Radial Mobility and Cytotoxic Function of Retroviral Replicating Vector Transduced, Non-adherent Alloresponsive T Lymphocytes

Tumor cells in a single cell suspension were pipetted into the wells of a Teflon-masked slide. The cells were allowed to adhere and then formed monolayers in a humidified 5% CO₂, 37°C incubator (Figure 1A). Established monolayers or ECM proteins derived from the monolayer could be harvested for these assays (Figure 1B). Effector T lymphocytes labeled with vital fluorescent dyes or transduced with vectors coding for EMD were seeded into the center of the well using a CSM. The ability of non-adherent effector T lymphocytes to migrate on the monolayer or ECM proteins over time was then evaluated and quantified with light and fluorescence microscopy. Here, we evaluated the migration and cytotoxicity of non-adherent effector alloCTL on a monolayer of partially-relevant target U-87MG glioma cells (Figures 3 and 5). These target cells, displaying only some of the Human Leukocyte Antigen (HLA) molecules (HLA-A2, B44) found on the original stimulators, were used to reduce the potent lytic activity that would be displayed by the alloCTL to the relevant 13-06-MG target cells (HLA-A1,2 and B44,57). We also observed the mobility of the alloCTL on ECM harvested from a monolayer of MDA-MB-231BR breast tumor cells (Figure 6). The alloCTL were generated by one-way MLTR according to methods previously published⁹. Effector alloCTL were harvested 4-5 days after one-way MLTR and transduced with RRV-GS4-EMD (Figures 3 and 5) or RRV-ACE-EMD (Figure 6)^10,11. Higher transduction levels were achieved with the GS4 vector utilizing a Gibbon ape leukemia virus envelope than with the amphotropic ACE vector. Using this radial migration assay, we noted that both transduced and nontransduced fluorescently-labeled alloCTL had migratory capacity. We also demonstrate that they retain their ability to be cytotoxic after undergoing manipulations for RRV transduction.

The ability of murine alloCTL to migrate was also assessed (Figure 4). alloCTL sensitized to haplotype H-2b/k displayed by TU-2449 cells were stained with eFluor 670 (purple) and a migration assay was performed using TU-2449 transduced with CMV-Straw-IRES-FLUC2 (red). When using fully relevant target, a lower number of alloCTL effector cells were seeded to prevent rapid lysis of the tumor cell targets. Using 1 hr as a baseline, T cells are largely localized at high density in the center of the well as a single cell suspension. The 48 hr micrographs show extensive migration of T cells toward the outer edge of the well and clustering of the alloCTL is readily apparent with destruction of the tumor cells at the site of their initial deposit. Proliferation of the TU-2449 cells is evidenced as the mStrawberry labeled cells increase in density over the well. This is better seen in surface fluorescence plots that were generated to visualize the distribution of purple cells in the wells (Figure 4, bottom panel). AlloCTL mobility is much more evident in the well at 48 hr compared to the 1 hr baseline. Similar to what was seen in the assays utilizing human alloCTL and glioma cells, a clearance of the tumor cells around the area of the initial deposit of effectors is noticeable at 48 hr, suggesting lysis of the tumor target cells.

A critical step in this assay is the growth of healthy monolayers and subsequent harvest of ‘skeletal’ ECM. Seeding of the tumor cells at 2.5 – 5.0 x 10⁴ cells per well, according to step 1.9 above, usually resulted in a roughly even monolayer for the tumor cell types we used. We found it crucial to allow the tumor cells to adhere at RT, otherwise the convection that occurs at 37 °C will push the cells toward the center of the wells. As well, pre-coating of the wells with poly-D-lysine and/or an ECM protein such as fibronectin or collagen can greatly increase success rate. Additionally, we had better success with these assays using monolayer-derived ‘skeletal’ ECM than with individual ECM components (i.e., fibronectin, collagen), which proved much less reliable. It is important to experiment with the number of non adherent cells to deposit through the sedimentation manifold as well, and if using a pure cell type, one might theoretically be able to define a motility coefficient with this assay. Because the well volume is small and some evaporation of medium occurs, the monolayer needs to be replenished daily with fresh medium. Media replacement also prevents accumulation of toxic lactate as the cells grow and divide. Failure to cultivate a healthy monolayer can result in shedding of the monolayer or ECM from the well surface.

The Coomassie blue stained ECM proteins derived from the U-87MG glioma cell line according to step 1.13 above are shown as a visual example of the adherent substrate layer left behind after digestion with Triton X-100 (Figure 1B). ECM production from a breast carcinoma cell line, MDA-MB-231BR cells, often was patchy and thin, resulting in an ECM layer that was more poorly adherent and sometimes lifted from the slide during the assay. The ECM protein adherence to the slide improved when tumor cells were maintained as confluent monolayers for 1-2 days before digestion. Throughout the process, maintenance of the Teflon seal surrounding the well was also important. Keeping the volume low and minimizing exposure of the Teflon to the Triton-X100 solution mitigated damage to the integrity of the Teflon seal.

One adaptation of this protocol might be use of a three dimensional extracellular matrix harvested from sections of OCT embedded tumor tissue¹⁴. However, it should be noted that the distance from the flat bottomed well to the slide surface is only 1 mm; thus, sections exceeding this thickness may be disrupted by manifold placement. Also, a method would need to be devised for making a slight burr hole in the tissue to accept the droplet of medium containing the non-adherent cells.

