-NLS mice were cultured with B16-R-GECO and observed under an epifluorescence microscope. Quantification of the FRET/CFP ratio in the NK cells that induced apoptosis (killer cells) and those that failed to induce apoptosis (non-killer cells) in the target cells. Data were pooled from six independent experiments and are shown as median ± SD; n=43 cells for killer cells and n=73 cells for non-killer cells. (B) Induction of apoptosis in the target cells by NK cells with or without ERK activation. Data are from six independent experiments. (C) NK cells were cultured with B16-R-GECO cells in the presence or absence of MEKi. Percentages of target cell death are shown. Data are pooled from three independent experiments and represented as means ± SDs. (D) NK cells derived from hyBRET-ERK-NLS mice are sorted by the expression of DNAM-1. The DNAM-1+ or DNAM-1− NK cells were cultured with B16-R-GECO cells. Data were pooled from two independent experiments and are represented as median ± SD; n=37 cells for DNAM-1+ cells and n=27 cells for DNAM-1− NK cells. (E) B16F10 cells or B16F10 Necl5−/− Nectin2−/− cells were stained with DNAM-1 Fc. The gray histogram is the background staining with a secondary Ab only. (F) The DNAM-1+ or DNAM-1− NK cells derived from hyBRET-ERK-NLS mice were cultured with B16F10 Necl5−/− Nectin2−/− cells. Data were pooled from two independent experiments and are represented as median ± SD; n=38 cells for DNAM-1+ cells and n=41 cells for DNAM-1− NK cells. (G) NK cells were cultured with B16F10 cells, or Necl5−/− Nectin2−/− B16F10 cell clones, A7, B7, and E7. Percentages of target cell death are shown. Data are pooled from three mice and represented as means ± SDs.
Figure 5—figure supplement 2
In vivo dynamics of ERK activity in NK cells after target cell contact.
(A) Quantification of the FRET/CFP ratio in the NK cells that exhibited ERK activation after target cell contact. Data are pooled from two independent experiments and are represented as means ± SD, n=18 cells. (B) Time intervals between ERK activation in NK cells and Ca2+ flux in melanoma cells. Data were pooled from three independent experiments. A red line represents the mean. n = 43 cells. NK, natural killer.
Figure 6 with 4 supplements
Shedding of Necl5 correlates evasion of NK cell surveillance.
(A, B) B16-Akaluc cells were injected into the tail vein and the expression level of Necl5 on survived tumor cells was analyzed at 24 hr after dissemination. The MFI of Necl5 in tumor cells injected 0 hr or 24 hr before is shown in (B). Red lines represent the median. Data were pooled from two independent experiments. (C) Schematic representation of the Necl5-ScNeo fusion protein. (D) The representative images of mScarlet/mNeonGreen ratio in the B16F10 cells expressing Necl5-ScNeo at 0.5 hr and 24 hr after injection are shown in the IMD mode. The quantified mScarlet/mNeonGreen ratio in the transmembrane in indicated time point is shown in (E). Data were pooled from three animals. Mice were treated in their drinking water with 5 mg/L warfarin at least for 5 days and intravenously injected with 5×105 B16-Akaluc cells. The BLI at 24 hr, which normalized to those at 1 hr after tumor injection for each mouse are shown. Red lines represent the median. Data were pooled from two independent experiments. (G, H) Mice were intravenously injected with 5×105 B16-Akaluc cells. At 1 hr before and 12 hr after tumor injection, edoxaban, dabigatran etexilate, or vehicle was orally administered to mice. For NK cell depletion, mice were pretreated with either control antibody or αAGM1. The BLI at 24 hr, which normalized to those at 1 hr after tumor injection for each mouse is shown. Red lines represent the median. Data were pooled from two independent experiments. (I) Flow cytometric analysis of B16F10 cells treated with recombinant thrombin for 3 hr. Data are representative of two independent experiments with three wells per group and are represented as means ± SD. (J) Mice were orally administrated with dabigatran etexilate 1 hr before tumor injection and analyzed as in (D, E). Data were pooled from three animals. (K) Evasion of NK cell surveillance by shedding of Nec-l5. NK, natural killer.
Figure 6—figure supplement 1
Edoxaban promotes the elimination of disseminated tumor cells.
Mice were intravenously injected with 5×105 B16-Akaluc cells. At 1 hr before and 12 hr after injection, edoxaban or vehicle was orally administered to mice. Representative macroscopic images of the lung metastasis and the number of metastatic nodules at day 14 are shown. Red lines represent the median. Data were pooled from two independent experiments. Control, n=6 mice; edoxaban, n=7 mice.
Figure 6—figure supplement 2
Lack of micro-thrombus around the disseminated tumor cells.
