Lectin-mediated reversible immobilization of human cells into a glycosylated macroporous protein hydrogel as a cell culture matrix

Abstract

3D cell culture is a helpful approach to study cell-cell interaction in a native-like environment, but is often limited due the challenge of retrieving cells from the material. In this study, we present the use of recombinant lectin B, a sugar-binding protein with four binding cavities, to enable reversible cell integration into a macroporous protein hydrogel matrix. By functionalizing hydrogel precursors with saccharose, lectin B can both bind to sugar moieties on the cellular surface as well as to the modified hydrogel network. Confocal microscopy and flow cytometry analysis revealed cells to be integrated into the network and to adhere and proliferate. Furthermore, the specificity and reversibility was investigated by using a recombinantly produced yellow fluorescent - lectin B fusion protein and a variety of sugars with diverging affinities for lectin B at different concentrations and elution times. Cells could be eluted within minutes by addition of L-fucose to the cell-loaded hydrogels to make cells available for further analysis.

Figure 1

Decoration of cells and hydrogels with YFP-LecB and biocompatibility. (A) The glycosylation of BSA was determined with an acid-schiff reaction. BSA was analysed on a SDS page (A top) and a glycoprotein detection gel (middle) for different glycosylation times (30, 60 and 120 min) where glycosylation is visualized by magenta bands in the gel (shown are only the relevant parts from full length gels in Fig. S4 Supplementary Information). Confocal images (lower panel) of YFP-LecB decorated hydrogels for untreated and glycosylated (30, 60 and 120 min) hydrogels which were incubated with YFP-LecB, followed by subsequent washing. Yellow color represents the bound YFP-LecB. (B) Lung cancer cells (A549) were incubated with YFP-LecB, washed and visualized with confocal microscopy to reveal lectin binding to the cells whereas red color represents the rhodamine-phalloidin stained cytoskeleton, blue the DAPI stained cell nucleus and yellow the YFP-LecB. (C) Cells were incubated with YFP-LecB and SytoxBlue to reveal the extent of YFP-LecB binding to the cells and its biocompatibility; YFP-LecB was proved to be biocompatible (overall viability > 95% after 24 h) and to bind to nearly all cells (bottom right).

Figure 2

Lectin-mediated cell adhesion. (A) 2*10⁵ cells were seeded onto a hydrogel surface in the absence or presence of different YFP-LecB concentrations from 0 to 400 µM. Cells could adhere to the surface in the presence of cell culture medium for 24 h, followed by staining of cell cytoskeleton (red phalloidin-rhodamine) and nucleus (blue DAPI). Cellular surfaces were analysed with confocal laser scanning microscopy. Adhesion was detected for YFP-LecB concentration of 300 µM and higher. (B) Typical A549 cells in the presence of 50 and 400 µM of YFP-LecB. All bars represent the standard deviation. The significance was tested with a one way anova with alpha = 0.05.

Figure 3

Hydrogel structure. Hydrogels were prepared by mixing both glycosylated protein backbone and linker, followed by polymerization. Hydrogels were incubated for 10 min with 300 µM YFP-LecB (red color) to show the protein binding to the glycosylated matrix. Hydrogels were visualized with confocal laser scanning microscopy. (A, Top left) Furthermore, hydrogels were freeze-dried to generate pores by evaporation of ice crystals from the matrix and analysed with confocal imaging (A, top right). To test interconnectivity of pores and specificity of YFP-LecB binding, hydrogels were diffused by a green fluorescent protein which is structurally similar to the yellow fluorescent protein. 200 µM of GFP could freely diffuse throughout the whole network and was only localized in the pores (A, bottom left and right). (B) A 3D image of a macroporous hydrogels in the presence of GFP.

Figure 4

Lectin release from the hydrogel. To test the reversibility of YFP-LecB from the macroporous hydrogel matrix, gels were decorated with YFP-LecB. Afterwards, decorated hydrogels were incubated with different concentration (0–1000 µM) of lectin binding sugars (D-mannose, L-fucose and D-galactose). After 2 h, the YFP-LecB concentration was measured via fluorescence in the supernatant. (B) To investigate the time which is needed to elute lectin from the hydrogel, hydrogels decorated with YFP-LecB were incubated with 400 µM of L-fucose and at regular intervals samples were taken from the supernatant and analysed for YFP-LecB fluorescence. All bars represent the standard deviation.

Figure 5

Release of A549 cells from the hydrogels. Hydrogels were polymerized and freeze-dried as described earlier and decorated with YFP-LecB. Cells were seeded into the hydrogels and incubated for 2 h to guarantee cell adhesion. (A) Cell loaded hydrogels were incubated with different concentrations of L-fucose and about 90% of cells could be released with 100 µM of L-fucose. (B) YFP-LecB after elution of the cells. Flow cytometry analysis was conducted, revealing that about 50% of all cells still bear YFP-LecB on the surface. (C) Cell-loaded gels were incubated with 100 µM of L-fucose, samples were taken at regular intervals and cells were counted. After approx. 10 min, 90% of the cells were eluted from the matrix. (D) Biocompatibility of materials and procedures used. Cells were eluted from the hydrogel and cytotoxicity was investigated with flow cytometry. (E) Cells were seeded and grown for 3 days in the hydrogels, eluted with 100 µM of L-fucose and counted with a Neubauer counting chamber to see proliferation of cells. (F) Cells were eluted from the hydrogel with 100 µM of L-fucose and re-cultured under 2D conditions over 4 days (left side) and the growth was compared with a fresh control (right side). (G) Adhesion of eluted cells after being re-cultured in 2D. The right side shows the average surface area of re-cultured cells compared to a fresh control. The right side shows the actin filament of re-cultured and fresh cells to observe possible differences in their morphology. All bars represent the standard deviation. The significance was tested with a one-way ANOVA with alpha = 0.05 for Fig. 5A, E and G and with a two-way ANOVA for Fig. 5F with alpha = 0.05.