Differential tumor biological role of the tumor suppressor KAI1 and its splice variant in human breast cancer cells

The tetraspanin and tumor suppressor KAI1 is downregulated or lost in many cancers which correlates with poor prognosis. KAI1 acts via physical/functional crosstalk with other membrane receptors. Also, a splice variant of KAI1 (KAI1-SP) has been identified indicative of poor prognosis. We here characterized differential effects of the two KAI1 variants on tumor biological events involving integrin (αvß3) and/or epidermal growth factor receptor (EGF-R). In MDA-MB-231 and -435 breast cancer cells, differential effects were documented on the expression levels of the tumor biologically relevant integrin αvß3 which colocalized with KAI1-WT but not with KAI1-SP. Cellular motility was assessed by video image processing, including motion detection and vector analysis for the quantification and visualization of cell motion parameters. In MDA-MB-231 cells, KAI1-SP provoked a quicker wound gap closure and higher closure rates than KAI1-WT, also reflected by different velocities and average motion amplitudes of singular cells. KAI1-SP induced highest cell motion adjacent to the wound gap borders, whereas in MDA-MB-435 cells a comparable induction of both KAI1 variants was noticed. Moreover, while KAI1-WT reduced cell growth, KAI1-SP significantly increased it going along with a pronounced EGF-R upregulation. KAI1-SP-induced cell migration and proliferation was accompanied by the activation of the focal adhesion and Src kinase. Our findings suggest that splicing of KAI1 does not only abrogate its tumor suppressive functions, but even more, promotes tumor biological effects in favor of cancer progression and metastasis.


Recording cell motion
Recordings of the breast cancer cell motion were performed using a stage top mini-incubator (INUSFP-MED-F1-PT, Tokai Hit Co., Ltd., Shizuoka, Japan), which can maintain the cells in the well plate in a humidified atmosphere of 95 % (v/v) air / 5 % (v/v) CO 2 at 37 (±1.0) °C. The temperature of the culture medium can be directly monitored by setting a thermocouple into the culture medium and then controlled by the temperature feedback system of the incubator (Tokai Hit). This mini-incubator was mounted on an inverted microscope (Eclipse Ti, Nikon) with an x-y scanning stage (Bios-T, Sigma Koki, Tokyo, Japan), which can control the culture plate position on a microscope stage within a resolution of ±100 nm. The imaging camera was a low noise, high-resolution (2758×2208 pixels), 12bit, 27fps, XCL-s600 (Sony Corporation). The SI8000 software can perform an auto-focusing procedure before each image is captured in order to compensate for focus drift.
Herewith, cell movement (such as velocity, acceleration, and frequency) may be detected and quantified over time periods from ms to tens of days [65,66]. For the motion detection in the present study, typically a frame integral of 1, lower-and upper-contrast thresholds of 9 and 255, and a mesh size of 1 were used. For normalization of motion analysis by motion area, a time filter size of one frame was used.

Motion vector analysis
Motion vectors of breast cancer cells in wound scratch assays were obtained using a block matching algorithm essentially as described elsewhere [67,68]. The use of a block-matching algorithm allows detection of cell movements at the submicron level with high temporal and spatial resolution [67,68]. Briefly, each frame was divided into square blocks of N×N pixels.
Then, for a maximum motion displacement of w pixels per frame, the current block of pixels was matched to the corresponding block at the same coordinates in the previous frame within a square window of width N+2w. Optimal values of N and w for the motion detection of breast cancer cells may vary with the observation magnification and the resolution of the employed camera. Here, we set N = 16 and w = 4 that were determined empirically based on the throughput speed of calculation and accuracy of the block matching detection. The best match on the basis of a matching criterion yielded the displacement of each block. The mean absolute error (MAE) was used as the matching criterion.
The matching function is given by where f t (m, n) represents the intensity at coordinates (m, n) in the current block of N×N pixels and f t-1 (m+i, n+j) represents the intensity at new coordinates (m+i, n+j) in the corresponding block in the previous frame.
We performed above calculation for every 4×4 pixels in the frame with 2752×2200 pixels, and obtained 378400 motion vectors ((2752×2200 pixels)/(4×4 pixels)). Spatial average of the motion-vector magnitude was defined by the following equation: where |V| = absolute value of motion vector, N ROI = number of valid motion vectors in ROI, and x i and y i represents the components the ith vector. By plotting the special average |V| against time, we can obtain information regarding the propagation rate or wound-scratch gapclosure rate of the breast cancer cells. Since we average the magnitude of the motion vectors as shown in the formula above, motion in any direction gives positive values. In the present paper, we simply termed "average of the magnitude of velocity" for the special average |V|. Wound-scratch gap-closure rate parameters were evaluated from the average of motion waveforms obtained during time lapse measurements of several hours to several days.