In vivo imaging of pulmonary capillaries of a hyBRET-ERK-NES mouse, which was injected B16F10 cells expressing tdTomato-CAAX 4 hr before imaging. The representative merged image of host cells (Green) and B16F10 cells (Magenta) is shown, with a schematic view of this region. Asterisks represent the laser ablated regions. Ac, alveolar cavity; Ec, endothelial cell. The image is representative of two independent experiments with 22 FOVs. See also Figure 6—video 1.
Figure 6—figure supplement 3
Elimination of Necl5−/− Nectin2−/− cells.
Mice were intravenously injected with 5×105 wild-type or Necl5−/− Nectin2−/− B16-Akaluc cells. The BLI at 24 hr, which normalized to those at 1 hr after tumor injection for each mouse is shown. Red lines represent the median. Data were pooled from three independent experiments.
Figure 6—video 1
Thrombus-formation by laser ablation around the disseminated tumor cells.
Related to Figure 6—figure supplement 2. Intravital imaging of the pulmonary capillary of hyBRET-ERK-NES mouse, which was injected B16F10 cells expressing tdTomato-CAAX 4 hr before imaging. Host cells and tumor cells are shown in green and magenta, respectively.
Tables
Table 1
Dynamics of NK cell killing of melanoma cells in the lung.
| Symbols | Parameters | Values | Units | References, equations, and comments |
|---|---|---|---|---|
| Histological and physiological parameters from published papers | ||||
| Bv | Blood volume | 1.7E−6 | m3 | Table II (Davies and Morris, 1993) |
| Lv | Lung volume | 3.7E−07 | m3 | Figure 2A for Week 8 mice (Gomes et al., 2019) |
| CO | Cardiac output | 2.0E−05 | m3/min | Abstract (Janssen et al., 2002) |
| Cp_l | Total lung capillary length | 1.1E+03 | m | Result section (Knust et al., 2009) |
| Parameters determined experimentally | ||||
| FOV | Field of view | 2.5E+07 | m2 | 0.5×0.5 mm2 |
| Plt_speed | Platelet speed | 0.0570.95 | m/min mm/s | Determined as described previously (Sano et al., 2016). |
| Cp_r | Capillary radius | 3.4E−06 | m | Measured on the images of Figure 2 |
| NK_d | NK cell diameter | 1.0E−5 | m | Measured on the images of Figure 2 |
| NK_bl | NK cell count in blood | 5.1E+10 | cells/m3 | Determined for C57BL/6 mice of 8–12 weeks old. |
| NK_FOV | NK cell number in a field of view | 16 | cells/FOV | Measured on the images of Figure 4 |
| NK_speed | NK crawling speed on capillaries | 4.8E−06 | m/min | Determined with time-lapse images of Figure 4 |
| NK_hit_obs | Observed NK cell hit probability | 8.0E−3 | cells/min | Determined with time-lapse images of Figure 4 |
| NK_kill | NK cell killing probability | 0.5 | Determined with time-lapse images of Figure 4. | |
| Ml_hl_BLI | Melanoma half-life based on BLI data | 146 | min | From bioluminescence images of Figure 1 at 4 hr. |
| Calculated parameters | ||||
| Cp_fr | Capillary flow rate | 1.60E−12 | m3/min | Plt_speed*π*Cp_r2 |
| NK_wbl | Whole blood NK cells | 8.7E+04 | cells | NK_bl*Bv |
| NK_density | NK cell density in lung | 6.4E+12 | cells/m3 | NK_FOV/FOV/NK_d |
| NK_lung | Total NK cell in lung | 2.4E+06 | cells | NK_density*Lv |
| NK_in | NK cell influx to lung | 1.0E+06 | cells/min | CO*NK_bl |
| NK_out | NK cell decay constant | 0.42 | /min | NK_in/NK_lung |
| NK_hl | NK cell half-life in the lung | 1.6 | min | ln2/NK_out |
| NK_fr | NK cell flow rate per capillary | 0.116.6 | cells/min cells/hr | NK_bl*Cp_fr |
| NK_dcp | NK cell density on capillary | 2.1E+03 | cells/m | NK_lung/Cp_l |
| NK_hit_cr | Crawling NK cell hit probability | 0.0100.6 | cells/min cells/hr | NK_dcp*NK_speed |
| Ml_t_2P_obs | Melanoma decay constant calculated by2P imaging | 4.0E−03 | /min | NK_kill*NK_hit_obs |
| Ml_hl_2P_obs | Melanoma half-life based on 2P imaging data | 173 | min | ln2/Ml_t_2P_obs |
| Ml_t_2P_crsp | Melanoma decay constant based on crawling speed | 5.1E−03 | /min | NK_kill*NK_hit_cr |
| Ml_hl_2P_crsp | Melanoma half-life based on crawling speed | 137 | min | ln2/Ml_t_2P_crsp |
-
Basic parameters: Macroscopic and histological data were based on previous papers (Gomes et al., 2019; Janssen et al., 2002; Knust et al., 2009). The speed of platelets in the lung capillaries (Plt_speed) was determined as described previously (Sano et al., 2016). The speed of platelets in the pulmonary capillaries was roughly one-third of the speed of platelets in the arteriole of mouse bladder, 3.1 mm/s (0.186 m/min) (Sano et al., 2016). The capillary radius, diameter of NK cells, and number of NK cells in a field of view were determined on at least three images. Plt_speed and Cp_r were used to calculate the capillary flow rate (Cp_fr). To determine the total number of NK cells in the blood, 50 µL of blood was collected from the right ventricle of 8- to 12-week-old C57BL/6 mice, lysed in ACK buffer (155 mM/L NH4Cl, 10 mM/L KHCO3, and 0.1 mM/L EDTA), and analyzed by flow cytometry. The CD3− NK1.1+ cells were counted as NK cells. This number of NK cells is two- to threefold larger than that reported previously by using C57BL/6J mice (Banh et al., 2012). The mean crawling speed on the endothelial cells is described in the text related to Figure 3E. The probability of an NK cell hitting a tumor cell was determined by a MATLAB script (Main_191017.m). The probability of an NK cell killing a tumor cell was 0.5, based on the probability of induction of calcium influx in the target B16 melanoma cells (Figure 4H).
