Tumor cell migration and invasion are regulated by expression of variant integrin glycoforms

Abstract

The ST6Gal-I glycosyltransferase, which adds alpha2-6-linked sialic acids to glycoproteins, is overexpressed in colon adenocarcinoma, and enzyme activity is correlated with tumor cell invasiveness. Previously we reported that forced expression of oncogenic ras in HD3 colonocytes causes upregulation of ST6Gal-I, leading to increased alpha2-6 sialylation of beta1 integrins. To determine whether ras-induced sialylation is involved in promoting the tumor cell phenotype, we used shRNA to downregulate ST6Gal-I in ras-expressors, and then monitored integrin-dependent responses. Here we show that forced ST6Gal-I downregulation, leading to diminished alpha2-6 sialylation of integrins, inhibits cell adhesion to collagen I, a beta1 ligand. Correspondingly, collagen binding is reduced by enzymatic removal of cell surface sialic acids from ras-expressors with high ST6Gal-I levels (i.e., no shRNA). Cells with forced ST6Gal-I downregulation also exhibit decreased migration on collagen I and diminished invasion through Matrigel. Importantly, GD25 cells, which lack beta1 integrins (and ST6Gal-I), do not demonstrate differential invasiveness when forced to express ST6Gal-I, suggesting that the effects of variant sialylation are mediated specifically by beta1 integrins. The observation that cell migration and invasion can be blocked in oncogenic ras-expressing cells by forcing ST6Gal-I downregulation implicates differential sialylation as an important ras effector, and also suggests that ST6Gal-I is a promising therapeutic target.

Figure 1. Generation of cell lines with downregulated ST6Gal-I

A) Oncogenic ras-expressing HD3 cells (control) were transduced with lentiviral vectors expressing one of five different anti-ST6Gal-I shRNA sequences. Control cell lysates or shRNA-expressing cell lysates (“sh.ST6”, samples 1-5) were Western blotted for ST6Gal-I. All of the shRNA-transduced cells showed downregulation of ST6Gal-I, with shRNA sequences 2 and 4 giving the greatest suppression of enzyme levels. Cells transduced with sequence #4 (representing a pooled population of clones) were used for all further experiments. B) Cells were incubated with FITC-conjugated SNA and evaluated by flow cytometry. The green lines represent labeling by a nonspecific IgG-FITC conjugate (negative control), whereas the red lines reflect SNA labeling. As shown, the SNA labeling of sh.ST6 cells was shifted to the left relative to control cells, indicating a lower level of cell surface α2-6 sialylation.

Figure 2. Downregulation of ST6Gal-I results in decreased sialylation of β1 integrins

A) Lysates from either control cells or cells expressing shRNA for ST6Gal-I (sh.ST6) were incubated with agarose-conjugated SNA-1, a lectin specific for α2-6 linked sialic acids. Glycoprotein/lectin complexes were precipitated by centrifugation. Glycoproteins were subsequently released from complexes by boiling in SDS-PAGE buffer, resolved by SDS-PAGE, and immunoblotted to detect β1 integrins (upper panel). In the lower panel, whole cell lysates were blotted for the β1 integrin to assess total levels of β1 expression. B) Bands on immunoblots of SNA precipitates or whole cell homogenates were quantified using densitometry and the ratio of α2-6 sialylated β1 integrin to total β1 integrin was plotted (of note, the “mature” band was used to obtain the value for total β1 integrin expression, since it is the mature isoform that is a substrate for sialylation). Results represent the means and SEMs of densitometric ratios generated from blots of 4 independent sets of cell lysates. C) Cells were incubated with a glycosylation-insensitive anti-β1 antibody and then subjected to flow cytometry (red lines). A nonspecific IgG was used as an isotype control (green lines). These results show that equivalent amounts of β1 integrin are expressed on the cell surface. D) β1 integrins were immunoprecipitated from control and sh.ST6 cell lysates, then blotted with MAA, a lectin specific for α2-3-linked sialic acids.

Figure 3. Cells expressing hyposialylated integrins adhere poorly to collagen I

A) Control or shRNA-expressing cells were seeded onto dishes coated with collagen I, and adhesion was quantified by a standard colorimetric assay. Cells with downregulated ST6Gal-I expression (sh.ST6), show an approximately 2 fold-reduction in adhesion to collagen I. Results represent means and SEMs of 3 independent experiments performed in triplicate. * = p<0.05 B) Cell surface sialic acids were removed from control HD3 cells (i.e., ras-expressing, but not transduced with shRNA for ST6Gal-I) by incubating with Vibrio cholerae sialidase using a concentration range of 1-100 mU/ml. After washing with PBS to remove the sialidase, the cells were seeded onto tissue culture dishes that had been precoated with collagen I. Cell adhesion was quantified using standard procedures. The binding of untreated (control) cells was set at 100% to minimize baseline variation. Where indicated (*), the adhesion to collagen of sialidase-treated cells was significantly less than that of control cells (p<0.05). Values represent means and SEMs of 3 independent experiments performed in duplicate or triplicate.

Figure 4. Cells with hyposialylated integrins demonstrate reduced migration on collagen I and decreased invasion through Matrigel

A) Control and shRNA-expressing cells were seeded into the upper wells of Boyden chambers lined with 8.0-µm PET membranes coated on the underside with collagen I. The lower chambers contained medium plus 2% FBS as a chemoattractant. After 24 hours, migration along a collagen I concentration gradient (haptotaxis) was measured using the vendor's protocol. As shown, the haptotactic migration of sh.ST6 cells was significantly less than that of control cells. (p< 0.05). Results represent means and SEMs of 3 independent experiments performed in triplicate. B) HD3 cells were seeded onto the tops of growth factor-reduced MatrigelTM invasion chambers (BD Biosciences) and allowed to invade for 48 hours. NIH3T3-conditioned medium was added to lower chambers as a chemoattractant. Cells invading through the Matrigel layer, and adhering to the underside of the membrane, were stained and quantified by a standard crystal violet staining method. As shown, expression of shRNA for ST6Gal-I significantly inhibits invasion (p<0.05). Results represent means and SEMs of 3 independent experiments performed in triplicate.

Figure 5. Differential α2-6 sialylation does not alter the invasiveness of β1 integrin-null GD25 cells

A) GD25 cells were stably transfected with ST6Gal-I enzyme tagged with a V5 epitope for easy detection. Western blots revealed the presence of the V5 tagged-protein in ST6Gal-I transfected cells (ST6), but not in parental cells (Par). B) Lysates from parental (Par) and ST6Gal-I transfected (ST6) GD25 cells were probed with SNA lectin to identify and compare α2-6 sialylated proteins. As shown, the ST6Gal-I transfected cells had a significantly higher number and amount of α2-6 sialylated proteins when compared to the parental cells, indicating that the transfected ST6Gal-I was functional. C) Cell invasion was monitored for GD25 cells, which are known to lack expression of the β1 integrin. Briefly, parental GD25 cells (which have low levels of endogenous ST6Gal-I), or GD25 cells forced to express ST6Gal-I, were monitored for invasion through Matrigel as described previously. No significant difference was observed in invasion for the GD25 cells, suggesting that expression of the β1 integrin is required for the differential effects of sialylation on cell invasion. Results represent means and SEMs of 3 independent experiments performed in triplicate.