Development of super-dense transfected cell microarrays generated by piezoelectric inkjet printing

Abstract

Super-dense transfected cell microarrays (TCMs) were created by a piezoelectric inkjet printer on a glass substrate that had been grafted with poly(ethylene glycol) (PEG). The micro-spots that contained plasmid and extra-cellular matrix (ECM) protein were separated from one another by a hydrophilic barrier generated by PEG. We successfully constructed the densest TCMs with spots of 50 μm in diameter and 150 μm in pitch.

---

Technologies for high-throughput and high-content analysis have improved dramatically over the past decade and are anticipated to become powerful tools with medical applications in drug discovery, theranostics, and development of regenerative therapies, for example. One form of cell microarray, the transfected cell microarray (TCM),^1,2 wherein plasmid DNA or siRNA, spotted on the surface of a substrate, is reverse-transfected locally into adherent cells, has become a standard tool for parallel cell-based “omics” analysis (cellomics).^3–5 The construction of TCMs of high density should have particular relevance to medical applications. However, the medically relevant exploitation of most current TCMs remains limited because of cross contamination of arrayed reagents and the migration of cells.⁶

It is usually necessary to maintain a distance of 500 μm or more between microspots on a TCM.^1–8 For higher-density microspots, methodological breakthroughs are required to prevent the migration of cells and to limit the diffusion of spotted materials among the micro-spots in the array. The micro-patterning of a glass substrate via generation of a hydrophobic or hydrophilic surface is very effective for the regulation of cell adhesion and prevention of the migration of cells among spots.^9,10 Thus, for example, micro-patterns have been generated by soft lithography⁹ and microfabrication,¹⁰ with subsequent spotting, by a dispenser, of the mixture of the relevant plasmid and the materials for induction of reverse transfection. However, the necessary two-step processes are often complicated and difficult to control and, as a result, they often limit the potential utility of the microarray since the mixture must be spotted on the tiny regions of the micro-pattern for cell adhesion with highly accurate positioning. Moreover, exploitation of the techniques of microfabrication and lithography is not always feasible in life science laboratories.

In this report, we describe the development of a super dense TCM. To create the TCM, we used an inkjet printer to spot a mixture of plasmid, ECM protein, and other reagents for induction of reverse transfection on a glass substrate that had been grafted with PEG. The micro-spots containing ECMs were separated from one another by a hydrophilic barrier generated by PEG, which has proven to be extremely effective in preventing the migration of cells and the cross contamination of reagents among adjacent spots. Moreover, during the generation of the TCM, we spotted the mixture for the reverse transfection of genes and with the ECM required for cell adherence and generation of non-adherent regions simultaneously in one step with an inkjet printer. Our method was very easy and allowed the preparation of super-dense patterns on the TCM. The densest TCM that we prepared had arrays with spots of 50 μm in diameter and 150 μm in pitch.

To construct very dense microarrays for reverse transfection of cells, we spotted 90-pL to 10-nL droplets that contained one of two kinds of plasmid, which encoded Venus and mCherry fluorescent protein, respectively, ECM (fibronectin or type I collagen) and the materials for induction of reverse transfection, namely, DMEM and Lipofectamine™ 2000 reagent (Life Technologies), alternately on the surface of a PEG-grafted glass substrate with a solenoid valve or piezoelectric inkjet printer (Fig. 1). After air drying the substrate, we added a suspension of HeLa cells in the culture medium for seeding and culture of cells. Plasmids encoding Venus or mCherry fluorescent protein were transfected into adherent cells on the contact surface of each spot.

Fig. 1 Schematic illustration of the construction of a super-dense TCM.

ECMs, such as fibronectin² and type I collagen,⁷ promote cell adhesion. Therefore, we constructed cell microarrays to evaluate the optimum concentrations of fibronectin (Fig. 2A) and type I collagen (Fig. 2B) in droplets for successful adhesion of cells to spots. We seeded HeLa cells on glass cover slips, on which spots of drying droplets with various concentrations of fibronectin (0.0005–0.4%) or type I collagen (0.0005–0.01%) had been arrayed on the hydrophilic non-adherent PEG-grafted glass surface. We confirmed that almost all HeLa cells, which rolled freely on the PEG-grafted surface as a result of convection currents in the medium, landed on and adhered to spots of ECM within 1 h of seeding. Therefore, 1 h after seeding, we washed off the suspension of remaining non-adherent cells. We found that the spotting of droplets that contained more than 0.01% fibronectin or 0.01% type I collagen was necessary for adherence of cells to the spots of ECM. Moreover, we confirmed that adherent cells on spots that had been prepared with more than 0.01% fibronectin or type I collagen were unable to migrate across non-spotted regions. When we tested spots of droplets that contained less than 0.005% fibronectin or type I collagen, we found that cells barely adhered to the spots. Spotting of droplets that contained fibronectin turned out to be more effective for the uniform overlaying of cells than type I collagen. In particular, droplets containing a low concentration (0.0005–0.001%) of type I collagen did not dry uniformly and, as a result, cells adhered in a toroid-like pattern. This problem might have been caused by the fact that the hydrophobic surface of triple helix had low affinity for the hydrophilic PEG-grafted glass surface.^11,12

Fig. 2 Phase-contrast images of cell microarrays. HeLa cells adhered to spots that contained A) fibronectin or B) type I collagen on a PEG-grafted glass surface. Ten-nl droplets of pure water containing 0.0005–0.4% fibronectin or 0.0005–0.01% type I collagen were spotted on the substratum and dried in air before seeding of cells. Upper is 4× and lower is 20× magnification.

