Migration Dynamics of Human NK Cell Preparations in Microchannels and Their Invasion Into Patient-Derived Tissue

Abstract

Natural killer (NK) cells are characterised by their ability to attack cancer cells without prior antigen stimulation. Additionally, clinical trials revealed great potential of NK cells expressing chimeric antigen receptors (CARs). Successful anti-tumour efficacy remains limited by migration and infiltration to the tumour site by NK cell preparations, which is linked to the scarcity in the knowledge of migration dynamics and invasion potential. Here, we applied a recently reported innovative microfluidic microchannel technology to gain insight into the intrinsic motility of NK cells. We assessed the baseline activated and proliferating NK cells in direct comparison with T cells and investigated their motility patterns in the presence of tumour cells. Additionally, we performed high-resolution 4D confocal imaging in patient-derived hyperplastic lymphatic tissues to assess their invasive capacity. Our data revealed that the invasion potential of NK cells was greater than that of T cells, despite their similar velocities. The flexibility of the NK cell nucleus may have contributed to the higher invasion potential. The motility of CD19-CAR-NK cell preparations was similar to that of non-transduced NK cells in hyperplastic lymphoid tissue, with improved targeted migration in tumour tissue, suggesting the suitability of genetically engineered NK cells for difficult-to-reach tumour tissues.

Keywords: CAR‐NK; DLBCL tissue; NK cells; T cells; hyperplastic lymphoid tissue; microchannel; motility.

FIGURE 1

Expansion and functionality of NK and T cells. (A) Graphical representation of the study workflow: NK and T cells were isolated from identical healthy human donors (buffy coats) and expanded for 7–21 days. Following quality controls, analysis of cell motility using a microchannel‐based assay and ex vivo 4D‐imaging of patient‐derived adenoid tissue was performed. (B) Flow cytometry‐based evaluation of cell purity of CD56⁺ NK and CD3⁺ T cells (n = 4). (C) Expansion of NK and T cells (n = 2). (D, E) Cytotoxic anti‐tumour functionality of NK cells against (D) NALM‐6 and (E) TMD8 at an E:T ratio of 1:1 and 2:1 after 20 h of co‐cultivation (n = 8–10).

FIGURE 2

NK and T cell motility in the presence of NALM‐6 and TMD8 in fibronectin‐coated microchannels. (A) Schematic illustration of the microfluidic microchannel assay. NK/T cells and tumour cells (NALM‐6 or TMD8) were seeded in their respective reservoir of the microchannels. NK/T cell motility was measured for 20 h with 4 min imaging intervals. Exemplary images of (B) the microchannel frame and (C) NK and T cells in the microchannel under basal conditions or in the presence of tumour cells. (D) The number of NK and T cells entering the microchannel under basal conditions (n = 10) or in the presence of NALM‐6 (n = 8) or TMD8 (n = 6). One donor is represented by n = 1–3 data points. Velocity (μm/min) of NK and T cells under (E) basal condition (n = 10), in presence of (F) NALM‐6 (n = 8) or (G) TMD8 (n = 6). One donor is represented by n = 1–3 data points. Statistical analysis was performed using a mixed‐effects model with Turkey's multiple comparison (number of tracks) or Wilcoxon matched‐pairs signed rank test (velocity). Statistics for the velocity data were calculated using the weighted averages of all donors. Statistical significance thresholds were set to *p ≤ 0.05, **p ≤ 0.005, ns p > 0.05 (p values are indicated), (D) if not indicated no significant differences were observed.

FIGURE 3

NK and T cell motility in patient‐derived human hyperplastic lymphatic tonsil tissue. (A) Schematic illustration of ex vivo 4D‐imaging of patient‐derived human hyperplastic lymphatic tonsil tissue. NK and T cells were added simultaneously to the tissue slices and imaged using 4D fluorescence microscopy. Videos were taken from different sections of the same tissue slice (n = 4–8), and all visible NK and T cells were tracked. (B) Number of NK and T cells infiltrating the tissue. Each data point represents data from a single video (n = 4–8), taken of a distinct section of a certain tissue slice with NK and T cells (n = 9). (C) Velocity (μm/min) of NK and T cells in hyperplastic lymphatic tonsil tissue from multiple videos (n = 4–8) of different sections of the same tissue slice. (D) Track length (μm) and (E) displacement (μm) of NK and T cells in hyperplastic lymphatic tonsil tissue. (C–E) Each data point represents the weighted average of all tracked cells, derived from one NK or T cell donor (n = 9) from multiple videos (n = 4–8) of different sections of the same tissue slice. (F) Exemplary image of adenoid tissue slice with endogenous CD19⁺ lymphatic cells (magenta) following addition of primary NK (blue) and T cells (green). Tracks of NK and T cells are indicated by thin lines. Individual tracks of (G) T cells and (H) NK cells from one representative donor tracked in hyperplastic lymphatic tonsil tissue. (I) Representative image of NK and T cells stained with DAPI (blue) and an anti‐lamin A/C antibody (magenta). (J) Percentage distribution of NK and T cells (n = 4) showing a high or low area of nuclear lamin. The cut‐off for low lamin was set to ≤ 30% and for high lamin to > 30%. Statistical analysis was assessed using the Wilcoxon matched‐pairs signed rank test for the number of tracks, velocity, track length, and displacement. Statistical significance thresholds were set to *p ≤ 0.05; **p ≤ 0.005; ns p > 0.05 (p values are indicated).

FIGURE 4

Motility of CD19‐CAR‐NK and NT NK cells under basal conditions in microchannel and in patient‐derived human hyperplastic lymphatic tonsil tissue. (A) Schematic illustration of the workflow for SB‐generated CD19‐CAR‐NK cells. (B) Anti‐tumour functionality of CD19‐CAR‐NK cells from one representative donor after 4 h of co‐culture with NALM‐6. (C) The number of cells entering the microchannel under basal conditions (n = 7). One donor is represented by n = 2 data points. (D) Velocity (μm/min) of CD19‐CAR‐NK and NT NK cells under basal conditions (n = 7) in microchannels. Each donor is represented by n = 2 data points. (E) Number of CD19‐CAR‐NK and NT NK cells infiltrating patient‐derived human hyperplastic lymphatic tonsil tissue (n = 6). Each data point represents the number of infiltrating cells from a tissue section over the course of one video. (F) Velocity of CD19‐CAR‐NK and NT NK cells in μm/min (n = 6). (G) Track lengths of CD19‐CAR‐NK and NT NK cells in μm (n = 6). (H) Displacement of CD19‐CAR‐NK and NT NK cells in μm (n = 6). (E‐H) Each donor is shown as n = 2–8 data points, representing for technical replicates performed on different sections of the same tissue slice. Individual tracks from one representative donor of (I) NT NK cells and (J) CD19‐CAR‐NK cells. (K) Exemplary images of a tissue slice containing endogenous CD20⁺ cells (magenta), CD19‐CAR‐NK cells (red) and NT NK cells (blue). Individual tracks of CAR‐NK and NT NK cells are indicated as thin lines (red and blue, respectively). (L) Mean duration of contacts with endogenous lymphatic cells in minutes (n = 3). Each donor is represented by n = 4–7 data points. Statistical analysis of the number of tracks was performed using the Wilcoxon matched‐pairs signed rank test with all single data points. The data on velocity, track length, displacement and duration of contacts were statistically assessed using the Wilcoxon matched‐pairs signed rank test with the weighted average of each donor. Statistical significance thresholds were set to *p ≤ 0.05; ns p > 0.05 (p values are indicated).