Preparation of a well-dispersed single cell suspension of the effector T lymphocytes in advance of sedimentation was also imperative. Gentle tituration of the non-adherent T cells to disperse aggregates to a single cell suspension was an essential step right before loading the alloCTL into the channels of the CSM. Use of non-enzymatic cell dissociation buffers should be employed if necessary. Further, the density of effector lymphocytes should be around 1.0 – 2.0 x 10⁵ cells/µl for sedimentation, as described in step 2.8 above. Low cell numbers may result in no sedimentation, while higher cell numbers can cause a large deposit that is easy to disrupt when moving the slide. It is also recommended that the serum concentration in the medium be at least 10% to provide adequate surface tension of the positive meniscus that reduces the possibility of spillover.

An inverted light microscope with fluorescence imaging capabilities and a 4x or 10x objectives is optimal for this assay. The time points for imaging can be determined according to individual experimental inquiries and cell types. The non-adherent CTL in particular migrate from the center of sedimentation toward the edge of the well more quickly on ECM proteins (4-8 hr) than over a cell monolayer (12-72 hr). Partially MHC relevant target cells were preferable to relevant targets for our alloCTL assay to encourage the visualization of motility and reduce cell-mediated cytolysis of the monolayers before significant migration can occur; use of a low effector to target ratio also slows the lysis. The use of partially-matched targets allows for better visualization of RRV transmission to the tumor cells as well.

Use of an inverted microscope affords multiple advantages over an upright scope. An inverted scope allows for imaging through the humidified chamber so that slides do not have to be disturbed, while also maintaining a sterile environment for the cells. Also, imaging through the chamber provides a barrier between the specimen and the technician, allowing for adherence to biosafety level II (BSL II) guidelines when working with human-derived tissue¹⁵.

If using an upright scope, the highest magnification achievable without disturbing the positive meniscus of the well is 20x. Unfortunately, use of a coverslip, or optics kit accessories to the cell sedimentation manifold, is not recommended, as the disruption to the meniscus typically results in the spreading of the sedimented non-adherent cells throughout the well. Imaging over long periods of time may require careful addition of medium to each well to replace that lost to evaporation, although this is more likely to be required during time-lapse imaging.

This protocol is distinct from methods that measure horizontal cell migration of adherent cells in Zigmond Chambers^16,17, the design of which would not allow for the visualization of non-adherent cell migration. Other assays that evaluate horizontal migration toward a chemotactic gradient through a substrate such as agar are generally used to study migration of one cell type^18-20. Use of the CSM provides several advantages to these methods as well as methods that measure vertical migration/invasion of cultured cells, such as in Boyden chambers or Matrigel transwell plates. First, cell-cell contact allows for a determination of target-effector cell interaction and injury by effector cells. The cytolytic ability of alloCTL-RRV-EMD initially stimulated by one-way MLTR with stimulator 13-06-MG tumor cells to the partially MHC-matched U-87MG cell line is evident in the center of the monolayer at 24 and 72 hr following sedimentation (Figures 3 and 5, dashed lines). As well, the degradation of TU-2449 cells by murine alloCTL is visible at 48 hr post-sedimentation (Figure 4). The monolayer is obliterated at the center of alloCTL deposit, but occurs more slowly over the remainder of the monolayer, owing either to the overlap in HLA expression of 13-06-MG and U-87MG in the assays with human cells²¹ or the low starting number of effector cells for the murine glioma. Also, the effector to target density is reduced as the effector alloCTL migrate away from the sedimentation site. In addition to migration and cytolysis, this protocol can be used to visualize the transfer of virus encoding fluorescent proteins. Here, transfer of RRV coding for EMD from the alloCTL to the tumor cells is evident after 72 hr (Figure 5, inset).

The CSM was suitably designed for examining non-adherent cell mobility on adherent cells. It’s design allowed for minimal disruption of the nonadherent cells with the adherent cell monolayer when the manifold was removed. Additionally, the center of sedimentation within the Teflon masked slide is identical in each well of the manifold allowing for reduced variability between experimental replicates. However, other options such as glass cylinders, cloning rings, or other machined devices could be validated for similar purpose.

A limitation of this model is that migration can only be assessed in vitro, which may not be truly reflective of the in vivo migration. Another limitation is that chemokine/chemokine receptor gradients cannot be established using the model. While interactions of immune cells with ECM or individual tumor cells may still occur and depend on what those cells secrete or express, this motility assay would not answer questions on whether the non-adherent cells will migrate specifically to or away from isolated factors.

An advantage of this assay is the ability to determine if a population of non-adherent cells maintain migratory functionality if the lymphocytes have or have not been transduced. In our example, the transduced subpopulation is visible through EMD protein expression, and it is clear that migratory functionality is maintained in this subpopulation. A wider application might be to test the effects of various concentrations of agents or drugs that might affect non-adherent immune cell migration. Also, other hematopoietic cells might be tested for mobility on various substrates or adherent cells derived from normal tissue. Finally, although it is not done here, seeding a combination of adherent cell types, i.e., stromal cells or dendritic cells with tumor, may provide analyses reflective of more complex in vivo environments.