To confirm the general applicability of our technique with another line of cells, we also tested NBT-L2b cells, which can adhere strongly to appropriate surfaces and migrate very rapidly (2 μm min⁻¹) and unidirectionally.^7,13 We obtained basically similar results with these cells (Fig. S1†).

Next, to construct an optimized TCM on a PEG-grafted glass surface, we evaluated rates of cell death and the efficiency of the reverse transfection of HeLa cells on drying spots of droplets that contained various concentrations (0.01–0.08%) of fibronectin or type I collagen, the plasmid that encoded Venus fluorescent protein, and other materials required for induction of reverse transfection on the TCM (Fig. 3). After 24 h of reverse transfection of the plasmid into HeLa cells on the TCM, we calculated the rate of cell death and the efficiency of reverse transfection of HeLa cells by the analyzing merged phase-contrast images and fluorescence images, which were due to the fluorescence of the Venus protein. On the spots that contained fibronectin, we observed the extension of lamellipodia from many cells, while the extension of lamellipodia was hardly evident on any spots of type I collagen (Fig. 3A). Moreover, the efficiency of the reverse transfection on the spots that contained fibronectin tended to be higher than type I collagen (Fig. 3C). Therefore, the adhesion and the extension of lamellipodia of cells might be important for the efficiency of the reverse transfection. As a result, fibronectin was more suitable for the adhesion of HeLa cells than was type I collagen. Droplets containing higher concentrations of fibronectin were better able to wet and spread on the PEG-grafted glass surface because of the hydrophilic properties of fibronectin in comparison to type I collagen. As a result, the areas of spots were bigger and the number of cells that adhered to a single spot was higher (Fig. 3B). By contrast, droplets containing higher concentrations of type I collagen were less able to wet the surface, because of the hydrophilicity of type I collagen and, as a result, it was difficult to obtain uniformly coated spots.

Fig. 3 Evaluation of the rate of cell death and the efficiency of reverse transfection of HeLa cells. A) Merged phase-contrast and fluorescence images of HeLa cells on spots that contained 0.01–0.08% fibronectin (upper) and type I collagen (lower) on PEG-grafted cover slips (20× magnification). Images of cells obtained by phase-contrast microscopy and fluorescence microscopy, which revealed the pseudo color of Venus fluorescent protein, were merged by an image-processing program (ImageJ). B) Average numbers of adherent living cells per spot (n = 8). Results were derived from phase-contrast images of HeLa cells at 20× magnification. C) Average rates of cell death (black columns) and the efficiency of reverse transfection (gray columns) per one spot (n = 8). Results were derived from phase-contrast and fluorescence images of HeLa cells at 20× magnification.

The efficiency of reverse transfection of HeLa cells on spots that contained fibronectin was higher than when spots contained type I collagen. In the case of spots of 0.01% fibronectin, the rate of cell death was high and cell adhesion was low. The best conditions for the preparation of TCMs for HeLa cells included the use of 0.08% fibronectin (Fig. 3B,C). We also evaluated these conditions using NBT-L2b cells (Fig. S2†). The efficiency of reverse transfection of NBT-L2b was highest on spots that contained 0.04% type I collagen. Thus, the optimal ECM protein for reverse transfection differed between the two cell lines of cells that we tested.

Next, we evaluated the possible cross contamination of arrayed reagents and the migration of cells among spots on the TCM. We recorded time-lapse images of four types of TCM with arrayed spots of droplets that contained fibronectin for 24 h after reverse transfection. We made a mosaic pattern of spots that contained either the plasmid that encoded Venus or the plasmid that encoded mCherry fluorescent protein on four types of TCM to evaluate the cross contamination of arrayed reagents for reverse transfection. We detected the fluorescence of Venus (green) and mCherry (red) inside many cells in each array 24 h after reverse transfection (Fig. 4). On spots of droplets that contained less than 0.04% fibronectin, the reverse transfection of cells with plasmids did occur but some cells (colored yellow) were transfected with both the plasmids that encoded Venus and the plasmid that encoded mCherry as a result of cross contamination of reagents among adjacent spots. By contrast, spots of droplets that contained 0.08% of fibronectin were not associated with any cross contamination. Thus, we were able to control the diffusion of reagents for reverse transfection by using a higher concentration of fibronectin. We also confirmed that no cells moved between spots during time-lapse imaging of four kinds of TCM for 24 h after reverse transfection (Supplemental Movie 1–4†).

