A simplified, 96-well–adapted, ATP luminescence–based motility assay

Cell migration is essential for normal and pathological processes such as development, tissue repair, immune function, inflammation, and tumorigenesis. Quantitative analysis of cell migration has been mainly undertaken with the so-called Boyden chamber system, formed of an upper compartment into which cells are injected, and which is separated from a lower compartment by a porous membrane with pores of various size depending on the cell type under study. Cells can migrate from the upper compartment through the porous membrane, along a concentration gradient formed by a chemotactic factor present in the lower chamber. Quantification of migrating cells depends on the cell type. Nonadherent cells that have migrated to the lower compartment are collected and counted by optical means or flow cytometry, which requires high cell numbers. Migrated-adherent (MA) cells remain attached to the underside of the porous membrane, and their quantification typically requires colorimetric or fluorometric assays (1–4). Automated technologies involving optical imaging and specific software is being developed within the industry for screening purposes (5). However, these various assays are often time-consuming, and exhibit limitations due to manual counting and inconsistent results, low signal-to-noise ratio and fluorescence interference, and high cost and the need for specific equipment (5). Most available assays, with the exception of fluorescence labeling assays, are limited to a 6-, 12-, or (at best) 24-well-plate format. Recently, an ATP luminescence–based motility-invasion assay (ALMI) was described that provides a reliable and flexible method to quantify cell motility using a luminescence microplate reader (6). However, the multiple steps and transfer of biological material required in this assay considerably complicate its usefulness and increase the related costs and potential bias. Briefly, in the ALMI assay, the cells remaining inside the chamber of the upper surface of the membrane [nonmigrated (NM) cells] are harvested with a sterile cotton swab, which is than inserted into a well of a 96-well plate containing the ATPLite reagent, and the applicator stick is clipped with pruning shears. To harvest MA cells, the blind well chamber is disassembled and the membrane, devoid of NM cells, is placed in a second well containing the ATPLite reagent. The migrated-nonad-herent (MNA) cells are harvested by resuspending the cells in the lower chamber of the blind well, transferring them to an Eppendorf tube, pelleting the cells in a microcentrifuge, and then aspirating the medium and adding the ATPLite reagent. Here we describe a more simple and direct way to carry out ALMI, which we have adapted to 96-well-plate format and which we call ALMA, for ATP luminescence–based motility assay.

The human breast cancer SKBR3 cell line (ATCC, Manassas, VA, USA) was used to set up the ALMA since this cell line was routinely assayed in the laboratory by means of a colorimetric assay (7). Cells (1 × 10⁵/100 µL) were serum-starved for 24 h (to prepare them for optimal response to the chemotactic factor) and distributed in 6.5-mm diameter Transwell inserts (Cat. no. 3422, Corning, Corning, NY, USA) precoated with 25 µg/mL rat tail collagen I (Cat. no. 11179179001; Roche, Meylan, France) (for 2 h at 37°C) and fitted to 24-well plates (Cat. no. 3526; Costar, Avon, France). EGF receptor (EGFR/HER2)–dependent cell motility was then assayed overnight in response to heregulin β (rh NRG-1β1; Cat. no. 396-HB; R&D Systems, Lille, France) and measured by two methods: one using colorimetry and manual counting and the other using ALMA (Figure 1). For the colorimetric method, NM cells from the upper chamber were scraped from the top side of the porous membrane using a cotton swab, followed by cell fixation in 4% paraformaldehyde and staining in 0.1% crystal violet (Cat. no. C3886; Sigma-Aldrich, Lyon, France) (Figure 1, A and B). Cells were counted manually using a microscope (Olympus IMT2; Olympus, France) (Figure 1C). For the ALMA method, migrated cells on the bottom of the insert were detached by incubation in receiver plates containing medium with 0.5% trypsin-EDTA (Cat. no. 25 300-054; Invitrogen, Cergy-Pontoise, France). Optimal cell detachment was checked by crystal violet staining of the membranes, which showed the absence of remaining migrated cells (not shown). Collected cells were then centrifuged (500× g for 5 min; Model no. 5417R; Eppendorf, Le Pecq, France), resuspended in 100 µL medium, and then transferred to an opaque 96-well plate (Greiner Lumitrac 200; Dutscher, Brumath, France), to which 50 µL CellTiterGlo reagent (Cat. no. G7571; Promega, Charboniére, France) was added. The plate was shaken for 2 min and luminescence was recorded after 10 min of incubation time, using a microplate luminometer (Centro LB960; Berthold, Thoiry, France) (Figure 1, A and D). The compared heregulin-induced cell motility, as determined by colorimetry and ATP luminescence, revealed a similar induction (8.2- versus 8.9-fold, respectively, with a Pearson correlation coefficient of 0.94), indicating that ALMA can provide a convenient alternative method to quantify cell motility as compared with manual counting.

