In this study, we established a pancreatic cancer cell migration assay system by using the TAXIScan device. We found that coating of scaffolds such as collagen and Matrigel on glass, similar to that in some published studies using other methods, was necessary for successful adhesion and migration. BxPC3 and PANC-1 cells migrated towards LPA in a dose-dependent manner, which was clearly inhibited by an LPA inhibitor, Ki16425. This is the first report of pancreatic cancer cell migration monitored by the TAXIScan system that enables analysis of multiple parameters, including directionality, velocity, and cell morphology. Additionally, this is the first report simultaneously comparing the TAXIScan and Boyden chamber methods. The Boyden chamber method has been used for over 50 years [
29], the limitations of this method have been pointed out by several researchers. In this method, a membrane of 10 μm thickness, having holes of 8 μm diameter (in this study) with random density, separates the upper and lower wells (see Additional file
7). It is thought that cells are able to sense differences in the chemoattractant concentration between these two wells. Although this method appears simple, it has certain limitations. (I) The density of holes may not be uniform. (II) The micro-structure inside the hole, e.g., a micro-channel of 10 μm length × 8 μm diameter, is unknown, and the chemoattractant gradient is not measurable. (III) A large number of cells is necessary for this assay (1.5 × 10
5 cells per well in this study). (IV) A considerable amount of chemoattractant is necessary (500 μL per well in this study), which is expensive. (V) The process of cell migration is not visible. (VI) The device only displays the numbers of migrated cells. (VII) The obtained data may have high background noise. (VIII) The density of cells migrating to the lower side of the membrane is not uniform. A few advantage of this method are as follows: (I) It has a simple structure; (II) the apparatus itself (without coating materials) is inexpensive; and (III) it is well known and widely used. On the other hands, the advantages of TAXIScan are as follows [
14] (see also Additional file
8): (I) it has an uniform micro-channel (260 μm length × 1000 μm width × 8 μm height); (II) the chemoattractant, which is placed on one end of the micro-channel, defuses uniformly through the channel, resulting in a stable concentration gradient [
14]; (III) a small number of cells is required for analysis (100 or less cells per channel); (IV) a small and inexpensive amount of chemoattractant is necessary (1 μL per channel); (V) migrating cells are observable; (VI) images obtained during migration are recorded automatically; (VII) data obtained from this assay including that on morphology, behavior, directionality, and velocity, are more informative. However, some demerits of TAXIScan are as follows: (I) although the running cost is low, the initial cost is high, and (II) it is not well-known yet. In fact, it may not be appropriate to position TAXIScan as an alternate to the Boyden method, because both methods utilize completely different equipment and data collection methods, and the quality of data obtained using these methods is entirely different (Additional files
7 and
8). However, because of lower requirement of samples and the collection of more informative data, the approach to cancer cell migration using TAXIScan is more useful than analysis using existing techniques such as the Boyden chamber method. With the TAXIScan system, the characteristics of pancreatic cancer cells can be analyzed in detail. Moreover, our system can be adopted for migration studies in other types of cancer cells.
In the Boyden chamber method, a certain number of cells without LPA was observed to migrate, indicating a high background (Fig.
2c), similar to that reported previously [
30‐
33]. This high background with the Boyden chamber method is considered to be due to the thickness of the membrane (10 μm in this study). In TAXIScan method, cells without LPA were observed to migrate for more than 10 μm (up to 100 μm) (Fig.
2a), explaining this phenomenon. From this point of view, we could argue that TAXIScan has a wider dynamic range to detect cell migration.
Herein, 4 pancreatic cancer cell lines were analyzed and only 2 of these cell-lines, BxPC3 and PANC-1, showed good migration towards LPA with reasonable co-evidence on the expression of LPA receptors. The reason why AsPC1 and MIAPaCa-2 cells do not migrate towards LPA is still unknown. BxPC3 and PANC-1 do express LPA
1, LPA
2, and LPA
3; however, these cell lines do not express LPA
4, LPA
5, and LPA
6 as observed during western blotting (Fig.
3b). The latter 3 receptors are likely not involved in cell migration but might be involved in other cellular functions.
LPA inhibitor, Ki16425, shown in this study is believed to block human LPA
1 and LPA
3 receptors [
28]; 10 μM of Ki16425 significantly blocked the migration of cancer cells [
13]. In our system, Ki16425 clearly inhibited BxPC3 cell migration towards LPA at 5-50 μM concentrations, indicating that TAXIScan and BxPC3 cells are the best tools for screening inhibitors of pancreatic cell migration. Utilizing such a new method, new molecules for regulating pancreatic cancer metastasis can be identified, and the limited treatment options and the poor prognosis of this disease can be overcome. Studies on neutrophils have tested various kinds of compounds and found that some compounds inhibit neutrophil function, leading to the successful selection of several effective molecules [
34]. Collectively, it can be concluded that the system established in our study can be a powerful tool for cancer research and drug discovery in seeking effectors and inhibitors for analyzing cancer cell function. We are currently looking for and screening such molecules that can regulate pancreatic cancer cell migration; some promising molecules will be reported in the near future.