Fig. 4 Merged phase-contrast and fluorescence images of HeLa cells on TCMs prepared with A) 0.01, B) 0.02, C) 0.04 and D) 0.08% fibronectin on PEG-grafted cover slips (20× magnification) recorded 24 h after reverse transfection. Fluorescence of Venus and mCherry were recorded with the appropriate excitation and emission filters and green and red pseudo-colors were assigned to Venus and mCherry, respectively. Yellow indicates the expression of both Venus and mCherry fluorescent proteins.

Finally, we attempted to construct a super-dense TCM (Fig. 5). We spotted 90-pL droplets of a mixture that contained the plasmid encoding either Venus or mCherry, 0.08% fibronectin, 0.01% gelatin and materials for induction of reverse transfection on a PEG-grafted glass surface with a piezoelectric inkjet printer (Microjet Corporation, Nagano, Japan). We used a piezoelectric inkjet printer rather than a thermal printer for spotting to avoid the thermal denaturation of the reagents. We successfully constructed super-dense TCMs with arrays of spots of 50 μm in diameter and 150 μm in pitch. We confirmed that from one to four HeLa cells adhered to each spot. We also confirmed the expression of Venus or mCherry fluorescent protein in more than 65% of spots, while we failed to detect the expression of a fluorescent protein in cells on about 35% of spots. We also confirmed the absence of cross contamination among spots.

Fig. 5 Merged phase-contrast and fluorescence images of HeLa cells on a super-dense TCM prepared with 0.08% fibronectin on a PEG-grafted cover slip 24 h after reverse transfection. A) Lower (4×) and B) higher (20×) magnification. Green and red pseudo-colors were assigned to Venus and mCherry, respectively.

In conclusion, we successfully constructed TCMs with, to our knowledge, the highest density of spots reported to date. We call these TCMs “super-dense” TCMs. In TCMs, the adhesion of a couple of cells per spot is most effective, because nuclear transport of a plasmid via the division of transfected cell is necessary for the expression of genes from the plasmid.¹⁴ Our results demonstrated that the adhesion of cells to spots within 1 h is essential for transfection with a plasmid at the interface of a spot and, thus, control of the rate of cell adhesion by the selection of an ECM suitable for each cell line is critical. Preparation of the TCM involves only spotting with an inkjet printer, and transfection on the TCM requires only the seeding of cells. Therefore, generation and exploitation of our TCM is extremely easy and inexpensive. The development of our super-dense TCM should facilitate the screening and functional analysis of genes that are based on cell phenotype and should also provide, in the future, a powerful tool for personalized medical diagnosis and treatment.

Acknowledgements

This work was supported by a Grant-in-Aid for Scientific Research on Innovative Areas from MEXT in Japan and by a grant from AIST in Japan.

References

1. J. Ziauddin and D. M. Sabatini, Nature, 2001, 411, 107–110.
2. T. Yoshikawa, E. Uchimura, M. Kishi, D. P. Funeriu, M. Miyake and J. Miyake, J. Control. Release, 2004, 96, 227–232.
3. J. M. Silva, H. Mizuno, A. Brady, R. Lucito and G. J. Hannon, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 6548–6552.
4. D. B. Wheeler, A. E. Carpenter and D. M. Sabatini, Nat. Genet., 2005, 37, 25–30.
5. A. L. Hook, H. Thissen and N. H. Voelcker, Trends Biotechnol., 2006, 24, 471–477.
6. H. Erfle, B. Neumann, U. Liebel, P. Rogers, M. Held, T. Walter, J. Ellenberg and R. Pepperkok, Nat. Protoc., 2007, 2, 392–399.
7. R. Onuki-Nagasaki, A. Nagasaki, K. Hakamada, T. Q. P. Uyeda, S. Fujita, M. Miyake and J. Miyake, Lab Chip, 2008, 8, 1502–1506.
8. J. K. Rantala, R. Mäkelä, A. Aaltola, P. Laasola, J-P. Mpindi, M. Nees, P. Saviranta and O. Kallioniemi, BMC Genomics, 2011, 12, 162.
9. A. L. Hook, H. Thissen and N. H. Voelcker, Biomacromolecules, 2009, 10, 573–579.
10. F. L. Geyer, E. Ueda, U. Liebel, N. Grau and P. A. Levkin, Angew. Chem., Int. Ed., 2011, 50, 1–5.
11. G. Suarez, M. Veliz and R. L. Nagel, Arch. Biochem. Biophys., 1980, 205, 422–427.
12. D. J. Prockop and D. J. S. Hulmes, ed. P. D. Yurchenco, D. E. Birk, and R. P. Mecham, Extracellular Matrix Assembly and Structure, pp. 47–90, Academic Press, San Diego, 1994.
13. R. Onuki-Nagasaki, A. Nagasaki, K. Hakamada, T. Q. Uyeda, S. Fujita, M. Miyake and J. Miyake, Methods Mol. Biol., 2010, 629, 193–203.
14. K. Hakamada, S. Fujita and J. Miyake, J. Biosci. Bioeng., 2010, 109, 62–66.

---

Footnote

† Electronic supplementary information (ESI) available: Supplemental results, experimental procedures and four kinds of movies of reverse transfection on TCM. See DOI: 10.1039/c2lc40709d

---