We next simplified ALMA and adapted it for a 96-well plate assay. Indeed, luminescence, as determined using the CellTiter-Glo reagent, correlates linearly with viable cell number, and offers the required sensitivity to detect as few as 20 cells; this amount was not affected by the presence of FCS and the trypsin-EDTA treatment for 30 min at 37°C (Figure 2A; compare the blue and pink curves). In the simplified ALMA (Figure 1A), serumstarved SKBR3 cells (5 × 10⁴/50 µL) were loaded into a HTS Transwell 96-well support (8-µm pore polyester membrane, Cat. no. 3374; Corning) coated for 2 h at 37°C with 25 µg/mL rat tail collagen I (Roche) and allowed to migrate overnight in a feeder well plate (Cat. no. 3383; Corning). The experiment was performed in parallel using two plates, to see how ALMA compared with manual counting. After this incubation period, the HTS Transwell insert was transferred to an opaque 96-well feeder plate (Cat. no. 3583; Corning) containing prewarmed (37°C) trypsin-EDTA for 30 min (this incubation time and/or temperature of incubation can be adapted, depending on the cell type or specific experimental conditions under study). FCS (10%) was added to inactivate trypsin, followed by the CellTiter-Glo reagent. Luminescence was recorded 10 min later using a microplate luminometer (Centro LB960; Berthold). Using this simplified ALMA protocol, heregulin (0.01–1 nM) induced a dose-dependent increase of SKBR3 cell migration (Figure 2B). Heregulin-induced SKBR3 cell migration was inhibited in a dose-dependent manner upon incubation of cells in the presence of gefitinib (0.5–15 µM; Astra Zeneca France, Rueil-Malmaison, France), a kinase inhibitor of the EGFR tyrosine kinase (Figure 2C). This indicated that the simplified ALMA protocol can be used to evaluate pharmacological inhibition of cell motility. To discriminate compounds acting specifically on cell migration from compounds that affect cell viability or display other pharmacological effects (such as ATP generation), a cell viability assay can be developed in parallel using CellTiter-Glo and the same experimental conditions as those developed for the ALMA protocol. Only those compounds affecting cell migration are considered hits. The simplified ALMA was also applied to monitor cell motility of other adherent cells, such as T47D, a breast tumor–derived cell line (Figure 2E).

The migration of nonadherent cells can also be determined by adding CellTiterGlo reagent to the lower chamber. This was illustrated using the widely-used murine BaF3 B cells and human Jurkat T cells (ATCC, Manassas, VA, USA), whose migration toward a FCS gradient was evaluated using ALMA (Figure 2F). Results indicate the possibility of using ALMA to study leucocyte migration. The simplified ALMA protocol described herein is very competitive, direct, and simple to carry out, thereby making it possible to develop quantitative functional screening assays while taking full advantage of bioluminescence (such as its sensitivity, linear dynamics of response, signal-to-noise ratio, and readability). Quantification of cell migration can be presented as (i) a relative light unit (RLU); (ii) the percent of migrated cells calculated by measuring, in parallel, the RLU produced by the total amount of cells loaded in the upper chamber, in the same culture conditions (e.g., Trypsin-EDTA and FCS), using CellTiter-Glo; or (iii) as migrating cell number per square millimeter of porous membrane, which can be obtained by developing, in parallel, a standard curve formed of increasing number of cells directly injected in the reading plate and analyzed using CellTiter-Glo (RLU as a function of cell number). Alternatively, the NM cell population in the upper chamber can be determined by adding CellTiter-Glo reagent, harvesting the lysed cell suspension, transfering it to a 96-well plate, and then recording the luminescence. An accurate measure of cell migration can then be determined as the percentage of MA and MNA cells (as determined in the lower chamber) of the total cell population (e.g., NM + MA + MNA cells), similarly to De la Monte et al. (6). Using this latter methodology, we examined the effect of overnight incubation with 1 nM heregulin as chemotactic factor on SKBR3 cell migration, and the percentages of NM, MA, and MNA cells were determined (Figure 2G). With the use of a microplate reader, this assay allows the user to obtain results from multiple replicate and various experimental conditions with a very limited number of experimental manipulations, in less than one hour.