-
Total number of lung NK cells: The total number of NK cells residing in the lung (NK_lung) was estimated from the mean NK cell density (NK_density) and total lung volume (Lv). The diameter of an NK cell (NK_d) was used as the thickness of the image plane. The number of total NK cells, 2.4 million, is markedly larger than the previous values, which ranged from 0.2 to 1 million (Bi et al., 2017; Grégoire et al., 2007; Yan et al., 2014). In previous studies, the whole lungs were lysed to count the blood cell number. It is possible that the recovery rate might have been low due to insufficient tissue lysis. As described in the main text, we observed comparable numbers of tumor cells and NK cells in each FOV, when 1.5×106 B16-SCAT cells were injected into NK-tdTomato mice, supporting the fidelity of the number of total NK cells determined in this study.
-
Dynamics of NK cells: Most of the pulmonary NK cells are within the vasculature (Figure 1—figure supplement 3), and the number of pulmonary NK cells overwhelms that of NK cells in the blood. Thus, the total number of lung NK cells can be used as the total number of NK cells in the lung vasculature. NK cell influx into the lung (NK_in) is obtained from cardiac output (CO) and NK cell count in the blood (NK_bl). If all NK cells stay in the lung with equal probability, the apparent transit time in the lung, or NK cell half-life in the lung, is calculated as 1.6 min from NK_in and NK_lung. This value is markedly smaller than the tracking duration period observed in Figures 3 and 4, indicating that the major population of NK cells in the blood go through the lung without adhesion to the endothelial cells. By using the capillary flow rate (Cp_fr) and NK cell count (NK_bl), the NK cell flow rate per capillary (NK_fr) is determined as 0.11 cells/min. Meanwhile, from the total length of capillaries (Cp_l) and the number of NK cells (NK_lung), NK cell density on the capillary (NK_dcp) is determined as 2.1 cells/mm. From NK_dcp and the crawling speed of NK cells, the probability of a tumor cell being hit by crawling NK cells (NK_hit_cr) becomes 0.010 cells/min. This value is approximately one-tenth of the flow rate of NK cells (NK_fr).
-
Dynamics of disseminated melanoma cells: The BLI signals from 1 to 12 hr (Figure 1B) were fitted with MATLAB using the following equation: .
-
With this fitting, the decay rate decreases with time. Because we characterized the NK cell interaction with tumor cells between 4 and 8 hr after tumor cell injection, we determined the half-life of B16F-Akaluc cells from 4 hr (Ml_hl_BLI) and obtained 146 min. Meanwhile, from the probability of an NK cell hitting a tumor cell (NK_hit_obs) and the probability of an NK cell killing a tumor cell (NK_kill), the half-life of tumor cells (Ml_hl_2P_obs) becomes 173 min. If we adopt the probability of a tumor cell hit based on the crawling speed of NK cells, the expected half-life of tumor cells (Ml_hl_2P_crsp) becomes 137 min. Considering the precision of parameters obtained from in vivo imaging data, we believe that the half-life of melanoma cells estimated from the 2P microscopy reasonably matched the half-life of melanoma cells determined by BLI.
Additional files
-
Transparent reporting form
- https://cdn.elifesciences.org/articles/76269/elife-76269-transrepform1-v2.docx
-
Source code 1
Matlab scripts and data sets.
- https://cdn.elifesciences.org/articles/76269/elife-76269-code1-v2.zip
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Functional visualization of NK cell-mediated killing of metastatic single tumor cells
eLife 11:e76269.
https://doi.org/10.7554/eLife.76269
